System and method for testing durability of gas heating furnace based on audio-visual collaborative diagnosis

Abstract

The invention relates to the technical field of gas heating furnace detection, in particular to a gas heating furnace durability test system and method for audio-visual collaborative diagnosis, and the system comprises a system test pipeline, an electrical control system, a camera visual collection and diagnosis subsystem, and an MEMS microphone voiceprint collection and diagnosis subsystem. The system testing pipeline comprises a gas testing pipeline, a domestic water pipeline, a heating water supply branch, a heating water return branch and a bathroom water outlet branch, all the pipelines are connected with the gas heating furnace through corresponding connectors, and instruments and valves used for monitoring, regulating and controlling fluid parameters are arranged on the pipelines. According to the invention, image recognition of the digital panel is realized through a specific process, digits on the digital panel of the gas heating stove can be accurately recognized in real time, a basis is provided for a durability test application program, full automation of durability test is realized, and manpower is released; a pickup module and a voiceprint diagnosis algorithm which are designed by taking an MEMS microphone as a core can detect abnormal sound when a component works as soon as possible.

Description

Technical Field

[0001] The present invention relates to the technical field of gas heating furnace detection, and in particular to a gas heating furnace durability performance testing system and method for audio-visual collaborative diagnosis. Background Art

[0002] Gas-fired hot water boilers (referred to as heating boilers) are commonly used heating equipment in northern households. Their reliability and durability are key factors in measuring product quality and have attracted much attention from manufacturers. Under typical northern usage conditions, from November to March of the following year, if the heating boiler is running 24 hours a day, the average annual operating time is about 3,600 hours. Once a heating boiler fails during use, it will have a serious adverse impact on people's lives.

[0003] To ensure product quality, manufacturers need to conduct durability tests on the entire heating furnace before it leaves the factory. This test is crucial to product reliability management, because durability testing can promptly discover quality problems that are not easily detected during product design, thereby eliminating quality defects at an early stage. The design life of domestic gas-fired heating water boilers is generally 10 to 15 years, and accordingly, its average trouble-free working time is required to be greater than 10,000 hours. However, if a 10,000-hour durability test is to be conducted, a test cycle of up to several months is required, which is time-consuming and costly.

[0004] To improve work efficiency and better record test information, factories usually use heating furnace durability test benches for semi-automatic testing. However, during the durability test, the heating furnace may fail, so it is necessary to arrange personnel on duty for timely processing. In terms of relevant standards, wall-mounted boilers need to be tested under three living water inlet pressure conditions: low pressure, normal pressure, and high pressure, and low flow, normal flow, and high flow tests must be performed under each pressure condition.

[0005] The existing related technologies have certain limitations. For example, the utility model patent with publication number CN202511981U discloses a universal test equipment for gas water heaters. Although a water pressure regulating valve is installed in the domestic water inlet pipe, it lacks a water flow controller and cannot meet the three flow test requirements at each pressure of the relevant standards. Moreover, the patent lacks intervention measures when an abnormal heating furnace is encountered during the test.

[0006] In addition, the invention patent with publication number CN108871835A discloses an intelligent detection and control system for a gas water heater or a wall-mounted boiler, which is mainly composed of an intelligent detection controller embedded in the water heater or the wall-mounted boiler, a cloud server and a background terminal. The system performs real-time diagnosis of water pumps, solenoid valves, thermostats, pulse igniters and other devices in the water heater or the wall-mounted boiler by means of electrical connection, and abnormal conditions are uploaded to the cloud server through a wireless module. However, this method requires deep integration with the water heater or the wall-mounted boiler. When the internal device model is changed, the intelligent detection controller needs to be re-debugged, the compatibility is poor, and the operation is cumbersome. In addition, the patent mainly focuses on the detection of electrical device failures, and lacks testing of the bathroom and heating performance of the water heater or the wall-mounted boiler.

