Fire control system for a device

Abstract

A laser cutting or engraving device comprising: a body having an enclosed chamber; a laser cutter for cutting or engraving a work material present within the enclosed chamber; a control unit for controlling operation of the laser cutter; and a fire detection and suppression unit mounted to the body to be located within the enclosed chamber, the fire detection and compression unit comprising one or more image sensors configured to take an image of the laser cutter as the laser cutter performs cutting or engraving of the work material; a fire suppression supply actuable to supply fire suppression material into the enclosed chamber; and a controller for receiving image data from the one or more image sensors and configured to perform a primary analysis of the image data to determine a likelihood of a fire event within the enclosed chamber, and upon the primary analysis determining that a fire event has likely occurred, the controller is configured to conduct a secondary analysis of the image data to confirm that a fire event has occurred and upon the secondary analysis confirming a fire event the controller is configured to actuate the fire suppression supply to suppress a fire associated with the fire event within the enclosed chamber.

Description

[0001] FIRE CONTROL SYSTEM FOR A DEVICE

[0002] RELATED APPLICATION

[0003] The present application claims priority from Australian Provisional Patent Application No. 2023903524 filed on 2 November 2023, the entire contents of which are incorporated herein by reference.

[0004] FIELD OF INVENTION

[0005] The present invention relates generally to control system for managing a fire event in a device, and in particular, to a control system for detecting a fire event and for extinguishing a fire event in a material processing device.

[0006] BACKGROUND OF THE INVENTION

[0007] A laser cutting machine is a device that is used across a wide range of industries to provide precision cutting or engraving of a variety of materials from metals, wood, glass and plastic. Such devices use a laser to focus a high energy beam of light onto a material or substrate to perform a high quality and accurate cut or engraving. The laser can be finitely moved over the surface of the material or substrate to form precision cutting or engraving on a surface of the material or substrate, as desired.

[0008] Laser cutting machines can come in a variety of sizes depending on the application. Due to the use of such a high intensity laser, such devices are typically enclosed to avoid potential hazards to the users from exposure to the laser light, high temperatures that could result in a fire, and toxic air contaminants that may be inhaled.

[0009] In such devices, the high intensity laser beam can produce extremely high temperatures and significant amounts of heat as the substrate material is burned away while cutting. In this regard, it is not uncommon for some materials to catch fire during cutting operations, which can generate fumes and smoke inside the device that can be hazardous and cause the device to catch on fire, creating a risk not only to the users of the device but also to the surrounding environment.

[0010] Due to this, a number of safety features have been employed in laser cutting machines and similar material removal machines that use smoke sensors and other thermal sensors to detect smoke or fumes within the device and initiate a shut-down procedure upon detection. In such devices the detection of smoke or fumes may also initiate delivery of a fire suppressant into the chamber of the machine to suppress any fire present therein. However, as the generation of smoke and vapour is relatively common when laser cutting any material or substrate, there is typically a high likelihood that the fire event detected by the sensors in the machine is not an actual fire or likely to cause an actual fire, thereby generating a “false alarm” upon activating a system shutdown or fire suppressant delivery into the machine. Such “false alarms” can cause significant issues in terms of loss of time and damage to the substrate being worked, which can have a significant financial impact on the user of the device. Optical infra-red light sensors have also been proposed for use in such applications, however such sensors have proven to be extremely unreliable due to the fact that they are affected by all light, including sunlight.

[0011] Thus, there is a need to provide an improved means for detecting a fire event in a device that provides greater analysis of the substrate or workpiece to minimise unnecessary intervention by the safety features of the device whilst maintaining safety.

[0012] The above references to and descriptions of prior proposals or products are not intended to be, and are not to be construed as, statements or admissions of common general knowledge in the art. In particular, the above prior art discussion does not relate to what is commonly or well known by the person skilled in the art, but assists in the understanding of the inventive step of the present invention of which the identification of pertinent prior art proposals is but one part.

[0013] STATEMENT OF INVENTION

[0014] The invention according to one or more aspects is as defined in the independent claims. Some optional and / or preferred features of the invention are defined in the dependent claims.

