fire control systems, fire suppression, fire detection, duct protection, fire protection for a comme

专利摘要:
a fire safety system particularly adapted for commercial kitchen applications provides exhaust level control for energy efficiency and intelligently responds to fires or fire hazards. in the modalities, the systems can provide advance warning of fire hazards or imminent fires. the modalities use probabilistic estimates with alarms that can be canceled. the modalities use classifiers that can use alarm cancellations as a mechanism for supervised learning.
公开号:BR112019019443A2
申请号:R112019019443
申请日:2018-03-20
公开日:2020-04-14
发明作者:V Livchak Andrey；W Schrock Derek；A Lyons Gregory；Sandusky Jimmy；Sunderlin Kyle；Harlow Nicholas；j meredith Philip
申请人:Oy Halton Group Ltd；
IPC主号:

专利说明:

FIRE DETECTION SYSTEM
Cross Reference to Related Orders [001] This order claims the benefit of US Provisional Order No. 62 / 473,747, filed on March 20, 2017, which is incorporated herein by reference in its entirety.
Basics [002] The known fire suppression systems used in cooker hoods or stoves are primarily concerned with dispensing flame retardant on the cooking surface to prevent oil or fat fires when an indicative fire temperature is measured in the chamber hood or in the pipeline network. Existing fire suppression systems operate by measuring a fixed absolute temperature in the hood chamber or in the duct network and activating an alarm or flame retardant release when a previously defined temperature has been reached. This type of approach, however, does not take into account changes in the exhaust temperature, nor the scenarios in which there is only a flare of regular cooking, instead of a fire. In addition, fire suppression systems detect only existing fires. There is a need in the prior art for improved fire detection systems, as well as prevention and systems with faster response times.
Summary [003] In modalities, network-based or rule-based methods combine multiple sensor inputs to generate a status indication that is used to control fire suppression and exhaust flow through a single set of sensor inputs. In modalities, at least one type of sensor that generates a predefined signal is used to detect the fire condition and the cooking state of the utensil, the predefined signal being applied to a controller that differentiates, in response to
Petition 870190093460, of 09/18/2019, p. 102/111
2/57 predefined signal, in combination with other sensor signals, at least two cooking states, each of the cooking states corresponding to at least two exhaust flows that the controller implements in response to the controller differentiation of the two states and whose predefined signal is used simultaneously to differentiate a fire condition, in response to the differentiation, the same controller activates a fire suppression mechanism, such as a water jet or chemical fire extinguisher.
[004] Objects and advantages of modalities of the described matter will become evident from the description below, when considered together with the attached drawings.
Brief description of the drawings [005] Modalities below will be described in detail below with reference to the accompanying drawings, where similar reference numbers represent similar elements. The attached drawings were not necessarily drawn to scale. Where applicable, some resources may not be illustrated to assist in describing the underlying resources.
[006] Fig. 1 shows an overview of signal processing and control for fire detection and suppression according to the modalities of the described material.
[007] Fig. 2 summarizes examples of sensors that can be used for one or more of the inputs in the system in Fig. 1.
[008] Figs. 3A-3F show the initial conditioning of different types of sensor signals and additional processing that can be used to reduce the dimensionality of the primary signals while providing meaningful information to a controller classifier.
[009] Fig. 3G shows conditioned inputs resulting from the initial processing of Figs. 3A-3F with an exemplary range of emissions according to the modalities of the described material.
Petition 870190093460, of 09/18/2019, p. 103/111
3/57 used to reduce the dimensionality of primary signals while providing meaningful information to a controller classifier.
[009] Fig. 3G shows conditioned inputs resulting from the initial processing of Figs. 3A-3F with an exemplary range of emissions according to the modalities of the described material.
[0010] Fig. 4A shows a kitchen with fire detection and suppression elements according to the modalities of the described material.
[0011] Fig. 4B shows a plan view of the utensils to show an interstice between the adjacent utensils and an interstice between the utensils and a wall.
[0012] Fig. 5 shows a combined detection and suppression device according to the modalities of the described material.
[0013] Fig. 6 shows a multiple exhaust suppression fire suppression system according to the modalities of the described material.
[0014] Figs. 7 and 8 show duct sections with a fire detection and suppression system according to the modalities of the described material.
[0015] Figs. 9A - 9C show a fire simulation device according to the modalities of the described material.
[0016] Figs. 10A and 10B show processed thermal images of a fire and normal cooking, respectively, according to the modalities of the described material.
[0017] Fig. 11 shows an application for a granulated thermal image formation system according to the modalities of the described material. [0018] Figs. 12A and 12B show front and side views of an exhaust hood with associated sensors, according to a modality of the material described.
[0019] Figs. 12C and 12D show symbolic processing diagrams of a fire detection system according to modalities
Petition 870190093460, of 09/18/2019, p. 11/13
4/57 of the subject described.
[0020] Fig. 12E shows, figuratively, sensor magnitudes of a modality of a fire detection system according to modalities of the described material.
[0021] Fig. 12F shows the measured sensor magnitudes of a fire detection system modality according to the described material modalities.
[0022] Fig. 13 shows an aerial view of an exhaust hood and a camera with detectable vapors according to the modalities of the described material.
[0023] Fig. 14 shows a block diagram of an example computer system that can be used as a basis for each or all of the modular computational elements discussed in relation to each modality of the subject described.
Detailed Description [0024] In the described modalities, simple control schemes, or more complex network-based or rule-based methods and algorithms, can combine one or multiple sensor inputs to generate a status indication that is used to control fire suppression and related answers. In addition, the modalities can provide exhaust flow control from the same set of sensor inputs. In modalities, at least one type of sensor that generates a predefined signal is used to detect the fire condition and cooking state of the utensil, the predefined signal being applied to a controller that differentiates, in response to the predefined signal, in combination with other sensor signals, at least two cooking states, each of the cooking states corresponding to at least two exhaust flows that the controller implements in response to the controller differentiation of the two states and whose predefined signal is used simultaneously to differentiate one
Petition 870190093460, of 09/18/2019, p. 11/141
5/57 fire condition, in response to the differentiation, the same controller activates a fire suppression mechanism, such as a water jet or chemical fire extinguisher.
[0025] With reference to Fig. 1, a controller 108 includes a combiner / reducer / classifier 106 that receives one or more conditioned inputs 105 derived from filtered signals 103 which are filtered 105, as by analogue filtering or other type of signal conditioning which may depend on the type of signal from sensor 101. Sensors 100 can be of a variety of types and can generate continuous emission sensor signals 101. Conditioned inputs 105 can combine to detect a fire that can indicate by means of a signal internal (not shown).
[0026] The fire detection indication can be attended to by generating one or more signals that represent a level of confidence associated with the fire indication. In the same way, the controller can also determine a type of fire, a location and a size of the fire, each with a corresponding level of confidence. All of these can be associated with internal indications, such as stored data or signals accessible to the controller 108.
[0027] Controller 108 mediates between fire indications, confidence levels, other data associated with the fire indication, such as location, severity, size, etc. and generates emission signals 109 to drive final emissions that can be applied to emission effectors 110 that implement some actions such as triggering the release of fire suppressant chemicals, selecting and directing the nozzles for the distribution of fire suppressors, selecting the type suppressor, etc. The types and degrees of freedom (DOF) of emission effectors 110 are discussed in the examples below.
[0028] A user interface 112 may provide one or more of the emission effectors 110 or may not be used for emission indications
Petition 870190093460, of 09/18/2019, p. 11/15
6/57 related to fires. User interface 112 can allow entry of modal inputs, such as the type of utensil or utensils being monitored, the status of the utensil, the type of fuel, the location of the utensil and other information related to sensors 100, the distribution effectors of suppressors, and other information that can be combined with the sensor inputs and that can influence the detection of a fire, the response to it and the confidence levels associated with the detection. The interface can include direct digital interfaces to other devices, such as a building management system, device interface, etc.
[0029] A communication interface 111 can provide functionality for certain broadcasts, such as activation of remote alarms, cell phone calls and instant messages. It can also provide Internet connectivity to allow remote control, system software updates, data updates and remote access to a state portal hosted locally or by a remote server park. The communication interface 11 can also provide communication support for the refinement of collaborative system software, where the feedback from the control system is grouped with the feedback from other similar systems and used by a service to improve the control software that can be used. distributed in software updates.
[0030] With reference now to Fig. 4A, an exhaust fan 716 draws air and vapors from a duct 627 connected to an exhaust chamber 624 that supports a filter 626 at an entrance of the same to suck vapors from a recess 622 of a hood. exhaust 621. There may be multiple chambers 624, filters 626 and ducts 627 in a variety of different arrangements and the present is shown as an example only. Several sensors are shown which will be discussed below. A fire suppressor, such as a chemical suppressor, is stored in a 680 pressurized container or other suitable subsystem. A 607 switchgear may, for example, include
Petition 870190093460, of 09/18/2019, p. 11/16
7/57 example, a 607A spray nozzle. Suppressors for water extinguishers, gas suppressants or dry can also be provided.
[0031] A 617A fire extinguisher system with 617A extinguisher heads may also be present in a commercial kitchen space 628. The 617A extinguisher heads may have fusible links that open spray valves when heated for a predefined period above one predefined temperature. Several people 601 can move around in the occupied space of kitchen 628 by performing various activities, including cooking, cleaning, storing, maintaining equipment, etc. The 601 team may also be involved in fire inspection tasks, fire suppression tasks in the event of a fire emergency, evacuation, etc. A cooking utensil, which can be one of many, is indicated in 620. The cooking utensil can generate cooking vapors that are exhausted by the exhaust hood 621.
[0032] Utensils 620 are generally positioned adjacent to each other with an interstice between adjacent utensils, as indicated in 630 in Fig. 4B, or adjacent to a wall that defines a space between the wall and the utensil, as indicated in 631 Such spaces 631 can be traps for dust, grease, old food and many other debris that can fuel a fire and thus create a hazard.
[0033] Controller 108 (600) can continuously or intermittently monitor one or more inputs derived from the various sensors 100, examples of which are collectively illustrated in Fig. 2. The sensors 100 can include temperature sensors, including a temperature sensor of duct gas 629 that can be positioned at a distance from a wall of a duct 627 to ensure that it measures the gas temperature continuously and responsively. Various known types of temperature sensors can be used here and elsewhere in the modalities described throughout the present application. This can include thermocouples, temperature sensors,
Petition 870190093460, of 09/18/2019, p. 11/171
8/57 resistance, resistance temperature sensors (RTDs), quartz oscillator thermometers, thermistors or any other type of temperature sensor.
[0034] Sensors 100 may also include one or more temperature sensors of any of the above types positioned and configured to measure the temperature of the duct wall at one or more locations on the duct wall, as indicated in 625. A temperature sensor Duct wall temperature can indicate heat generated by a fire fueled by deposits in the duct itself. For example, in grill exhaust systems, such as gas or wooden grills, oil or creosote droplets that escape normal filtration can accumulate on the walls of duct 627. A burning ember or several embers can ignite this deposit and cause fires at very high temperatures. These can also be detected by the 629 duct gas temperature sensor and / or instances of it that are located downstream from a location where a duct fire is likely to occur. Generally, a 626 filter, such as a grease filter, will prevent fuel from encrusting the 627 duct walls, but the filters are imperfect. Burning embers can also escape from a filter, such as a grease-type filter (impact type).
[0035] One or more temperature sensors 604 can be provided in the recess 622 of the exhaust hood 621. This can provide an early indication of a thermal peak associated with combustion. The exhaust vapors that accompany cooking and inactive heating of a cooking surface, such as a grill, have a predictable pattern for them. For example, turning the meat on a grill can produce spikes of steam, which produces a high volume of smoke, but once the cooking mode is established, the exhaust flow can be controlled to ensure that the flow is at a rate of predetermined design to be able to handle such a load. However, when burning additional fuel is added to the
Petition 870190093460, of 09/18/2019, p. 11/181
9/57 situation, much higher temperatures can be indicated inside the hood.
[0036] Fryers are another source of possible fires. Frying has a predictable time pattern for vapors rising in recess 622 of exhaust hood 621. When an oil fire occurs, a large temperature rise can occur within recess 622 of exhaust hood 621 in a short time. The provision of multiple temperature sensors 604 can allow the position of the fire along a long hood to be determined. Multiple sensors can also allow the detection of a temperature length scale, since a large fire will be detectable by multiple remote sensors. Temporal and spatial resolution can be used to indicate the size and location of a fire.
[0037] In terms of processing temperature information, the temperature fluctuation time scale can indicate the energy of the thermally driven fire flow. A power spectral density (PSD) function measured cumulatively from a floating temperature signal using a low mass temperature sensor can be analyzed to indicate turbulent energy. In a fire, where a higher turbulent energy is being generated and manifests itself as temperature variations, the components of higher frequency in the PSD indicate a higher turbulent energy. Several examples of fires can be analyzed to identify spectral limits in the PSD, absolute or relative, that indicate a fire. Relative limits can refer, for example, to a power ratio in a low band to that in a predefined band less than the low band that is below a limit. Likewise, several temperature sensors spaced a short distance away can reveal high-intensity turbulence generated by burning gas in a similar way. The scales of duration and time associated with burning can have unique characteristics that
Petition 870190093460, of 09/18/2019, p. 11/191
10/57 can not be distinguished in terms of simply turbulent energy and these characteristics can be discovered experimentally and distinguished in terms of a PSD in any case. A grainy representation of the PSD can provide a small number of DOF that can be applied to a classifier in order to detect fires.
[0038] One or more gas species detectors 658 can be provided to sample and analyze ambient air in order to detect the presence of flammable gas, oxygen levels, carbon monoxide, carbon dioxide, volatile, organic compounds specific volatiles associated with the uncontrolled burning of common materials in kitchen fires and other gaseous species.