[0007] In summary, the existing gas heating water boiler testing technology has shortcomings in meeting standard testing requirements, coping with fault handling and system compatibility, and needs further improvement and perfection. Summary of the invention

[0008] The purpose of the present invention is to overcome the above problems or at least partially solve the above problems, and to propose a gas heating furnace durability performance testing system and method with audio-visual collaborative diagnosis.

[0009] To achieve the above object, the present invention provides the following technical solution: a gas heating furnace durability performance test system with audio-visual collaborative diagnosis, comprising:

[0010] System test pipeline, the system test pipeline includes a gas test pipeline, a domestic water pipeline, a heating water supply branch, a heating return water branch and a bathroom water outlet branch. Each pipeline is connected to the gas heating furnace through a corresponding interface to realize the gas supply, water circulation and drainage functions, and the pipeline is provided with instruments and valves for monitoring and regulating fluid parameters;

[0011] The electrical control system collects and controls the digital flow meter, temperature sensor, pressure sensor, pneumatic valve, gas pressure regulating valve, and water flow proportional valve installed on the test pipeline of the system through PLC. The PLC communicates data with the industrial computer through Ethernet. The industrial computer runs a program for data processing and system control.

[0012] The camera visual acquisition diagnosis subsystem includes a camera installed on the test bench station with its lens facing the digital panel of the gas heating furnace. The camera is powered by POE and connected to the industrial computer through the network. The industrial computer runs an image recognition program to collect image information of the digital panel of the gas heating furnace and perform recognition, and send the recognition result to the durability test application program.

[0013] The MEMS microphone voiceprint acquisition and diagnosis subsystem includes a sound pickup module with a MEMS microphone as the core, which is magnetically attached to a gas heating furnace. The sound pickup module is connected to an industrial computer through an acquisition gateway. The acquisition gateway realizes analog-to-digital conversion and transmits real-time audio information to the industrial computer. The voiceprint diagnosis algorithm runs in the industrial computer to collect sound information when the gas heating furnace is working and perform auscultation diagnosis, and send the diagnosis results to a durability test application. The voiceprint diagnosis algorithm includes sound detection, sound preprocessing, acoustic feature extraction, voiceprint feature extraction model and feature similarity matching module, and its workflow includes training, registration and reasoning stages.

[0014] In a preferred embodiment, the inlet of the gas test pipeline is connected to the factory gas port, and the gas test pipeline leads to at least one gas branch. The gas branch is sequentially provided with a manual ball valve, a gas pressure regulating valve, a gas pressure gauge, a gas flow meter, a pneumatic valve, a digital gas pressure gauge, and a gas temperature sensor. The gas branch outlet is connected to the gas inlet of the heating furnace. The gas pressure regulating valve adjusts the valve through a stepper motor, the stepper motor is connected to the motor driver, and the motor driver is connected to the PLC through the serial port.

[0015] In a preferred embodiment, the inlet of the domestic water pipeline is connected to the cold water outlet of the factory, at least one domestic water branch is led out from the domestic water pipeline, the domestic water branch provides cold water for the heating furnace, a manual ball valve and a filter are installed on the domestic water branch, and three pressure regulating branches are separated after the filter, namely high-pressure, normal-pressure and low-pressure branches, wherein the high-pressure branch is installed with a pressure gauge, a pneumatic valve and a booster pump in sequence, the normal-pressure branch and the low-pressure branch are installed with a pressure gauge, a pneumatic valve and a water pressure regulating valve in sequence, and the three pressure regulating branches converge to the domestic water pipe. A digital flow meter is installed on the domestic water branch, and the flow meter is divided into four flow branches, namely high flow, normal flow, low flow branch and heat flow exchange branch. Pneumatic valves and proportional valves are installed on the high flow, normal flow and low flow branches respectively, and then pressure sensors and temperature sensors are installed on them after they converge on the domestic water branch. The terminal outlet of the domestic water branch is connected to the domestic water inlet of the heating furnace. After the pneumatic valve is installed on the heat flow exchange branch, it is connected to the plate heat exchanger, and the drainage branch is led out from the plate heat exchanger and connected to the factory drain outlet.