[0015] Accordingly, in one aspect of the invention there is provided a laser cutting or engraving device comprising: a body having an enclosed chamber; a laser cutter for cutting or engraving a work material present within the enclosed chamber; a control unit for controlling operation of the laser cutter; and a fire detection and suppression unit mounted to the body to be located within the enclosed chamber, the fire detection and compression unit comprising one or more image sensors configured to take an image of the laser cutter as the laser cutter performs cutting or engraving of the work material; a fire suppression supply actuable to supply fire suppression material into the enclosed chamber; and a controller for receiving image data from the one or more image sensors and configured to perform a primary analysis of the image data to determine a likelihood of a fire event within the enclosed chamber, and upon the primary analysis determining that a fire event has likely occurred, the controller is configured to conduct a secondary analysis of the image data to confirm that a fire event has occurred and upon the secondary analysis confirming a fire event the controller is configured to actuate the fire suppression supply to suppress a fire associated with the fire event within the enclosed chamber.

[0016] The controller may comprise a processor unit that performs the primary analysis of the image data by receiving updated image data from the one or more image sensors and performing computer vision analysis of the image data over time to determine that a fire event has likely occurred.

[0017] The controller may further comprise a neural network processor that receives the image data and compares the image data against predetermined image categories to confirm whether the image data is representative of a fire event.

[0018] The predetermined image categories may be images classified as being representative of a “no flame” category, “hot component” category, or “flame” category.

[0019] If the neural network processer determines the image data is representative of a “flame” category, the controller may actuate the fire suppression supply to supply fire suppression material into the enclosed chamber.

[0020] The fire suppression supply may comprise a trigger mechanism that is actuated by the controller to supply fire suppression material into the enclosed chamber.

[0021] The trigger mechanism may be configured to pierce a cannister of fire suppression material stored under pressure within the cannister to release the fire suppression material from the cannister into the enclosed container.

[0022] The one or more image sensors may comprise a visual camera and / or a thermal image camera. The one or more image sensors may comprise a Forward Looking Infrared (FLIR) camera.

[0023] In one embodiment, upon the secondary analysis confirming a fire event the controller may be configured to also perform any one or more of: deactivating the laser cutter and move the laser cutter to a safe location within the work chamber, away from the fire, to prevent damage to the laser cutter; disable any exhaust fans and disconnect any fume extractors to prevent more air from entering the work chamber; issue an audible warning alarm sound; and send a message to the control unit of the device advising of a fire event.

[0024] Accordingly, in another aspect of the invention there is provided a method detection and suppression of a fire in a laser cutting or engraving device comprising: collecting image data of an interface between a laser cutter and a work material being cut by the laser cutter; performing a primary analysis of said collected image data to determine a likely fire event at the interface between the laser cutter and the work material; upon detection of a likely fire event at the interface between the laser cutter and the work material, conducting a secondary analysis of the collected image data of the interface between a laser cutter and a work material using a neural network processor that receives the image data and compares the image data against predetermined image categories to confirm whether the image data is representative of a fire event; and upon determining that the secondary analysis of the collected image data confirms that there is a fire event at the interface between the laser cutter and the work material, releasing a fire suppressant material to the interface between the laser cutter and the work material.

[0025] The primary analysis of said collected image data may comprise performing a vision analysis of the collected image data.

[0026] The vision analysis of the collected image data may comprise reviewing the image data frames and search the image data frame for the presence of predetermined regions of interest in the pixels present in the image data frame.

[0027] The predetermined regions of interest in the pixels present in the image data frame may comprises a region of coloured pixels having an area exceeding a predetermined size. The predetermined regions of interest in the pixels present in the image data frame may be tracked across successive image data frames to determine whether the predetermined regions of interest in the pixels is increasing, reducing, remaining constant or moving.

[0028] If the predetermined regions of interest in the pixels present in successive image data frames is increasing, remaining constant and / or above a predetermined area the primary analysis may determine that a fire event at the interface between the laser cutter and the work material is likely.

[0029] The secondary analysis of the collected image data of the interface between a laser cutter and a work material may comprise the neural network processor employing an algorithm using a convolutional neural network (CNN) that classifies the image data into different event categories, including, “no flame”, “hot component”, and “flame”.