[0039] Temperature sensors 604 can be distributed in a rectangular or hexagonal matrix over a two-dimensional field within the recess 622 of the exhaust hood 621. The positions shown in Fig. 4A are only figurative. As indicated above, temperature sensors 604 can be used to indicate a temperature that varies slowly or a fluctuating temperature from which statistics can be derived and used for classification to indicate a fire. Examples of temperature sensors with low thermal inertia are RTDs and thermocouples, as well as thermistors.
[0040] Radiant emissions, or light energy in the thermal range, can be detected and used for fire detection and / or to discriminate a non-fire condition, given other indications by other fire sensors. Sensors 102 may additionally include one or more radiant temperature sensors positioned and intended to detect the average temperature of a region (field of view or FOV), as indicated in 610. There may be multiple 610 radiant temperature sensors facing several regions or FOVs. For example, one indicated in 632 can be directed at a portion of the cooking surface of the
Petition 870190093460, of 09/18/2019, p. 11/20
11/57 utensil 620, while another is positioned to detect flames in the recess 622 of the exhaust hood 621. The FOV can be narrow or wide. In modalities, FOV is selectable. The signal provided by the radiant temperature sensors 610 can be an instantaneous signal in real time from which information can be obtained by the controller from the unstable signal from it.
[0041] The radiant temperature of a region can be spatially resolved by one or more 613 infrared cameras. With brief reference to Figs. 10A and 10B, infrared images digitally processed from different advantages of a fire at a 705 grid station and a normal 706 grid are shown. Several features that can be extracted from the bubble and the fast Fourier transform (FFT) analysis can be seen. First, a closed region at high temperature 710 (bubble in the image analysis vernacular) is defined by the IR luminance that limits the image. A classifier can detect from the area, the shape (simple statistics, such as aspect ratio, temporal persistence and similar characteristics) of the bubble can be extracted in the combiner / reducer / classifier 106, optionally in combination with other sensor inputs, to provide a fire indication and associated data, including confidence level, location, size and intensity. It can also be seen that a region 714 is distinguished by a strong fluctuation of spatially resolved luminance that can be distilled as a statistic, such as a spatial frequency distribution of the region. As indicated above, the energy in the PSD for a visible frequency band would be very high. Although variations in a similar band are visible and widely distributed in Fig. 10B, the variation in luminance amplitude is not as high. In addition, this may not be accompanied by a time fluctuation that, as indicated above, can be captured in a statistic and applied to the controller.
Petition 870190093460, of 09/18/2019, p. 11/211
12/57 [0042] An optical or infrared video imaging device can be mounted in order to detect the emergence of fire or vapors containing hot water vapor from a flat view of a hood. The 599 camera can be selected based on a wide range of optical and near infrared frequencies. A recognition algorithm implemented by controller 108 can recognize fire or smoke from the hood. Vapors and, of course, fire, should not be visible above the 621 hood under normal circumstances. The expansion of the combustible vapor under the hood forcibly exits the regulated exhaust form, so that fire and hot smoke can be easily detected. In addition, the use of an image or video capture device, such as the 599 camera, for this purpose allows quantifying the volume of exhaust gases, if not extreme, to a certain extent, as the radiant temperature and the area of a projection of a radiant plume can be quantified with the help of image processing. With reference to Fig. 13, a hood 470 is shown in plan view with a video camera 471 positioned above the hood to provide an overhead view, for example, locating it close to the ceiling, as shown at 599 in Fig. 4A. The density of vapors can be determined by subtracting the background pattern and the level of infrared illumination if the detected vapors are obscuring, for example, carbon dioxide. Or if there is heat from incandescent gases (fire), they can also be detected at visible and infrared wavelengths. Using processes established in the image processing and measurement fields, the magnitude of opacity or luminance can be quantified for each of the multiple regions 472, 476, 475 and 477 by image processing and analysis to provide an estimate of the vapor depth for each one of the multiple regions 472, 476, 475 and 477, whose areas can also be quantified by image processing and analysis. The projected flow of
Petition 870190093460, of 09/18/2019, p. 11/22
13/57 an exhaust hood may be sufficient to prevent the occasional escape of any substantial plumes of vapors due to occasional spikes in vapors, such as splashing water or fat in a hot cooking source, such as grills fueled by hot fuel. In general, any significant escape of vapors would therefore be the result of unusual misuse of the equipment, fire or failure of the exhaust system. In one method, the controller can delay the emission of a fire indication or the control of an emission effector, such as an alarm or fire suppression system, to provide time for a manual override to be entered by an operator. The controller may, according to the method modality coded by instructions executable by the processor, provide a warning signal indicating that a provisional fire detection has occurred, thus alerting operators to the need for an override input to prevent an alarm emission or deletion.
[0043] Video scene classification techniques can be applied to recognize dangerous situations before they actually generate a fire. Hazard classification can cause a controller to generate a warning using any of the mechanisms described without necessarily or immediately triggering a fire suppression system in response. For example, an infrared image of a bubble in a scene, where the hot bubble is determined to have a temperature that is rising towards a predefined flash point of oil and where no activity is indicated by analyzing the movement of the scene, it would be a simple classification problem that can be defined in a classifier by explicit rules or implemented using supervised learning. Such a scene would be an indication of a possible fire waiting to happen.
[0044] In one embodiment, this fire threat warning system can have a thermal video camera connected to a controller.
Petition 870190093460, of 09/18/2019, p. 11/23
14/57
The controller can be programmed to store and execute steps that implement a scene classifier and image processing to simplify the scene's features to allow quick sorting. For example, the classifier may have an image processor that applies a temperature limit to successive images of the video signal to identify a hot region of a captured scene and also to generate data indicating a temperature in the hot region. The temperature can be monitored over time and its rate of change is estimated to determine if it is changing in a positive direction. The classifier can also estimate the movement in the scene. The motion estimate can be provided by compression chips not directly requested. The magnitude of the movement data in the scene can indicate whether a person is present and on duty or not. If a person is present and moving around the scene, a warning signal may be delayed in relation to a condition in which no movement is present. The classifier can generate a danger signal when the movement in the scene is below a limit and the temperature is above a predefined limit stored by the controller. The controller can activate an alarm in response to the danger signal. As described elsewhere, the controller can accept an override command from a user interface and disable the alarm in response to the override command.
[0045] Fig. 4A shows an override switch 699 that transmits an override command to controller 600. It should be noted that an override command can be applied via any type of user interface 603. It can be provided as a simple switch or a more complex user interface that can take many forms and can also provide additional data that can be used to support the training of the supervised learning algorithm that generates the warning signal or the alarm signal as a result of interpreted conditions fur
Petition 870190093460, of 09/18/2019, p. 11/24
15/57 supervised learning algorithm to indicate a fire. First, a switch like 699 can be located remotely from a likely fire location, for example, on a wall near a door or at least away from kitchen appliances. This can allow immediate unimpeded access to the switch in the event of a small but quickly contained fire. Another way in which an override command can be applied to controller 600 is via a smart phone 668 interface or another mobile interface. In one embodiment, the controller (or a service connected via the network to the user interface) is programmed to issue information about the provisional fire detection and request feedback from the user to help dismantle the conditions that led to the generation of the provisional fire detection. For example, a feedback collection program to allow more information for supervised learning can generate a selectable list of options that can be generated on a display, for example, Q small fire contained manually; Q without fire, unknown cause of provisional fire detection; Q without fire, known cause of provisional fire detection - defective detector; Q without fire, known cause of provisional detection of unusual fire-events (for example, such as construction, crowd of people moving in the kitchen); Q ^without fire, unknown cause of provisional fire detection. The feedback collection program can request more information conditionally based on responses to selections. The training of supervised learning algorithms can still be advanced by simple signs of cancellation.
[0046] Radiant temperature or local imaging can also be used in a duct. Fig. 7 shows a modular unit 650 that can be supported on a tube or other support 653 installed in a duct to detect burning embers floating in it using a 640 infrared camera with a wide angle lens for a wide FOV. a
Petition 870190093460, of 09/18/2019, p. 11/25
16/57 connected controller 644 performs image processing on the images in real time and controls a valve 647 to generate a brief spray from a nozzle 642 to erase the ember or embers. The nozzle 642 can have a rotating head to allow the objective to be optimized at installation. The sprayer can be a water sprayer from a dry or gaseous suppressor. The same device can be installed in installations other than ducts. For example, an independent device, as shown in Fig. 5, to identify small fires or embers, can be placed near an area at risk for this. Here, an example is shown mounted on a pipe to supply water in this example. A rotating nozzle 654 is tilted by actuators 652 and 651 through two axes of movement. A 650 wide-angle infrared camera can identify targets and apply suppressor locally to a region in its field of view. As the wide-angle infrared camera 650 is colocalized with the rotating nozzle 654, any fire or flame that is in the FOV will be treated by the spray cone emitted by the rotating nozzle 654, mapping the FOV properly to the action of the rotating nozzle 654. Note that, in addition to burning embers, the 650 infrared camera can also indicate burnt material accumulated on the duct wall and use a fire suppression sprayer to stop it.
[0047] The restricted targeting of an incipient fire is an advantageous feature of fire suppression for several reasons. First, the fire suppressor can damage a protected installation, such as a kitchen, or at least create dangers and an expensive cleaning problem. Therefore, confinement of the suppressor in the areas that need it is a desirable feature. In addition, due to problems created by the use of the suppressor, this can make facility managers or employees reluctant to allow the system to handle fires when detected. Fire suppression can be defeated or the
Petition 870190093460, of 09/18/2019, p. 11/26
17/57 decreased system responsiveness (through manipulation of setpoints and others) by teams reluctant to suffer its consequences. Another example is where the only option available to personnel is to operate a manual fire switch or confirm the presence of a fire detected by a fire detection and control system. Thus, limited and restricted suppressor application, such as by means of modalities such as those in Figs. 5 and 7, can reduce these concerns.
[0048] Controller 600 can be connected to a remote or mobile user interface, such as a 668 smart phone, via a 667 communications module. The 667 communication module can be a network or Internet interface, such as a modem and can include a router or switch. The communications module 667 can be a transceiver and the mobile user interface, a radio terminal.
[0049] A type of sensor that lies between the radiant temperature sensor at location 208 and an infrared camera 212 is a radiant grainy image sensor 210 that gathers only a small number of pixels with a path sensor. This sensor can have a ten-by-ten array of infrared sensors. The grainy image sensor 210 can have built-in image processing functionality and can produce a low bandwidth data stream that can be filtered more easily in real time. The resolution can be selected to detect variation in a luminance FOV, so that it can detect even certain frequencies that correspond to a limit and not higher. Thus, image processing capable of identifying energy in a region such as 714 can be provided without further resolution. The radiant grainy image sensor 210 can also allow the location of hot regions to be targeted by a selected fixed nozzle or targeting a mobile spray nozzle. Fig. 11 shows an arrangement in which the radiant grainy image sensor 210 with a wide-angle focus optic 731 (which
Petition 870190093460, of 09/18/2019, p. 11/271
18/57 can be simpler in alternative modes (for example, a Fresnel lens) due to the low resolving power of the sensor) is used to control a matrix 733 of separate nozzles 730. When a fire 732 is detected in a zone ( one of zones A, B, C, D, E, F, G), the fire being located in zones DE, which are indicated in region 210 of the radiant grainy image sensor 210, and then a controller opens the nozzle corresponding to this region, that is, the nozzle 735. The granulation of infrared detection calculates the average luminance over a region with minimal image processing. In this modality, the principle of localized and limited suppressor application can be applied.
[0050] In some embodiments, other means by which the suppressor may be limited include phased application in which a first type of suppressor considered more harmless than a second is applied initially and the second type of suppressor is applied only after the detection system indicate a failure to provide complete control. An example of an innocuous suppressor can be water, and a less innocuous one can be a chemical foam suppressor. Another means by which the suppressor can be limited is through the application of the suppressor mediated by feedback, so that, as the suppressor is applied, the effect on the fire is assessed and the amount of suppressor is limited to an amount necessary to decrease the fire condition to a preset detected level. Yet another method of control may include predictive suppressor application, where the total amount of suppressor is responsive to an indication of the size of the fire. This type of control algorithm can be identified as a forward control.
[0051] In some modalities, sensors responsive to light in the visible spectrum can also be used in the fire detection and suppression system. Fig. 2 shows examples of sensors including visible light imaging of the cooking surface 226 that
Petition 870190093460, of 09/18/2019, p. 11/28
19/57 can be obtained by a 619 visible light camera, for example. The luminance of the visible light at the location (without image formation) of a region of a cooking surface 228 can be detected by a photosensor 613, for example. It can have a focus optics or a light guide to protect the light from outside a FOV from being received by the 613 photosensor.
[0052] The luminance detection at the location received from the cooking surface 228 can also provide useful information for the detection of certain types of fires, such as fires with fat or where fat and water are accidentally combined. This can create a large flame that generates radiation of light and heat that is also distinguishable in terms of its time profile. The combination of these two signals, a luminance peak and a thermal peak, can be more informative than either alone, as a peak of visible light can potentially be caused by the reflection of a light source.
[0053] The occupied space of kitchen 628 can be photographed 230 continuously by a visible light camera 612 and / or by an infrared camera 616. One or more cameras of visible light 612 and / or infrared cameras 616 can be positioned in the kitchen to monitor the region adjacent to an appliance, such as appliance 620. In addition or alternatively, one or more visible light cameras 612 and / or infrared cameras 616 can be positioned to monitor various distinct locations around a kitchen, such as several adjacent utensils, doors, occupied regions, components of mechanical systems, such as fan motors, ventilation registers, spaces above the suspended ceiling, spaces on the inner walls, the interstices between the utensils, as indicated in 630, and the interstices between the utensils and the walls, as indicated in 631. See Figs. 4A and 4B.
[0054] Video compression provides a variety of simple and fast mechanisms for reducing data in a kitchen scene to extract
Petition 870190093460, of 09/18/2019, p. 11/29
20/57 data indicating an activity level. The level of activity can be combined with other information, such as contemporary audio content (from a 615 microphone) to allow detection of a human response to fire or panic by the occupants of kitchen 601. In a simple modality, block compression can discard high coefficients of color depth and high discrete cosine transform (DCT) and perform motion estimation on decimated blocks to provide an estimate of the total motion energy in the scene as an indicator of activity level, randomness or directionality, features that when applied to a classifier they can discriminate scenes that contain panic or purposeful behavior that meets an emergency from other normal activities. In this way, the controller can infer the presence of a possible fire from the reactions of the kitchen staff. As indicated, this information in combination with other information, such as infrared sensors facing a cooking surface, can provide highly reliable estimates of the presence of a fire. Again, this is just an example. Here again, selecting the dimensions of the input feature space is important to avoid problems such as over or under adjustment.
[0055] An IR 614 sensor can also be positioned in the kitchen to monitor the region adjacent to an appliance, such as the 620 appliance. The additional or alternative 614 infrared sensors can be positioned to monitor various discrete locations around a kitchen, such as various adjacent utensils, doors, occupied regions, components of mechanical systems, such as fan motors, ventilation registers, spaces above the suspended ceiling, spaces on the interior wall, the interstices between the utensils, as indicated in 630, and the interstices between utensils and walls, as indicated in 631. An IR sensor can also be positioned to detect the radiant temperature in a duct by a sensor
Petition 870190093460, of 09/18/2019, p. 11/30
21/57 of IV 623 positioned in a duct or to detect infrared light emanating from a source inside the duct, as indicated by the 623 IR sensor. This sensor can also be replaced by an infrared camera. The 623 infrared sensor or the infrared camera can detect burning embers, for example.
[0056] 232 radiation opacity or concealment sensors, such as those used in smoke alarms, can also provide relevant input for a signal combination controller, as discussed. For example, an opacity or radiation obscuration sensor 649 can be positioned outside a hood to detect a significant smoke leak. The 649 concealment sensor receives light from the environment or from a specific source within or near it to establish a baseline level of light intensity received. When smoke enters a space between the light source or sources and the light sensor, an emission signal level indicates the change in the opacity of the space between the light and the sensor. A variety of devices can be used for a light sensor, such as photocells, charge-coupled devices, photomultipliers and camera sensors. The controller can be configured to detect a duration of high smoke levels and fire based on duration and opacity or a combination of them. As discussed, this signal can be combined with others to identify a fire, locate it and determine its size in combination with other factors. In some embodiments, the smoke can absorb visible or thermal radiation, so it can be used as an inhibitory signal in combination with one or both of the signals in the controller response.
[0057] Other types of opacity or light blocking sensors, such as those used as optical triggers in security systems, can be used to detect the accumulation of debris. For example, this 633 lock sensor can be located in narrow spaces or areas that
Petition 870190093460, of 09/18/2019, p. 11/311
22/57 are usually hidden from view or difficult to inspect, which can be cleaned infrequently. An example of a space behind the utensil 620 adjacent to a wall 636 is shown. See also Fig. 4B and the interstices 630 between the adjacent utensils and the interstices 631 between a wall 601 and the adjacent utensils 620. Similar interstices where residues or dust that can accumulate are interstices between adjacent exhaust hoods or between a wall and an adjacent exhaust hood, as shown in 670 in Fig. 4A. It should be noted that there may be interstices between the exhaust hoods that are not adjacent to a wall, for example, so-called canopy hoods can be installed so that there are interstices between them where debris can accumulate.
[0058] With reference now to Fig. 3A, temperature sensors of all types, such as duct gas temperature 202, hood temperature 204 and duct wall temperature 206 can be filtered through a low-pass filter 302 before sampled for conversion to a digital signal for further processing. The low-pass filter 302 can be an analog filter for smoothing distortion. The digital processing of temperature signals can include the generation of a continuously updated PSD function 306 based on a middle core such as a Gaussian, so that the PSD function represents a short-term statistical representative of the contemporary distribution of variation in the temperature signal . A simpler type of 308 power band processing can be used to detect fluctuations, such as by a digital filter that indicates when the temperature changes faster than a predefined rate or contains frequency energy in a specific band or energy in several bands . These can be applied as independent DOF signals to the controller. Temperature fluctuations, as will be indicated by this digital processing, can distinguish cooking from fire. For example, when heat is generated in a turbulent flow,
Petition 870190093460, of 09/18/2019, p. 11/31
23/57 like a fire versus steam-driven plug flows, resulting from turning food on a grid, the frequency characteristics in terms of magnitude and wavelength will be characteristically different. Higher temperatures generate higher speeds (and higher Grashoff numbers) and therefore higher frequency energy in thermally induced turbulence. Therefore, the informative content of the unstable spectral components of temperature can cooperate with other parameters to distinguish normal states, such as cooking and preheating of kitchen utensils, from abnormal fires, where uncontrolled combustion occurs.
[0059] Radiant energy, such as radiant temperature signals 208 and visible light 228, shown in Fig. 3B, can be processed in the same way as the temperature signals described with reference to Fig. 3A. Here again, the spectral information of the time-varying signal, as well as the running steady-state signal, are relevant independent DOF signals that can help to discriminate normal processes and events, in a cooking environment, from fires. Again, this information can be combined with information from other sensors to influence a final determination of the presence, location, type and intensity of a fire. Here again, the frequency distribution of fluctuations in radiant energy can carry information about energy in a thermally driven or uncontrolled combustion flow. The smoke indicated by visible light or radiation masking sensors can produce a time-varying mask that causes infrared light to be absorbed in a variable way by smoke (a floating mask effect in which changes in masking are caused by same turbulent effects and therefore can also indicate turbulent energy). Changes and fluctuations in the radiant temperature of a cooking surface can indicate the turning or placing of food on a grill or the movement of hands or kitchen utensils or cooking containers.
Petition 870190093460, of 09/18/2019, p. 11/33
24/57 cooking on the cooking surface. Such fluctuations would be accompanied by sporadic drops in radiant temperature and low luminance in steady state, unless accompanied by a fire, so that the combination of these three signals is useful for the detection and discrimination of fire from more normal events in a context of cooking.
[0060] The information in the image, as discussed above, can be processed automatically to provide inputs to the controller that contribute to distinguish the fire from other events. The thermal image can be taken using a coarse thermal imaging device 210, as discussed with reference to Fig. 11. As indicated, this can provide location information that covers a large area with a single sensor instead of multiple radiating sensors. An infrared camera 212 can also cover a larger area and provide more relevant information, as also discussed above. A duct wall imaging device 214 that forms thermal images of the interior of a duct wall can be provided at one or more locations in a duct, as discussed with reference to the embodiment of Fig. 7.
[0061] Other transient and stable forms defined by the variation of the radiant temperature with decimation of the level of color or luminance (for example, a posterization filter, a limit filter and / or a boundary filter) can be recognized and classified to distinguish as characteristics of normal and fire conditions. This image processing can be done digitally by several known methods, implementing an image processing component 306 for the conditioning of the image signals from the various imaging devices discussed above (for example, coarse thermal imaging device 210, imaging camera infrared 212, duct wall image forming device 214, etc.).
Petition 870190093460, of 09/18/2019, p. 11/34
25/57 [0062] In some modalities, sensors such as those used in smoke alarms 232 can be filtered to produce spectral information, as discussed above. The temporal variation of the signal is a direct reflection of the turbulent energy in the air that passes through the sensor, with the smoke acting as a tracer. Here again, the range of higher frequency components in the signal based on the time produced by combustion or thermally driven flow at high temperature can be used to identify a fire and its location. Although Fig. 4A shows a single 649 radiation opacity or concealment sensor, multiple sensors can be used around a kitchen to help provide location information. Thus, the controller can compare spectral information at the various locations to determine by interpolation where the peak fire intensity is located inside the kitchen.
[0063] With reference now to Fig. 3D, the audio signature of a fire can be generated by a microphone 216 with proper processing and analysis of the primary audio signal, for example, by low-pass or band-pass filtering 302 and conversion to digital processing 310 by analog to digital (A / D) conversion 304. The audio signal can recognize the pattern of the sound of a fire or recognize the words spoken by the occupants to develop an indication of a fire and its characteristics. Fires generate sounds that can be detected and this information can be used to indicate the size, location and type of the fire. Microphones 216 can be placed around occupied zones, as well as unoccupied zones to detect sounds in those zones. Even the non-verbal sound that cannot be converted into words can be an indicator of emergency situations, such as a fire. Screams or screams can be distinguished by sophisticated or simple pattern matching 310. The sounds of people involved in cooperative or emotionally charged panic behaviors have features
Petition 870190093460, of 09/18/2019, p. 11/35
26/57 machine-recognizable that can be reduced and supplied to a local controller to facilitate appropriate action. See, for example, US patent publication US20080309761 to Kienzle, et al. Speech mechanisms for text can be used and they do not need to be processed locally, for example, speech can be transmitted through network channels, such as the Internet, to be processed into symbolic information, indicating the content or meaning of the speech that can be returned to local controllers for response. In addition, the audio recognition of a fire can provide useful information for firefighters, as described in US Patent No. 7,953,228 to Faltesek, et al.
[0064] With reference to Figs. 3E and 3F, modal inputs from utensils 218 and / or user manual inputs 220 can be provided to a controller for control actions. In one embodiment, a requirement for the presence of an operator may be imposed by the system, where appropriate. In this modality, a type of fuel used in a cooking operation can be solid fuel, which may require continuous monitoring and presence by an operator (for example, a cook or another employee) for safety reasons. This solid fuel cooking mode information can be combined with image formation from, for example, visible light camera 612 or audio signals from microphones 615 that can be processed to determine if this requirement is being met. If continuous presence is not maintained, an emission signal can be generated or some other improvement action can be invoked, such as stifling the fire by reducing combustion air or other actions such as the initial suppression of a fire suppression sequence. in stages. Image processing and recognition to detect personnel is a mature technology. For example, US patent 6,611,206 to Eshelman, et al. describes biotype, recognition of human presence, etc. These technologies have also been used by
Petition 870190093460, of 09/18/2019, p. 36/111
27/57 security systems for many years. For example, Honeywell offers video surveillance technology to monitor the activities of people in busy locations.
[0065] In some modalities, other types of sensor signals can be used and converted 304 into digital for detailed processing, such as simple digital filtering until recognition of complex events (classification). Sensors of volatile organic compounds (VOCs) can be used to detect the presence of slow combustion products, or inefficient uncontrolled fires can be revealed by the presence of aromatic hydrocarbons, such as ethane, methane, etc. For example, slow burning residues can emit these VOCs. As discussed, the signal from a VOC sensor can be combined with other signals to determine the presence or incipient formation of fires at an early stage or even the presence of a hazard that is yet to become a fire. An electronic nose can provide a profile of mixtures of gaseous species that can also be used to recognize fire and fire hazards. For example, flammable vapors can be detected. A 632 voltage sensor in equipment or architectural elements that can change shape due to non-uniform heating or high temperature can act as indirect indicators of the presence of an uncontrolled fire. Voltage sensors can be included in utensils, hoods, lighting accessories, work surfaces, ceiling panels, light reflectors and other elements present in an occupied area.
[0066] Fig. 9A shows a side view of a display device 662 with a display panel 660 that can be directed in selectable directions, rotating on a base 664 to allow display panel 660 to be directed at a sensor or image forming device to evaluate a response control system response to it. The display can emit radiation of a controlled luminance in a range
Petition 870190093460, of 09/18/2019, p. 37/111
28/57 selected colors or wavelengths, including visible and / or infrared. The patterns shown in the front view in Fig. 9C, Fig. 9B, showing simply in the same view without projecting any pattern, can simulate a fire pattern. Here, 664 bubbles can move, change shape and speed, etc., to simulate the image of a predefined type of fire. In one embodiment, the display panel 660 simply displays a video of an actual fire. In the modalities, the display has an array of high power directional infrared sources. The patterns of bubbles 664 (zones distinguished by a distinct change in luminance or color) and their movements and changes in size and shape can be statistically distinguished by the characteristic scales of time (~ t) and length (~ 1) commonly used in statistical description stochastic phenomena, such as turbulent or burning flows. Bubbles can also appear and disappear. In a use case, the display panel is tilted so that it points towards an imaging device. The size of the 660 display panel can be selected to cover the FOV of the imaging device or greater. Other features can be displayed, such as simulated burning ember, simulated slow burning debris, etc. The 662 display device can be used to test and calibrate the system under conditions installed in the real world. The display device 662 can be configured during normal kitchen operation to test the system also under real working conditions.
[0067] Fig. 4A also shows a fat sensor 638 which is described in US Patent 8487776 to Livchak, et al. The patent describes a variety of different devices that detect the rate of accumulation (or predict it from the accumulation of a sample) on a duct wall to determine whether a duct is at risk of fire. In modalities, a fouling detector estimates a fouling magnitude on a duct surface by measuring fouling on a substitute surface (without