[0030] BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The invention may be better understood from the following non-limiting description of preferred embodiments, in which:

[0032] Fig. 1 is a perspective view of laser cutting or engraving device for use with the system of the present invention;

[0033] Fig. 2 is a plan view of a fire detection and suppression unit in a charged form for use on a laser cutting or engraving device of Fig. 1;

[0034] Fig. 3 is a plan view of a fire detection and suppression unit in a used form for use on a laser cutting or engraving device of Fig. 1;

[0035] Fig. 4 is a schematic block diagram of a control system for controlling the fire control system of the present invention.

[0036] DETAILED DESCRIPTION OF THE DRAWINGS

[0037] Preferred features of the present invention will now be described with particular reference to the accompanying drawings. However, it is to be understood that the features illustrated in and described with reference to the drawings are not to be construed as limiting on the scope of the invention.

[0038] The present invention will be described below in relation to its use in a laser cutting machine, such as a desktop device that can be used in a school or workshop environment. However, it will be appreciated that system of the present invention could be employed in any type of machine where temperatures are generated at an interface with a workpiece that could cause a fire event.

[0039] Referring to Fig. 1, a laser cutting device 10 in accordance with an embodiment of the present invention is depicted. The laser cutting device 10 generally comprises a main body 12 that forms a working chamber 14 therein, within which a substrate or material to be worked is received. The main body 12 has a lid 13 hingedly mounted thereto that encloses the working chamber 14, to form a working chamber 14 that is substantially sealed from the surrounding environment.

[0040] A laser cutter 15 is mounted within the working chamber 14 and is mounted upon a gantry so as to be movable within the chamber 14 over a base 16 thereof. The substrate, or material to be worked, is positioned upon the base 16 such that the laser cutter 15 is able to pass over the substrate to direct the laser beam onto the surface of the substrate to facilitate cutting or engraving of the substrate.

[0041] The device 10 will house an electronic control unit (not shown) that will contain a microprocessor and associated computer components to control the operation of the laser cutter and the associate electronic components housed within the device 10. The control unit will have wireless and wired connection to enable connection with a remote computer unit or server to receive control signals and to transmit control signals in the maimer as required.

[0042] Referring to Fig. 2 and Fig. 3, a fire detection and suppression unit 20 in accordance with one embodiment of the present invention is depicted. The unit 20 is configured to form a panel that is mounted in an inner side wall of the body 12 of the laser cutting device 10 of Fig. 1. The unit 20 has an integral controller (not shown) that will contain a microprocessor and associated computer components to control the operation of the fire detection and suppression unit 20 in the manner as discussed below.

[0043] The unit 20 comprises a plate member 22 that provides a surface to which the components of the unit 20 are mounted. A gas sensor 23 is provided to extend into the chamber 14 of the device 10 to detect the presence of a gas therein. The gas sensor 23 is preferably a chlorine sensor that detects the presence of chlorine gas in the chamber 14. An image sensor 24 is also mounted to the plate member 22 to capture images of the substrate during use. The image sensor 24 is preferably a Forward Looking Infrared (FLIR) camera that constantly takes an image of the interface between the laser cutter 15 and the substrate being cut, during use of the device 10. The image sensor 24 may include a plurality of image capturing sensors including a visual camera as well as a thermal imaging camera.

[0044] A fire suppression supply 25 is also mounted to the plate member 22 as depicted in Fig. 2 by the dashed lines. The fire suppression supply 25 is preferably in the form of a CO2 cannister that is charged with CO2 gas to be released into the chamber 14 upon detection of a fire, in accordance with the present invention. It will be appreciated that the fire suppression supply 25 may be supplied with a variety of different fire suppression gasses or materials for different applications, as will be appreciated by those skilled in the art.