Petition 870190093460, of 09/18/2019, p. 38/111
29/57 duct) of the sensor. This is achieved by measuring the amount of light returned to a light sensor from a light source on the substitute surface that is exposed to the contaminating vapor stream. The signal from such a device can be applied to a controller to combine with other data to determine whether a fire is present, imminent or at risk of occurring.
[0068] Fig. 4A also shows the hood lighting 611. In the modalities of the material described, the hood lighting accessory 611 can be intermittent, its intensity varied or its color changed to indicate a condition such as fire risk, incipient fire or other alarm conditions. The colors emitted by the lighting fixture 611 can each correspond to a condition or level of risk. For example, the color can change to a yellow color or alternate between normal and yellow intermittently to indicate a risk, while it can do the same with red if a more urgent condition exists. A risky condition may be the absence of an attendant at a station covered by a hood, for example, where solid fuel is used. Another risk may be the blocking of a blocking sensor, such as the 633 blocking sensor, caused by the accumulation of debris. Thus, the color acts as a warning indicating that an action must be taken. An urgent condition may be the detection of a fire in a nearby location. These signals can be combined with the text output in the 603 user interface, explaining the meaning and type of warning and the action to be taken.
[0069] Fig. 6 shows a kitchen exhaust system 700 with several hoods 710, 712, 714 and 716 all connected to a common chamber or duct 713 from which the air is sucked in by a 717 fan. fire suppression contained under pressure in a container 718 is sprayed over a fire of one or more respective nozzles 726, fed by a distribution head 720 and drops 726, in a fire location selected by the controller. Any of the detection systems
Petition 870190093460, of 09/18/2019, p. 39/111
30/57 described here can be used to control the suppression response and which nozzle 726 is opened. The air flow from the hoods 710, 712, 714e716 to the chamber or duct 713 is regulated by the dampers 724, each corresponding to one of the ducts 710, 712, 714 and 716. The dampers 724 can be of a type suitable for high temperatures, which are adjustable to balance the exhaust flow between the hoods 710, 712, 714e716. Where hoods 710, 712, 714 and 716 are connected to an exhaust duct or common chamber, each hood 710, 712, 714 and 716 must be balanced against the others, so that each escape at the minimum rate that guarantees capture and containment of contaminants. The ducts that carry aerosol of fat may have specially adapted dampers, capable of dealing with the dangers caused by the precipitation of fat. A suitable configuration for the 724 muffler is described in US Patent Application Publication No. 2015/0300653 by Livchak, et al. This damper is continuously variable, grease tolerant, robust and heat resistant, in addition to being able to completely seal the 722 duct.
[0070] In a fire response, several actions can be taken by coordinating the opening and closing of selected actions of the 724 dampers. In a first response to a fire detection, all dampers are opened and the 717 fan flow is increased the maximum. This response focuses on the removal of vapors. The last response may respond, by the controller, to a certain type of fire, for example, one distinguished more by the generation of dangerous or poisonous smoke than by the risk of sudden ignition. In another answer, a different type of fire can be better controlled by pulling the vapors from the immediate location of the fire and directing all the suction to that location, closing all but a 724 damper. These two responses highlight a value of highly discriminatory detection systems , using sophisticated classification techniques based on multiple sensor inputs with filtering
Petition 870190093460, of 09/18/2019, p. 40/111
31/57 appropriate, as described in the various examples given in this document. The activation of the dampers can be controlled in response to a location and type of fire. A fire that generates dangerous smoke can be distinguished by low heat and widely distributed opacity and other data such as audio with word recognition or choking recognition, motion compensation table data indicating panic, VOC detection, confined high temperature field or low temperature and / or other parameters. A fire that presents a risk of spread can be distinguished by the same characteristics with different values and combinations of values. In a type of fire prone to spread, the controller can open the muffler closest to the fire and close those away from the fire, so that the suction of other mufflers does not help the fire to spread.
[0071] With reference to Fig. 3G, as described above, a variety of digitized and conditioned reduced signals can be applied as inputs to a signal combiner implemented in a software program on a 105 processor. The classifier can emit symbols indicating several recognized classes and confidence levels for each of the classes that can be further reduced to a signal that commands one or more actions by one or more emission effectors 107. The classifier and the upstream and downstream processing can be performed by a controller 600 (Fig. 4A) or at least in part by a controller located remotely implemented on a network or on a server connected to the Internet.
[0072] Fires can be classified by exploiting information stored on data storage devices located locally or remotely. Classifiers are known to use supervised learning to generate associated classes and confidence levels. These can be used to detect fires. The sensors
Petition 870190093460, of 09/18/2019, p. 41/111
32/57 actively monitor different indicators. The controller stores data samples from the different sensors. The system determines the probability of a single emission of the sensor or a combination of them, or a reduced indicator of them, as a running statistic, alone or in combination with other data, indicating a fire. The system can use information stored in various locations and developed over a period of time to improve its responsiveness and its ability to eliminate false positive detections. The confidence levels can be multiple, each associated with a single sensor or subcombination with the controller calculating a combined confidence level. Alternatively, a single level of confidence can be developed from the joint probability of all combined inputs, such as those provided by a network classifier. Other types of combiners / reducers / classifiers 106 can be used such as fuzzy logic, neural network evaluation, rule-based systems, Bayesian classifiers, model-based classifiers, unsupervised learning algorithms, etc. Different types and levels of alarm can be generated in response to a fire or hazardous conditions in response to one or more confidence levels of the system.
[0073] The sampled values of the various entries, the reduced data derived from them and various other data, as received from utensils, entered by users (modal data) and algorithms and models are stored in the data and / or memory stores to provide access to them by the classifier. The stored data can also include threshold reference values, image processing filters and other data.
[0074] Confidence levels are inherent to probabilistic classifiers, such as Bayesian classifiers. Rule-based classifiers can include high levels of confidence for each issue. Rules-based models can also calculate confidence levels using
Petition 870190093460, of 09/18/2019, p. 42/111
33/57 formulas, for example, confidence can be taken as a product logarithm discounting the confidence levels of each of the multiple independent inputs (independent DOFs) or their point vector product.
[0075] Figs. 12A and 12B show simplified front and side views of an exhaust hood 320 with auxiliary components described below. Referring now to Figs. 12A and 12B, a system for exhaust flow control and fire detection has at least one (in the illustrated mode, two) canopy temperature sensor (s) 400. Two canopy temperature sensors 400 CT1 and CT2 are shown, but one or a larger number of canopy temperature sensors 400 can be provided. A gas temperature sensor in duct 401 indicates the temperature of the vapors sucked through a fixed duct. The gas temperature sensor of duct 401 can be mounted on a rod to thermally insulate it from duct wall 320. As are commonly used in commercial kitchen ventilation systems, 321 grease filters, such as deflector filters, can be positioned in an exhaust flow path. The vapors are sucked in and through the duct 320 by an exhaust fan. Infrared temperature sensors 407 can also be provided. In the described modality, two infrared temperature sensors 407 identified as IV1 and IV2 are provided, but a smaller or greater number can be used in other modalities. The infrared temperature sensors 407 can be directed to the respective contiguous or overlapping regions of an appliance cooking area (such as a stove or oven - not shown) under the fume hood 320. The fume hood 320 can be of a rear shelf or canopy type configuration. The sensors generate emission signals that are applied to a programmable controller 108 that can include communications 111 and user interface 112. The sensors can include signal filters.
Petition 870190093460, of 09/18/2019, p. 43/111
34/57
A / D converters can be included in sensors or in controller 108. Temperature sensors can be resistance thermometers (RTDs), thermistors, thermocouples, quartz oscillators or any other type of temperature sensor.
[0076] To minimize the exhaust flow required to obtain complete capture and containment, it is known that a controller detects a demand (steam load) and controls the exhaust flow in proportion to the steam load. This is often called control based on exhaust flow demand. This can be achieved by modulating the position of a damper or by controlling the speed of an exhaust fan. The infrared, duct temperature and canopy temperature sensors described above, or a subset of them, can be used as inputs to regulate the exhaust flow. For example, US Patent 9494324 to Livchak et al describes a control system that detects that the appliance is turned off when infrared sensors and duct temperature sensors indicate a level limit. The controller signals an idle condition of the utensil (for example, a grill is heated by burners, but no food is cooking) when the infrared temperature sensors 407 indicate a threshold temperature and controller 108 progressively regulates the volume rate over a range flow rates in proportion to the temperature indicated by the duct housing temperature sensor 401. If the 407 infrared temperature sensors indicate a fluctuating rise or fall in temperature that is beyond a limit, that is, the absolute value of the rate of change exceeds a predefined stored limit, controller 108 increases the flow to a design flow. The design flow is a maximum rate designated for the specific hood. This is an example of an exhaust flow control scheme that uses infrared and duct gas temperature sensors. In this specific modality, the infrared temperature sensors 407 and gas duct 401 are combined with
Petition 870190093460, of 09/18/2019, p. 44/111
35/57 canopy temperature sensors 400 to provide additional fire detection functionality.
[0077] With reference now to Figs. 12D and 12F, a linear combiner type fire detector is shown that combines signals from the canopy temperature sensors 400 (here identified as 501C and 501D), the exhaust duct temperature sensor 401 (here identified as 501E) and the infrared temperature sensors 407 (here identified as 501A and 501B) to generate a fire detection signal. This can trigger a response such as an alarm or a fire suppression response or any of the various responses identified above. The emission levels of sensors 501A to 501E can be added (after conversion and A / D normalization), numerically by processor 108 to generate a composite emission signal. The sum operation here is indicated at 505. The composite send signal compared by controller 108 to a predefined limit stored internally to generate a limit signal as indicated symbolically at 506. The limit send signal can be applied to a unit 507 to generate an emission or alarm signal 508. The limit signal can also be applied to other effectors, such as a fire suppression system, to cause the suppressor to be applied to a region that is monitored locally by the 501A sensors a 501E. Fig. 12F shows a figurative graph of a typical signal. The canopy temperature remains regular and low through normal heating, cooking and accidental flames that occur in the kitchen, such as cooking hamburgers on a grill. This moderate behavior is the result of an adequately regulated exhaust flow. The curves shown exaggerate the smoothness and the instantaneous measurements show random fluctuations and small changes in the flames and steam that rise from cooking food. The sum signal from all sensors increases when a fire occurs and when the sum signal reaches a limit of the
Petition 870190093460, of 09/18/2019, p. 45/111
36/57 limit 506, the emission indicating a fire is generated. It should be noted that, although the process for adding signals is linear, the use of a 506 limit is such that the final emission is non-linear.
[0078] Fig. 12E shows a combiner in which each emission of the sensor 501A to 501E is limited by the limits 504A to 504E and the limited emissions added by an adder 505 to generate an emission that is limited 506. This is a type of system of voting in which if a sufficient number of sensors (determined by the 506 limiter limit) votes that there is a fire, the final emission of the 506 limiter indicates a fire.
[0079] An illustration of the effect of the composition is shown in Fig. 12F. The present is an illustration of observations of various real data from a test environment to summarize what happens during idleness, cooking, non-fire cooking flames and fire explosion in a real commercial kitchen environment. At the beginning of a cooking operation, a gas grill is connected. The grill gradually heats up and the exhaust flow control system regulates the flow to increase the flow as the steam load increases. At 464, the duct temperature signal limit emits a duct limit temperature which is a short peak due to imperfect feedback control. This is not a harmful condition and the occasional appearance of a signal that exceeds a limit does not generate a composite emission of the linear combiner of Fig. 12D. That is, the added emission does not exceed the limit, as it would when several sensors are generating a combination of emission signal.
[0080] Referring now to Fig. 12G, data are shown for an additional control scheme implemented by a controller mode 108. A fire detection signal is generated from the signals from the canopy temperature sensor 400 (Canopy_Temp_l) , exhaust duct temperature sensor 401 (Duct_Temp) and infrared temperature sensors 407 (IRl_Temp, IR2_Temp) to generate a
Petition 870190093460, of 09/18/2019, p. 46/111
37/57 fire detection signal. In this mode, an independent fire detection signal is generated by each sensor. The independent fire detection signal is indicated as 0 for false and 100 for true. A curve for each is indicated as IR1FD, IR2FD, CT1FD. The duct temperature is not used to generate a fire detection signal. The value of the respective infrared fire detection signals is 0, unless the DC signal of the respective infrared sensor indicates a radiant temperature of at least 250 F (121 ° C) AND a radiant temperature rise rate of at least 5 , 4 F / s (degrees per second) where the emission changes to 100% or True and the DC signal from the canopy temperature signal exceeds 150 F (65.5 ° C). In an additional modality, the absolute value of the rate of increase or decrease in the radiant temperature is compared to this limit and an emission of True (100%) is generated. The fire detection signals are combined by the controller, generating a composite fire detection signal of True if all individual fire detection signals are true at any given time. In a variant, the two infrared fire detection signals are OR-ed before being composed in this way. Thus, the composite fire detection signal is True if (IR1FD OR IR2FD) AND CT1FD is True. In addition, in the modalities, a delay operator can be applied to the instantaneous signals of each of the infrared fire detection signals. The magnitude of the delay, the limit for the rate of change of the radiant temperature, the limit radiant temperature and the canopy temperature can be provided as an adjustable value for an installer and accessed and changed through the user interface. The graphs of the fire detection signals indicating a fire (100%) appear as black curves that show peaks, as indicated in 471. The true / false indications of the individual fire detection signals may be delayed so that , when they become
Petition 870190093460, of 09/18/2019, p. 47/111