[0045] A trigger mechanism 26 is also mounted in the unit 20 and comprises a linear actuator member 27 pivotally connected to an arm member 28 having a piercing pin 29 attached thereto. The piercing pin 29 is configured to pierce an open end of the fire suppression supply 25 upon being triggered, to suppress any fire present in the device 10. In Fig. 2, the trigger mechanism 26 is depicted in an armed state where the linear actuator member 27 is in an extended position such that the piercing pin 29 is positioned adjacent an opening of the fire suppression supply 25. Fig. 3 depicts the trigger mechanism 26 in the triggered state whereby a fire event has been detected in the chamber 14 and the fire suppression supply 25 has been activated to release fire suppressant into the chamber 14. In this instance, the linear actuator member 27 of the trigger mechanism 26 has been retracted thereby pulling the arm member 28 in a direction toward the fire suppression supply 25 such that the piercing pin penetrates the opening of the fire suppression supply 25 to release the fire suppressant from the canister and into the chamber 14 of the device 10. An electric circuit may be provided about the trigger mechanism 26 to determine the state of the trigger mechanism 26, such that upon triggering and release of the fire suppressant from the fire suppressing supply 25, the control system of the device 10 can be updated to request replacement of the fire suppressing supply cannister 25. It will be appreciated that the maimer in which the fire suppressing supply 25 is triggered can vary, as will be appreciated by those skilled in the art.

[0046] Referring to Fig. 4, a schematic block diagram of the control system for controlling the fire detection and suppression unit 20 of the device 10 is depicted. Each unit 20 has a controller 18 that includes a microprocessor that is capable of receiving signals from the image sensors 24 and gas sensors 23 and to process signals to determine whether the data received indicates the presence of fire event requiring intervention by the fire suppression unit 20. Each controller 18 of each unit 20 comprises a processor unit 34 that receives the images from the image sensors 24 and performs a computer vision analysis to determine a potential fire event. When a potential fire event is detected the processor unit 34 feeds the images to a neural network processor 35 for review and analysis. The neural network processor 35 assesses the image received from the image sensors 24 to determine if they are representative of a fire event that warrants action and suppression. Upon the neural network processor 35 performing the analysis and confirming the fire event, the neural network processor 35 updates the peripheral controller 36 to trigger action and fire suppression.

[0047] The neural network processor 35 employs an algorithm using a convolutional neural network (CNN) that classifies images into different event categories, such as “no flame”, “hot component”, and “flame”. The neural network processor 35 is trained by collecting sample labelled data of each event category. The data is then randomly split into training, validation, and testing sets.

[0048] The neural network processor 35 is then specified with various layers, such as:

[0049] • Convolutional layers which apply filters to create filter maps;

[0050] • Activation layers which typically use ReLU (Rectified Linear Unit) activation functions after each convolution;

[0051] • Pooling layers to reduce the dimensions of the feature maps;

[0052] • Flattening layers to flatten the data from the pooling layer into a format the dense layers can understand; and

[0053] • Dense or fully connected layers to perform the classification on the features extracted from the convolutional layers.

[0054] The neural network processor 35 is compiled with a chosen loss function, optimiser, and metrics for monitoring model performance during training. The neural network processor 35 is then trained on the training dataset using a batch size (number of images processed before updating weights), epochs (iterations for training), and is validated using the validation dataset.

[0055] The neural network processor 35 is then evaluated using the test dataset and tuned based on results. Tuning can encompass changing elements such as learning rate, batch size, and model architecture. The model is then retrained after this, for use as a secondary analysis, as required.

[0056] Upon power-up of each unit 20, the associated controller 18 will initiate a check routine that will seek to initialise all internal libraries, variables and functions and establish communication with the device 10.

[0057] Upon initialisation of the controller 18, communication messages will be exchanged with the device 10 at regular intervals, which may be every 100ms. Such messages may include data associated with the device serial number, identification number of the CO2 cartridge installed and whether the cartridge is unused, device ready signals, error codes and the like.

[0058] The controller 18 will then initialise the image sensors 24, which may include initialising the thermal camera and define the operating parameters thereof. The operating parameters may include temperature range, gain control mode, thresholds and image format. Other IOCTLS (input output controls) of the camera may also be established, such as the region of interest for focussing the thermal camera to obtain the images.

[0059] Prior to use of the device, the controller 18 will perform a homing routine to ensure that the trigger mechanism 26 of the fire suppression and detection unit 20 is in the correct position and that the fire suppression supply 25 is ready for use and not in a triggered state.