38/57 true, they remain true for a minimum period of time, so that the chance of overlapping individual fire detection signals is increased. The time delay can be an adjustable parameter through the 112 user interface. The final emission of the fire detection can have a block that, once activated, remains activated until the rest due to some event, such as a user input or completion predefined suppression regime.
[0081] According to the first modalities, the described material includes a system to detect a fire. A plurality of sensors are connected to a controller. The controller implements one or more signal filters to process signals from the plurality of sensors and apply a result to a classifier implemented in the controller. The classifier emits a fire detection signal and a confidence level and applies the emission to a response system.
[0082] Any of the first modalities can be modified to form modalities in which the response system includes a fire suppression system. Any of the first modalities can be modified to form modalities in which the response system includes a fire suppression that uses a chemical suppressor to extinguish a fire. Any of the first modalities can be modified to form modalities in which the response system includes a fire suppression that uses a gas suppressor to extinguish a fire. Any of the first modalities can be modified to form modalities in which the response system includes a fire suppression system that uses a liquid suppressor to extinguish a fire.
[0083] Any of the first modalities can be modified to form modalities in which the classifier emits data that distinguishes a fire. Any of the first modalities can be modified to form modalities in which the fire is distinguished from
Petition 870190093460, of 09/18/2019, p. 48/111
39/57 according to a type of fuel. Any of the first modalities can be modified to form modalities in which the fire is distinguished according to the size of the fire. Any of the first modalities can be modified to form modalities in which the fire is distinguished according to the amount of smoke. Any of the first modalities can be modified to form modalities in which the fire is distinguished according to a temperature. Any of the first modalities can be modified to form modalities in which the response system receives the data that distinguishes a fire and selects one or at least two response modes in response to the data that distinguishes a fire. Any of the first modalities can be modified to form modalities in which the response modes differ in terms of a type of suppressor.
[0084] Any of the first modalities can be modified to form modalities in which the response modes differ in terms of the amount of suppressor. Any of the first modalities can be modified to form modalities in which the response modes differ in terms of a type of suppressor dispensing rate. Any of the first modalities can be modified to form modalities in which the response modes differ in terms of a type of waiting interval before dispensing a suppressor. Any of the first modalities can be modified to form modalities in which the response modes differ in terms of whether the response can be interrupted by activating a staff from an override control input. Any of the first modalities can be modified to form modalities in which the response modes differ according to the data received from a user interface indicating a mode of a cooking operation. Any of the first modalities can be modified to form modalities in which the response modes differ
Petition 870190093460, of 09/18/2019, p. 11/11
40/57 according to data received from a user interface indicating a type of fuel used for a cooking operation. Any of the first modalities can be modified to form modalities in which the plurality of sensors includes a temperature sensor and a luminance sensor. Any of the first modalities can be modified to form modalities in which the plurality of sensors includes a gas temperature sensor and a radiant temperature sensor.
[0085] Any of the first modalities can be modified to form modalities in which the plurality of sensors includes an image forming device positioned in a duct and the response system includes a water spray. Any of the first modalities can be modified to form modalities in which the image forming device applies an image to one or more signal processors that are adapted to detect a gas ember. Any of the first modalities can be modified to form modalities in which the imaging device includes a visible and / or infrared light camera.
[0086] Any of the first modalities can be modified to form modalities in which the image forming device applies an image to one or more signal processors that are adapted to detect a duct fire. Any of the first modalities can be modified to form modalities in which the sensors include a fouling detector configured to estimate a fouling magnitude on a duct surface by measuring fouling on a substitute surface (without duct). Any of the first modalities can be modified to form modalities in which the response system includes a light bulb. Any of the first modalities can be modified to form modalities in which the response system includes a light bulb.
Petition 870190093460, of 09/18/2019, p. 50/111
41/57 [0087] Any of the first modalities can be modified to form modalities in which the light bulb is positioned to illuminate a cooking surface and is located inside the recess of an exhaust hood. Any of the first modalities can be modified to form modalities in which the light bulb generates multiple colors in response to the fire detection signal. Any of the first modes can be modified to form modes in which the light bulb generates multiple colors in response to the fire detection signal and the level of confidence.
[0088] According to the second modalities, the subject described includes a system for detecting a fire. A plurality of sensors are connected to a controller. The controller implements one or more signal filters to process signals from the plurality of sensors and apply a result to a classifier implemented in the controller. The classifier emits a fire detection signal and applies the emission to a response system.
[0089] Any of the second modalities can be modified to form modalities in which the response system includes a fire suppression system. Any of the second modalities can be modified to form modalities in which the response system includes a fire suppression that uses a chemical suppressor to extinguish a fire. Any of the second modalities can be modified to form modalities in which the response system includes a fire suppression that uses a gas suppressor to extinguish a fire. Any of the second modalities can be modified to form modalities in which the response system includes a fire suppression system that uses a liquid suppressor to extinguish a fire.
[0090] Any of the second modalities can be modified to form modalities in which the classifier issues data that
Petition 870190093460, of 09/18/2019, p. 51/111
42/57 distinguish a fire. Any of the second modalities can be modified to form modalities in which the fire is distinguished according to a type of fuel. Any of the second modalities can be modified to form modalities in which the fire is distinguished according to the size of the fire. Any of the second modalities can be modified to form modalities in which the fire is distinguished according to the amount of smoke. Any of the second modalities can be modified to form modalities in which the fire is distinguished according to a temperature. Any of the second modalities can be modified to form modalities in which the response system receives the data that distinguishes a fire and selects one or at least two response modes in response to the data that distinguishes a fire. Any of the second modalities can be modified to form modalities in which the response modes differ in terms of a type of suppressor.
[0091] Any of the second modalities can be modified to form modalities in which the response modes differ in terms of the amount of suppressor. Any of the second modalities can be modified to form modalities in which the response modes differ in terms of a type of suppressor dispensing rate. Any of the second modalities can be modified to form modalities in which the response modes differ in terms of the type of waiting interval before dispensing a suppressor. Any of the second modalities can be modified to form modalities in which the response modes differ in terms of whether the response can be interrupted by activating a personnel from an override control input. Any of the second modalities can be modified to form modalities in which the response modes differ according to the data received from a user interface indicating a mode of a response.
Petition 870190093460, of 09/18/2019, p. 11/111
43/57 cooking operation. Any of the second modalities can be modified to form modalities in which the response modes differ according to the data received from a user interface indicating a type of fuel used for a cooking operation. Any of the second modalities can be modified to form modalities in which the plurality of sensors includes a temperature sensor and a luminance sensor. Any of the second modalities can be modified to form modalities in which the plurality of sensors includes a gas temperature sensor and a radiant temperature sensor.
[0092] Any of the second modalities can be modified to form modalities in which the plurality of sensors includes an image forming device positioned in a duct and the response system includes a water spray. Any of the second modalities can be modified to form modalities in which the image forming device applies an image to one or more signal processors that are adapted to detect a gas ember. Any of the second modalities can be modified to form modalities in which the imaging device includes a visible and / or infrared light camera.
[0093] Any of the second modalities can be modified to form modalities in which the image forming device applies an image to one of the more signal processors that are adapted to detect a duct fire. Any of the second modalities can be modified to form modalities in which the sensors include a fouling detector configured to estimate a fouling magnitude on a duct surface by measuring fouling on a substitute (ductless) surface. Any of the second modalities can be modified to form modalities in which the response system includes a light bulb.
Petition 870190093460, of 09/18/2019, p. 53/111
44/57 [0094] Any of the second modalities can be modified to form modalities in which the light bulb is positioned to illuminate a cooking surface and is located inside the recess of an exhaust hood. Any of the second modalities can be modified to form modalities in which the light bulb generates multiple colors in response to the fire detection signal. Any of the second modalities can be modified to form modalities in which the light bulb generates multiple colors in response to the fire detection signal and the level of confidence.
[0095] According to the third modalities, the matter described includes a duct protection system. At least one sensor is connected to a controller. The sensor is mounted in a duct. The sensor includes a radiant temperature sensitive element responsive to burning material on a wall or material carried by the gas flowing in the duct. A suppressor dispensing tube has a valve and a suppressor dispensing nozzle, the nozzle being positioned in the duct. The controller controls the suppressor dispensing valve in response to an emission from the sensor.
[0096] Any of the third modalities can be modified to form modalities in which the duct is a section of modular duct that can be adapted in a duct system.
[0097] Any of the third modalities can be modified to form modalities in which the dispensing tube is a water pipe.
[0098] Any of the third modalities can be modified to form modalities in which the sensor includes a thermal imaging device. Any of the third modalities can be modified to form modalities in which the thermal imaging device includes an infrared camera.
[0099] According to the fourth modalities, the material described includes a fire protection system for a commercial kitchen.
Petition 870190093460, of 09/18/2019, p. 54/111
45/57
A light blocking sensor (optionally) includes a retro-reflector and includes a light source and a photosensor. The retro-reflector can resume the light from the light source that can be placed on the photosensor or the photosensor can be opposite the light source. The light blocking sensor is positioned in a space adjacent to a kitchen appliance where debris accumulates and emitting an indication of light blocking caused by the accumulation of debris. [00100] Any of the fourth modalities can be modified to form modalities in which the light blocking sensor is positioned between a wall and a kitchen utensil and responds to a light path between the photosensor and the retro reflector or the light source . Any of the fourth modalities can be modified to form modalities in which the light path extends over the majority of a dimension of the utensil. Any of the fourth modalities can be modified to form modalities in which the utensil is a commercial fryer. Any of the fourth modalities can be modified to form modalities in which the debris includes grease and dust.
[00101] According to the fifth modalities, the subject described includes a fire control system. A controller has at least one sensor connected to the controller. The controller has a display and input element that displays an indication for selecting a type of kitchen appliance and receives a selection from a user that indicates a selection indicating a type of appliance. The controller stores mode data in response to the selection and controls the display element to emit a fire indication in response to data and mode data received from at least one sensor.
[00102] Any of the fifth modalities can be modified to form modalities in which the at least one sensor includes a video camera. Any of the fifth modalities can be modified to form modalities in which the emission of the video camera is applied to
Petition 870190093460, of 09/18/2019, p. 55/111
46/57 a video stream classifier implemented by the controller. Any of the fifth modalities can be modified to form modalities in which the classifier issues presence indication data if a person is present in a kitchen utensil in response to model data and presence indication data. Any of the fifth modes can be modified to form modes in which the controller generates an alarm if a person is not present when a particular mode is indicated by the mode data. Any of the fifth modalities can be modified to form modalities in which the determined mode is associated with the burning of a predefined type of fuel that must be monitored continuously. Any of the fifth modalities can be modified to form modalities in which the predefined type of fuel is a solid fuel.
[00103] According to the sixth modalities, the subject described includes a fire suppression system that emits a step response to a fire in response to an indication of severity or type of fire.
[00104] According to the seventh modalities, the subject described includes a fire suppression system. A controller has a fire detection element that includes smoke and energy emission sensors. The system has one or more controllers for separate exhaust hood dampers, each damper controlling the flow through a respective hood connected to a common exhaust passage. The fire detection element is responsive to the first and second types of fires, the first producing a greater volume of smoke than the second, the second being associated with a greater tendency to spread quickly than the first. The controller operates the dampers in response to a type of fire detected by the fire detection element.
[00105] Any of the seventh modalities can be modified to form modalities in which the first type of fire is indicated by high
Petition 870190093460, of 09/18/2019, p. 56/111
47/57 emission of radiant energy above a predefined magnitude. Any of the seventh modalities can be modified to form modalities in which the first type of fire is indicated by high emission of radiant energy above a predefined magnitude in combination with a smoke level below a predefined magnitude. Any of the seventh modalities can be modified to form modalities in which the first type of fire is indicated by a ratio of radiant energy emission to a smoke level above a predefined magnitude. Any of the seventh modalities can be modified to form modalities in which the controller, in response to the first type of fire, opens all the dampers simultaneously. Any of the seventh modalities can be modified to form modalities in which the controller, in response to the second type of fire, opens a subset of the dampers simultaneously and closes other dampers. Any of the seventh modalities can be modified to form modalities in which the controller, in response to one of the types of fire, maximizes an exhaust flow from the common exhaust passage.
[00106] According to the eighth modalities, the subject described includes a fire suppression system with a narrow cone spray nozzle connected to a fire suppressor source. A wide-angle imaging device is adapted to emit a signal indicating the location of a fire and to apply a signal indicating the same to a controller. The narrow cone spray nozzle has a steering actuator to allow the controller to aim the narrow cone spray nozzle at the fire site.