[0060] Once the device 10 has commenced use, a thermistor may check the temperature of the chamber 14 of the device 10 every 10 seconds and dynamically update the image sensors 24 to ensure that the FLIR camera’s temperature range values are sufficient to reduce the amount of background noise.

[0061] During laser cutting, the controller 18 of the device will conduct a primary analysis of the process to identify a flame event. Such a primary analysis will extract the latest image data reading from the image sensors 24 and analyse the image data using a “blob” tracking algorithm. The blob tracking algorithm will review the image frame received and search the frame for the presence of “blobs” in the frame. A “blob” can be described as a hot area of interest and comprises a region of coloured pixels in the frame. The control unit will keep the data associated with the largest “blob” identified and if multiple blobs are detected, all blobs other than the largest blob will be disregarded. For each detected blob, the controller 18 will determine the size (number of pixels) and x and y positions of the centre of the blob will be recorded. The control unit will also measure the maximum temp of the blob using the thermal data received.

[0062] To determine if a fire event has occurred, the controller 18 will continue to process the image data from the image sensors 24 over time. This is achieved by determining if the blob size is increasing across consecutive intervals, which could indicate a growing sized flame. The control unit can also analyse the image date to determine if the blob is moving by analysing whether the position of centre of the blob in the received frame has not changed significantly from the previous received frame. This could indicate a hot mechanical component not associated with a flame or a small flame that is following the beam of the laser.

[0063] As part of the primary analysis, the controller 18 can determine if:

[0064] • a blob has been continuously detected for a predetermined period of time ( for example - >10 seconds);

[0065] • if the data indicates growth of the blob ( for example - past 1 sec average > 10 secs); and

[0066] • if the average flame size is above the minimum detection threshold.

[0067] If these conditions have been determined, a flame event will be identified by the primary analysis and a secondary analysis of the data will be initiated.

[0068] As discussed above, the secondary analysis of the data will commence if:

[0069] • The average pixel size over the past second is greater than the extinguish threshold.

[0070] • The max temp detected is above the safe threshold (for example - above 190°C); and

[0071] • The blob isn’t moving.

[0072] • The time continuously detected reaches a threshold.

[0073] The secondary analysis will be conducted by the neural network processor 35 analysing the images using the neural network algorithm to confirm whether a flame event has been detected, or whether the event is merely a hot mechanical component or false alarm.

[0074] When both the Primary and Secondary analysis methods confirm a fire event, the controller 18 of the unit 20 will function as follows:

[0075] • Deactivate the laser and move the laser head to a safe location within the work chamber 14, away from the fire, to prevent damage;

[0076] • Disable exhaust fans and disconnect any fume extractor to prevent more air from entering the work chamber 14;

[0077] • Activate the fire suppression supply to flood the work chamber 14 with fire suppressant to extinguish the fire;

[0078] • Issue an audible warning alarm sound. • Send a message to the cutting device 10 advising of a fire event.

[0079] Following detection and suppression of a fire event, the controller 18 of the unit 20 will need to perform a system reset and a message will be sent to the cutter device 10 informing the user of the device 10 of the need to replace the fire suppression supply 25.

[0080] It will be appreciated that the fire control system of the present invention conducts an initial review of image data to indicate the likelihood of a flame event having been detected and then reviews the data through using a neural network algorithm to confirm that the event is a fire event that requires action. Such a two-phase analysis system minimises the event of a false alarm and ensures that the device 10 is working in an efficient maimer.

[0081] Throughout the specification and claims the word “comprise” and its derivatives are intended to have an inclusive rather than exclusive meaning unless the contrary is expressly stated or the context requires otherwise. That is, the word “comprise” and its derivatives will be taken to indicate the inclusion of not only the listed components, steps or features that it directly references, but also other components, steps or features not specifically listed, unless the contrary is expressly stated or the context requires otherwise.

[0082] It will be appreciated by those skilled in the art that many modifications and variations may be made to the methods of the invention described herein without departing from the spirit and scope of the invention.