[00107] Any of the eighth modalities can be modified to form modalities in which the actuator is a panoramic tilting mechanism. Any of the eighth modalities can be modified to form modalities in which the fire suppressor is a fire suppressor.
Petition 870190093460, of 09/18/2019, p. 57/111
48/57 foam. Any of the eighth modalities can be modified to form modalities in which the wide-angle imaging device is an infrared camera. Any of the eighth modalities can be modified to form modalities in which the wide-angle imaging device is a visible light camera.
[00108] According to the ninth modalities, the subject described includes a method for suppressing a fire that includes the application of a fire suppressor to a fire locally, detecting a fire response to the suppressor application. The method includes releasing a building fire extinguisher system from a holding condition, preventing water spraying in response to a detection result.
[00109] Additional variants of the method may include receiving a manual entry through a user interface indicating a cancellation signal. Still other variants can be such that the release is more responsive to the cancellation signal. In still other variants, the fire suppressor is a chemical suppressor. In yet other variants, the fire is in a kitchen.
[00110] In any form, including claims, an electronic nose can be provided in conjunction with a classifier to recognize fire risks and provide early warning of an incipient fire or fire risk. In the modalities, this system can allow the entry of an annulment or other confirmation to cancel the warning signal. The warning signal can be issued via the user interface, for example, as a message or via a specific alarm tone or visible indicator for such purposes.
[00111] In all modalities, a conventional support fire detection and suppression system can be provided in conjunction with any of the other systems.
Petition 870190093460, of 09/18/2019, p. 58/111
49/57 [00112] In any of the previous and claimed modalities, the control system used to detect fires can be used to detect the normal, but variable, state of a polluting source, such as a cooking utensil cooking surface, and adjust the exhaust flow to minimize the waste of air conditioning in the occupied space, ensuring capture and containment. It should be apparent that many of the sensors described here, as well as the technology for classifying a state as a fire, can be used to determine the normal state of a kitchen appliance, for example, to obtain exhaust control.
[00113] It will be recognized that the modules, processes, systems and sections described above can be implemented in hardware, software programmed hardware, software instructions stored in a non-transitory computer-readable medium or a combination of the above.
[00114] The devices and methods that receive signals from the sensor and emission information that discriminate several possible conditions indicated by the values of the sensor signals, such as fire conditions or vapor load conditions, can generally be identified as classifiers or recognition filters of patterns. In either mode, these mechanisms can use sophisticated processor-based algorithms that produce estimates of possible conditions and confidence estimates for each. In such methods, a main process can be provided to classify an input state vector, a vector that is the set of multiple reduced inputs from multiple sensors. With reduced inputs, it means that raw input data, like the many pixels in a video image, are converted into quantitative and symbolic tokens to provide relevant information with less potential for overfitting. So, for example, a raw video image of 3 million pixels can be reduced by image processing and pattern recognition to a count of people currently in a photographed scene. Contact Information
Petition 870190093460, of 09/18/2019, p. 59/111
Additional 50/57 can be provided, such as the average speed of human movement or if it fits a normal movement pattern (for example, by recognizing gait, the pattern recognizer can distinguish between running, which is abnormal, and walking ). Another example might be where a camera's video stream is used to achieve maximum brightness in a scene or contrast limit for a region of the scene that exceeds a brightness limit, which can indicate the vigor of a fire. A similar problem arises with regard to audio data. Thus, a front terminal pattern recognition process can include processing stages instead of just a single sorting process, and processing can include, in addition to A / D conversion, filtering (for example, image filtering and image selection). resources), decomposition of the orthogonal function, motion vector analysis, division of a two-dimensional state vector into zones and other processes. In addition to input reduction, there is also a problem that certain states involve history and a current condition cannot be simply determined by taking a snapshot of the states of all inputs. Instead, a history must be accumulated. The extraction of the motion vector from the video is an example. Gait recognition is another, as is the recognition of audio features. The effect of all this is that classification is a multi-step process.
[00115] Thus, the sensor data that can be applied to a control system based on machine learning generally has so many degrees of freedom that it is difficult to train a robust classifier or pattern recognizer. And the problem of simplifying the resource space (input vector) used to train or create a classifier is permanent, which creates unique challenges in each application. Unique opportunities and challenges include choosing which types of sensors to use, where to locate them, which information content in each type of sensor is more
Petition 870190093460, of 09/18/2019, p. 60/111
51/57 relevant to the recognition challenge and how to reduce the raw data to extract this type of information through processing and, optionally, one or more intermediate stages of filtering and pattern / resource recognition and a final classifier. Pattern classifier / recognition processes and devices can utilize various algorithms and hardware elements, which are known, are currently being developed or will be developed in the future.
[00116] The recognition of objects, such as human beings in images, using pattern recognition approaches is a technology known from computer vision and includes face recognition. The known technology can use 3D scanners (infrared, such as Microsoft Kinect, autonomous vehicles and product inspection systems). Examples are capable of face recognition and pedestrian detection. Many of these approaches are known. Some use machine learning to build detectors or filters through supervised learning of training images. Simpler systems can apply limits to define simplified fields (partitioned by chroma or luma) in a scene that is often called bubbles in the field of video analysis. These filters are scanned over an input image or video stream to identify the best match patterns. An algorithmic optimization of an adjustment (a type of regression, for example) can produce a pattern of better fit, as well as an estimate of a quality of fit (error), so that these pattern matching systems produce a classification (best fit) and a confidence estimate (a measure of the quality of the fit). The quality of the fit can derive from the value of an objective function after optimization, which is the quality of the fit between the best fit pattern and the target. Adjusting a pattern to a scene's moving resources provides additional information from a time sequence of images of a scene,
Petition 870190093460, of 09/18/2019, p. 61/111
52/57 allowing to estimate the movement of recognized objects. Thus, a numerical statistic indicating the speed and direction of movement can be obtained from a video stream. Even very simplified recognition algorithms can be used, for example, if you simply derive discrete bubbles and their speed and directions of movement, this can be a reasonable indicator of the level of activity of personnel in a scene. This can, for example, be used to discriminate between normal movement patterns associated with employees and a panic situation caused by a fire emergency.
[00117] Fig. 14 shows a block diagram of an exemplary computer system according to the modalities of the subject described. In various embodiments, all or part of the 1000 system can be included in a medical treatment device / system, such as a renal replacement therapy system. In these modalities, all or part of the 1000 system can provide the functionality of a controller of medical treatment devices / systems. In some embodiments, all or part of the 1000 system can be implemented as a distributed system, for example, as a cloud-based system.
[00118] System 1000 includes a computer 1002, such as a personal computer or workstation or other computing system that includes a processor 1006. However, alternative modalities may implement more than one processor and / or one or more microprocessors, devices microcontroller or control logic, including integrated circuits, such as ASIC.
[00119] Computer 1002 additionally includes a bus 1004 that provides communication functionality between various modules of computer 1002. For example, bus 1004 can allow the communication of information / data between processor 1006 and a memory 1008 of computer 1002, from so that the 1006 processor can recover
Petition 870190093460, of 09/18/2019, p. 62/111
53/57 data stored in memory 1008 and / or execute instructions stored in memory 1008. In one embodiment, these instructions can be compiled from source code / objects provided according to a programming language such as Java, C ++, C #, .NET, Visual Basic ™ language, LabVIEW or other structured or object-oriented programming language. In one embodiment, the instructions include software modules that, when executed by the 1006 processor, provide renal replacement therapy functionality in accordance with any of the modalities described herein.
[00120] Memory 1008 can include any volatile or non-volatile computer-readable memory that can be read by computer 1002. For example, memory 1008 can include a non-transitory computer-readable medium, such as ROM, PROM, EEPROM, RAM, flash memory, disk drive, etc. Memory 1008 can be removable or non-removable media.
[00121] Bus 1004 can also allow communication between computer 1002 and a display 1018, a keyboard 1020, a mouse 1022 and a speaker 1024, each providing the respective functionality according to various modalities described in this document, for example example, to set up a treatment for a patient and monitor a patient during a treatment.
[00122] Computer 1002 can also implement a 1010 communication interface to communicate with a 1012 network to provide any functionality described here, for example, to alert a healthcare professional and / or receive instructions from a healthcare professional, reporting conditions of patient / device in a distributed system to train a machine learning algorithm, record data in a remote repository, etc. The 1010 communication interface can be any interface known in the art to provide wireless and / or
Petition 870190093460, of 09/18/2019, p. 63/111
54/57 wire, such as a network card or a modem.
[00123] The bus 1004 can also allow communication with a sensor 1014 and / or an actuator 1016, each providing the respective functionality according to several modalities described in this document, for example, to measure signals indicative of a patient condition / device and to control the operation of the device accordingly. For example, sensor 1014 can provide a signal indicating a fluid viscosity in a fluid circuit in a renal replacement therapy device, and actuator 1016 can operate a pump that controls fluid flow in response to signals from the sensor 1014.
[00124] A method for detecting and / or suppressing fire can be implemented, for example, using a processor or system, as described with reference to Fig. 14, configured to execute a sequence of programmed instructions stored in a computer-readable medium non-transitory. For example, the processor may include, but is not limited to, a personal computer or workstation or other computing system that includes a processor, microprocessor, microcontroller device or is composed of control logic including integrated circuits, such as a Integrated Circuit of Specific Application (ASIC). The instructions can be compiled from the source code instructions provided according to a programming language such as Java, C ++, C # .net or similar. The instructions can also include code and data objects provided according to, for example, Visual Basic ™, LabVIEW or another structured or object-oriented programming language. The sequence of programmed instructions and associated data can be stored on a non-transitory computer-readable medium, such as a computer memory or storage device that can be any suitable memory device, such as, but not limited to, read-only memory (ROM),
Petition 870190093460, of 09/18/2019, p. 64/111
55/57 programmable read-only memory (PROM), electrically erasable programmable read-only memory (EEPROM), random access memory (RAM), flash memory, disk drive and the like.
[00125] In addition, modules, processes, systems and sections can be implemented as a single processor or as a distributed processor. In addition, it must be considered that the steps mentioned above can be performed on a single processor or distributed (single and / or multiple cores). In addition, the processes, modules and submodules described in the various figures to and from the above modalities can be distributed across multiple computers or systems or can be located on a single processor or system. Exemplary structural modalities suitable for implementing the modules, sections, systems, means or processes described here are provided below.
[00126] The modules, processors or systems described above can be implemented as a general purpose programmed computer, an electronic device programmed with microcode, a wired analog logic circuit, software stored in a computer-readable medium or signal, a optical computing, a networked system of electronic and / or optical devices, a special-purpose computing device, an integrated circuit device, a semiconductor chip and a software module or object stored in a computer-readable medium or signal, for example example.
[00127] Method and system modalities (or their subcomponents or modules) can be implemented in a general purpose computer, a special use computer, a programmed microprocessor or microcontroller and peripheral integrated circuit element, an ASIC or other integrated circuit, a digital signal processor, a wired electronic or logic circuit, as a discrete element circuit, a programmed logic circuit, as a programmable logic device (PLD), logic matrix
Petition 870190093460, of 09/18/2019, p. 65/111
56/57 programmable (PLA), field programmable port matrix (FPGA), programmable matrix logic device (PAL) or similar. In general, any process capable of implementing the functions or steps described here can be used to implement modalities of the computer program method, system or product (software program stored in a non-transitory computer-readable medium).
[00128] In addition, modalities of the computer program method, system and product described can be readily implemented, in whole or in part, in software using, for example, object-oriented or object-oriented software development environments that provide portable source code which can be used on a variety of computer platforms. Alternatively, modalities of the computer program method, system and product described can be implemented partially or completely in hardware using, for example, standard logic circuits or a very large-scale integration project (VLSI). Other hardware or software can be used to implement modalities, depending on the speed and / or efficiency requirements of the systems, the specific function and / or the software or hardware system, microprocessor or microcomputer used. Computer program method, system and product modalities may be implemented in hardware and / or software using any systems or structures, devices and / or software known or more developed by those skilled in the applicable technique from the description of the function provided here and with a general basic knowledge of control systems, signal processing, machine intelligence and / or computer programming techniques.
[00129] In addition, modalities of the computer program method, system and product described can be implemented in software run on a programmed general purpose computer, a special use computer, a microprocessor or similar.
Petition 870190093460, of 09/18/2019, p. 66/111
57/57 [00130] It is therefore apparent that, in accordance with the present description, methods and devices for fire detection and suppression systems are provided. Many alternatives, modifications and variations are permitted by the present description. The characteristics of the described modalities can be combined, rearranged, omitted, etc., within the scope of the invention to produce additional modalities. In addition, some features can sometimes be used to advantage without the corresponding use of other features. Consequently, Applicants intend to adopt all of these alternatives, modifications, equivalents and variations that are within the spirit and scope of the present invention.

权利要求:
Claims (34)
[1]
1. Fire detection system, characterized by the fact that it comprises:
a plurality of sensors;
a response system; and a controller connected to the plurality of sensors, the controller implementing one or more signal filters to process signals from the plurality of sensors and apply a result to a classifier implemented in the controller, the classifier emitting a fire detection signal and a confidence level and applying the emission to the response system, where the plurality of sensors includes an image forming device and a duct, and the response system includes a water spray, and the image forming device applies an image to said one or more signal filters that are adapted to detect a gas ember.
[2]
2. System according to claim 1, characterized by the fact that the response system includes a fire suppression system.
[3]
3. System according to claim 1, characterized in that the response system includes a fire suppression that uses a suppressor to extinguish a fire, wherein the suppressor includes at least one of a chemical suppressor, a gas suppressor or a liquid suppressor.
[4]
4. System according to claim 1, characterized by the fact that the classifier emits data that characterizes a fire.
[5]
5. System according to claim 4, characterized by the fact that the fire is distinguished in accordance with at least one of a type of fuel, a size of the fire, an amount of smoke
Petition 870190093460, of 09/18/2019, p. 104/111
2/7 or a temperature.
[6]
6. System according to claim 5, characterized by the fact that the response system receives said data that distinguishes the fire and selects one or at least two modes of response in response to said data that distinguishes the fire.
[7]
7. System according to claim 6, characterized in that the response modes differ in terms of at least one of a suppressor type, a suppressor quantity, a type of suppressor dispensing rate, a type of interval waiting before the suppressor is dispensed or if the response can be interrupted by personnel activation of an override control input.
[8]
8. System according to claim 6, characterized in that the response modes differ according to the data received from a user interface, indicating at least one in a cooking operation mode or a type of fuel used by the cooking operation.
[9]
System according to any one of claims 1 to 3, characterized in that the plurality of sensors includes a temperature sensor and a luminance sensor.
[10]
System according to any one of claims 1 to 3, characterized in that the plurality of sensors includes a gas temperature sensor and a radiant temperature sensor or an electronic nose.
[11]
11. System according to claim 1, characterized by the fact that the image-forming device includes a visible and / or infrared light camera.
[12]
12. Fire detection system, characterized by the fact that it comprises:
a plurality of sensors;
Petition 870190093460, of 09/18/2019, p. 105/111
3/7 a response system; and a controller connected to the plurality of sensors, the controller implementing one or more signal filters to process signals from the plurality of sensors and applying a result to a classifier implemented in the controller, the classifier emitting a fire detection signal and a confidence level and applying the emission to the response system, wherein the plurality of sensors includes an image forming device positioned in a duct, and the response system includes a water spray, and the image forming device applies an image to said one or more signal filters that are adapted to detect a duct fire.
[13]
13. Fire detection system, characterized by the fact that it comprises:
a plurality of sensors;
a response system; and a controller connected to the plurality of sensors, the controller implementing one or more signal filters to process signals from the plurality of sensors and applying a result to a classifier implemented in the controller, the classifier emitting a fire detection signal and a confidence level and applying the emission to the response system, where the plurality of sensors includes an image forming device positioned in a duct, and the response system includes a water spray, and the sensors include a fouling detector configured to estimate a magnitude of fouling on a duct surface by measuring fouling on a substitute surface (without duct).
[14]
System according to any one of claims 1, 12 or 13, characterized in that the response system includes a lighting lamp.
[15]
15. System according to claim 14, characterized
Petition 870190093460, of 09/18/2019, p. 106/111
It is due to the fact that the light bulb is positioned to illuminate a cooking surface and is located inside a recess in an exhaust hood.
[16]
16. System according to claim 15, characterized by the fact that the illumination lamp generates multiple colors in response to said fire detection signal.
[17]
17. System according to claim 16, characterized by the fact that the illumination lamp generates multiple colors in response to said fire detection signal and said level of confidence.
[18]
18. Fire detection system, characterized by the fact that it comprises:
a plurality of sensors;
a response system; and a controller connected to the plurality of sensors, the controller implementing one or more signal filters to process signals from the plurality of sensors and applying a result to a classifier implemented in the controller, the classifier emitting a fire detection signal and applying the emission to the response system, wherein the plurality of sensors includes an image forming device positioned in a duct, and the response system includes a water spray, and the image forming device applies an image to said one or more filter filters signal that are adapted to detect a gas ember.
[19]
19. System according to claim 18, characterized in that the response system includes a fire suppression system.
[20]
20. System according to claim 18, characterized by the fact that the response system includes a fire suppression that
Petition 870190093460, of 09/18/2019, p. 107/111
5/7 uses a suppressor to extinguish a fire, where the suppressor is a chemical suppressor, a gas suppressor or a liquid suppressor.
[21]
21. System according to claim 18, characterized by the fact that the classifier emits data that distinguishes a fire.
[22]
22. System according to claim 21, characterized in that the fire is distinguished in accordance with at least one of a fuel type, a size of the fire, an amount of smoke or a temperature.
[23]
23. System according to any one of claims 18 to 22, characterized in that the response system receives said data that distinguishes a fire and selects one or at least two modes of response in response to said data that distinguishes a fire .
[24]
24. System according to claim 23, characterized in that the response modes differ in terms of at least one of a suppressor type, a suppressor quantity, a type of suppressor dispensing rate, or a type of suppressor waiting interval before dispensing the suppressor.
[25]
25. System according to claim 23, characterized by the fact that the response modes differ in terms of whether the response can be interrupted by activating a staff from an override control input.
[26]
26. System according to claim 23, characterized in that the response modes differ according to the data received from a user interface, indicating at least one in a cooking operation mode or a type of fuel used by the cooking operation.
[27]
27. System according to any one of claims 21 to 26, characterized in that the plurality of sensors includes a temperature sensor and a luminance sensor.
Petition 870190093460, of 09/18/2019, p. 108/111
6/7
[28]
28. System according to any one of claims 18 to 20, characterized in that the plurality of sensors includes a gas temperature sensor and a radiant temperature sensor.
[29]
29. System according to claim 18, characterized by the fact that the image-forming device includes a visible and / or infrared light camera.
[30]
30. Fire detection system, characterized by the fact that it comprises:
a plurality of sensors;
a response system; and a controller connected to the plurality of sensors, the controller implementing one or more signal filters to process signals from the plurality of sensors and apply a result to a classifier implemented in the controller, the classifier emitting a fire detection signal and a confidence level and applying the emission to the response system, where the plurality of sensors includes an image forming device positioned in a duct, and the response system includes a water spray, and the image forming device applies an image to said one or more signal filters that are adapted to detect a duct fire.
[31]
31. Fire detection system, characterized by the fact that it comprises:
a plurality of sensors;
a response system; and a controller connected to the plurality of sensors, the controller implementing one or more signal filters to process signals from the plurality of sensors and applying a result to a classifier implemented in the controller, the classifier emitting a detection signal
Petition 870190093460, of 09/18/2019, p. 109/111
7/7 fire and applying the emission to the response system, where the plurality of sensors includes an image forming device positioned in a duct, and the response system includes a water spray, and the sensors include a scale configured to estimate a scale magnitude on a duct surface by measuring a scale on a substitute surface (without duct).
[32]
32. System according to any one of claims 18, 30 or 31, characterized in that the response system includes a lighting lamp.
[33]
33. System according to claim 32, characterized in that the lighting lamp is positioned to illuminate a cooking surface and is located within a recess in an exhaust hood.
[34]
34. System according to claim 32, characterized by the fact that the illumination lamp generates multiple colors in response to said fire detection signal.