Claims

The claims defining the invention are as follows:

1. A laser cutting or engraving device comprising: a body having an enclosed chamber; a laser cutter for cutting or engraving a work material present within the enclosed chamber; a control unit for controlling operation of the laser cutter; and a fire detection and suppression unit mounted to the body to be located within the enclosed chamber, the fire detection and compression unit comprising one or more image sensors configured to take an image of the laser cutter as the laser cutter performs cutting or engraving of the work material; a fire suppression supply actuable to supply fire suppression material into the enclosed chamber; and a controller for receiving image data from the one or more image sensors and configured to perform a primary analysis of the image data to determine a likelihood of a fire event within the enclosed chamber, and upon the primary analysis determining that a fire event has likely occurred, the controller is configured to conduct a secondary analysis of the image data to confirm that a fire event has occurred and upon the secondary analysis confirming a fire event the controller is configured to actuate the fire suppression supply to suppress a fire associated with the fire event within the enclosed chamber.

2. A laser cutting or engraving device according to claim 1, wherein the controller comprises a processor unit that performs the primary analysis of the image data by receiving updated image data from the one or more image sensors and performing computer vision analysis of the image data over time to determine that a fire event has likely occurred.

3. A laser cutting or engraving device according to claim 2, wherein the controller further comprises a neural network processor that receives the image data and compares the image data against predetermined image categories to confirm whether the image data is representative of a fire event.

4. A laser cutting or engraving device according to claim 3, wherein the predetermined image categories are images classified as being representative of a “no flame” category, “hot component” category, or “flame” category.

5. A laser cutting or engraving device according to claim 3, wherein if the neural network processer determines the image data is representative of a “flame” category, the controller actuates the fire suppression supply to supply fire suppression material into the enclosed chamber.

6. A laser cutting or engraving device according to claim 1, wherein the fire suppression supply comprises a trigger mechanism that is actuated by the controller to supply fire suppression material into the enclosed chamber.

7. A laser cutting or engraving device according to claim 6, wherein the trigger mechanism is configured to pierce a cannister of fire suppression material stored under pressure within the cannister to release the fire suppression material from the cannister into the enclosed container.

8. A laser cutting or engraving device according to claim 1, wherein the one or more image sensors comprises a visual camera and / or a thermal image camera.

9. A laser cutting or engraving device according to claim 1, wherein the one or more image sensors may comprise a Forward Looking Infrared (FLIR) camera.

10. A method detection and suppression of a fire in a laser cutting or engraving device comprising: collecting image data of an interface between a laser cutter and a work material being cut by the laser cutter; performing a primary analysis of said collected image data to determine a likely fire event at the interface between the laser cutter and the work material; upon detection of a likely fire event at the interface between the laser cutter and the work material, conducting a secondary analysis of the collected image data of the interface between a laser cutter and a work material using a neural network processor that receives the image dataand compares the image data against predetermined image categories to confirm whether the image data is representative of a fire event; and upon determining that the secondary analysis of the collected image data confirms that there is a fire event at the interface between the laser cutter and the work material, releasing a fire suppressant material to the interface between the laser cutter and the work material.

11. A method according to claim 10, wherein the primary analysis of said collected image data comprises performing a vision analysis of the collected image data.

12. A method according to claim 11, wherein the vision analysis of the collected image data comprises reviewing the image data frames and search the image data frame for the presence of predetermined regions of interest in the pixels present in the image data frame.

13. A method according to claim 12, wherein the predetermined regions of interest in the pixels present in the image data frame comprises a region of coloured pixels having an area exceeding a predetermined size.

14. A method according to claim 13, wherein the predetermined regions of interest in the pixels present in the image data frame are tracked across successive image data frames to determine whether the predetermined regions of interest in the pixels is increasing, reducing, remaining constant or moving.

15. A method according to claim 14, wherein if the predetermined regions of interest in the pixels present in successive image data frames is increasing, remaining constant and / or above a predetermined area the primary analysis will determine that a fire event at the interface between the laser cutter and the work material is likely.

16. A method according to claim 10, wherein the secondary analysis of the collected image data of the interface between a laser cutter and a work material comprises the neural network processor employing an algorithm using a convolutional neural network (CNN) that classifies the image data into different event categories, including, “no flame”, “hot component”, and “flame”.

Images