类似技术:

---

公开号 | 公开日 | 专利标题

---

BR112019019443A2|2020-04-14|fire control systems, fire suppression, fire detection, duct protection, fire protection for a commercial kitchen, fire detection and control and fire threat warning, and method for suppressing a fire fire

---

US10744356B2|2020-08-18|Fire suppression systems, devices, and methods

---

JP2007280167A|2007-10-25|Air conditioner

---

JP2019179573A|2019-10-17|Fire detection system and fire detection method

---

JP6325287B2|2018-05-16|Fire detection system and fire detection method

---

JP2018136977A|2018-08-30|Fire detection system and fire detection method

---

KR102032549B1|2019-11-08|Integrated intelligent safety management system

---

RU117684U1|2012-06-27|ADAPTIVE FIRE ALARM SYSTEM

---

US11127267B2|2021-09-21|Smart fire detection system

---

Mensch et al.2019|Evaluating sensor algorithms to prevent kitchen cooktop ignition and ignore normal cooking

---

JP6321403B2|2018-05-09|Fire detection system and fire detection method

---

Fennelly2020|Fire alarm systems

---

Keller2013|Multi-spectral System for Autonomous Robotic Location of Fires Indoors

---

Kouchinsky2007|Development of video image detection algorithm for smoke plumes

同族专利:

---

公开号 | 公开日

---

KR102311356B1|2021-10-12|

---

AU2018240189B2|2020-07-16|

---

CN110637330B|2021-12-10|

---

CA3056786A1|2018-09-27|

---

ZA201906201B|2021-05-26|

---

MX2019011156A|2019-10-17|

---

PE20191572A1|2019-10-29|

---

CN110637330A|2019-12-31|

---

CL2019002660A1|2019-12-27|

---

EP3602510A1|2020-02-05|

---

US20200054905A1|2020-02-20|

---

AU2018240189A1|2019-10-10|

---

KR20190139229A|2019-12-17|

---

JP6894002B2|2021-06-23|

---

SG11201908210TA|2019-10-30|

---

WO2018175495A1|2018-09-27|

---

JP2020521193A|2020-07-16|

---

CO2019010356A2|2019-10-09|

引用文献:

---

公开号 | 申请日 | 公开日 | 申请人 | 专利标题

---

---

JP3919928B2|1998-04-01|2007-05-30|文化シヤッター株式会社|Heating apparatus and control method thereof|

---

US6170480B1|1999-01-22|2001-01-09|Melink Corporation|Commercial kitchen exhaust system|

---

KR200188197Y1|1999-11-12|2000-07-15|주식회사화인텍|Modulized auto-fire extinguisher|

---

US6611206B2|2001-03-15|2003-08-26|Koninklijke Philips Electronics N.V.|Automatic system for monitoring independent person requiring occasional assistance|

---

US7953228B2|2003-11-18|2011-05-31|Honeywell International Inc.|Automatic audio systems for fire detection and diagnosis, and crew and person locating during fires|

---

US7221260B2|2003-11-21|2007-05-22|Honeywell International, Inc.|Multi-sensor fire detectors with audio sensors and systems thereof|

---

KR101035518B1|2005-02-23|2011-05-23|주식회사 대우일렉트로닉스|Apparatus for detecting fire in otr|

---

CA2539856C|2005-03-16|2010-11-23|Halton Company Corporation|Fume treatment method and apparatus using ultraviolet light to degrade contaminants|

---

US20060227237A1|2005-03-31|2006-10-12|International Business Machines Corporation|Video surveillance system and method with combined video and audio recognition|

---

US20080036593A1|2006-08-04|2008-02-14|The Government Of The Us, As Represented By The Secretary Of The Navy|Volume sensor: data fusion-based, multi-sensor system for advanced damage control|

---

JP2010530964A|2007-06-13|2010-09-16|オーワイハルトングループリミテッド|Duct oil / fat deposit detection apparatus, system and method|

---

GB2450732B|2007-07-04|2009-09-02|Food Industry Technical Ltd|Air control system and method|

---

CA2640840C|2007-10-09|2016-01-26|Oy Halton Group Ltd.|Damper suitable for liquid aerosol-laden flow streams|

---

US20090128355A1|2007-11-21|2009-05-21|Urbin Mark|Device for visibly marking a water output means and method of use|

---

JP5215694B2|2008-03-10|2013-06-19|シンポ株式会社|Duct closing device and smokeless roaster provided with the same|

---

SG171458A1|2008-12-03|2011-07-28|Halton Group Ltd Oy|Exhaust flow control system and method|

---

PE20150551A1|2012-06-07|2015-05-06|Halton Group Ltd Oy|FIRE SUPPRESSION SYSTEMS, DEVICES AND METHODS|

---

KR20140061563A|2012-11-12|2014-05-22|동아대학교 산학협력단|Fire alarm device|

---

KR101551715B1|2013-12-30|2015-09-10|목원대학교 산학협력단|System and method for prventing fire of bulding|

---

JP2015210826A|2014-04-24|2015-11-24|ラムデオ プラディープ|Smoke multi-gas detector alarm and transmitter device|

---

KR101640152B1|2014-05-28|2016-07-15|이공감|Complex fire detector and fire monitoring system comprising the same|

---

US10512809B2|2015-03-16|2019-12-24|Fire Rover LLC|Fire monitoring and suppression system|

---

US9466195B1|2015-08-06|2016-10-11|State Farm Mutual Automobile Insurance Company|Video flame detection system and method for controlling a range|

---

CN205508040U|2016-03-16|2016-08-24|泽川力合（天津）科技发展有限公司|Kitchen safety coefficient of control kitchen conflagration|

---

CN205904195U|2016-07-13|2017-01-25|姜自强|Exhaust pipe belt cleaning device|US11195010B2|2018-05-23|2021-12-07|Smoked Sp. Z O. O.|Smoke detection system and method|

---

FI20185482A1|2018-05-25|2019-11-26|Safera Oy|Stove guard that makes use of different wavelengths|

---

US10885755B2|2018-09-14|2021-01-05|International Business Machines Corporation|Heat-based pattern recognition and event determination for adaptive surveillance control in a surveillance system|

---

TWI694382B|2019-01-04|2020-05-21|財團法人金屬工業研究發展中心|Smoke detection method with deep vision|

---

US20200217550A1|2019-01-08|2020-07-09|Johnson Controls Technology Company|Hvac infrared detection systems and methods|

---

KR101989044B1|2019-02-21|2019-06-13|한국토지주택공사|Fire detector having a reset function interlocked with a lamp and method for warning a fire using the same|

---

WO2020234827A1|2019-05-22|2020-11-26|Tyco Fire Products Lp|Fire detection system with multiple stage alarms|

---

KR102097294B1|2019-07-19|2020-04-06|지와이네트웍스|Method and apparatus for training neural network model for detecting flame, and flame detecting method using the same model|

---

KR102166150B1|2020-01-22|2020-10-15|서정엔지니어링|Operation control system of sprinkler based on environment factor and setting factor|

---

KR102166152B1|2020-01-22|2020-10-15|서정엔지니어링|Operation control system of sprinkler based on environment factor and setting factor|

---

CN111672043A|2020-04-29|2020-09-18|广东电网有限责任公司东莞供电局|Automatic identification fire extinguisher|

---

CN111632331A|2020-06-01|2020-09-08|厦门艾士迪半导体有限公司|Intelligent kitchen fire extinguishing system and fire extinguishing method|

---

CN111951508A|2020-07-03|2020-11-17|北京中安安博文化科技有限公司|Fire classification method, device, medium and electronic equipment|

---

KR102179900B1|2020-07-07|2020-11-17|정희섭|Fire control system apparatus capable of controling fire for loading space and operating method thereof|

---

CN111784986B|2020-07-13|2021-02-09|和宇健康科技股份有限公司|Intelligent security alarm method based on big data|

---

DE102020210481A1|2020-08-18|2022-02-24|BSH Hausgeräte GmbH|Method for monitoring a cooking process and control device|

---

CN113076797B|2021-02-24|2022-01-18|江苏濠汉信息技术有限公司|Charging station electric vehicle fire alarm method and system based on intelligent video identification|

---

CN113350718A|2021-04-13|2021-09-07|西安石油大学|Remote control ultrasonic wave platform of putting out a fire based on FPGA|

法律状态:
2021-10-19| B350| Update of information on the portal [chapter 15.35 patent gazette]|

优先权:

---

申请号 | 申请日 | 专利标题

---

US201762473747P| true| 2017-03-20|2017-03-20|

---

PCT/US2018/023432|WO2018175495A1|2017-03-20|2018-03-20|Fire safety devices methods and systems|

---