[0016] In a preferred embodiment, the heating water supply branch is connected to the heating water supply outlet of the gas heating furnace, the heating water supply branch is connected to the heating water supply outlet of the gas heating furnace, the heating water supply branch is sequentially provided with a temperature sensor, a water replenishment branch tee, a pressure relief test point, a drainage branch, a water flow proportional valve and a pneumatic valve, and the end of the heating water supply branch is connected to the heating water outlet of the plate heat exchanger; the heating return water branch is led out from the heating return water outlet of the plate heat exchanger, and is sequentially provided with a water blowing branch tee, an expansion water tank, a digital flow proportional valve and a pneumatic valve. A gas flow meter, a pressure sensor, and a temperature sensor are installed. A compressed air water blowing branch is led out at the tee of the water blowing branch and a pneumatic valve is installed. A pressure relief point is led out next to the expansion water tank connection port. The end of the heating return water branch is connected to the heating return water outlet of the gas heating furnace; the bathroom water outlet branch is connected to the bathroom water outlet of the gas heating furnace. The bathroom water outlet branch is successively installed with a temperature sensor, a pressure gauge, a pneumatic valve, and a drainage tee. The end of the bathroom water outlet branch is connected to the factory drain, and the drainage tee is connected to the drainage branch.

[0017] In a preferred embodiment, in the electrical control system, the program running on the industrial computer includes a human-computer interaction interface program, a database and a model. The human-computer interaction interface program is used to realize the interaction between the user and the system, the database is used to store the collected data and diagnostic results, and the model is used to analyze and process the collected data.

[0018] In a preferred embodiment, in the camera visual acquisition diagnosis subsystem, the camera is a high-definition camera that can acquire high-definition image information of the digital panel of the gas heating furnace in real time. The image recognition program can recognize the display content on the digital panel, including fault codes, operating status parameters, etc., and classify and store the recognition results.

[0019] In a preferred embodiment, in the MEMS microphone voiceprint acquisition and diagnosis subsystem, the MEMS microphone has high sensitivity and low noise characteristics, and can accurately collect sound information when the gas heating furnace is working. The voiceprint diagnosis algorithm can extract and analyze the features of the collected sound information, determine whether there is abnormal sound in the gas heating furnace, and record and alarm the diagnosis results.

[0020] A method for testing the durability of a gas heating furnace with audio-visual collaborative diagnosis is applied to the above-mentioned gas heating furnace durability testing system with audio-visual collaborative diagnosis. The durability test application running in the industrial computer supports three modes, namely, a start-stop cycle mode of a domestic hot water function, a start-stop cycle mode of a heating function, and a start-stop cycle mode of switching between a domestic hot water function and a heating function. The start-stop cycle mode of the domestic hot water function includes the following steps:

[0021] Gas preparation steps: manually open the gas inlet ball valve, programmatically open the gas branch and monitor the gas flow and gas temperature;

[0022] Test steps for the first cycle (N1): manually open the domestic water inlet ball valve, the program starts counting, the program opens the high-pressure branch, closes the normal pressure and low-pressure branches, the program opens the high-flow branch, closes the normal flow and low-flow branches and the heat flow exchange branch, opens the bathroom outlet branch, monitors the flow, pressure, temperature of the domestic water branch and the temperature of the bathroom outlet branch, the gas heating furnace starts the domestic hot water function, burns for the set time, closes the bathroom outlet branch and stops for the set time, automatically opens the bathroom outlet branch, starts the gas heating furnace to burn for the set time, closes the bathroom outlet branch and stops for the set time, and the program counts and records;

[0023] Subsequent cycle test steps (N2-N9): Perform subsequent cycle N2-N9 tests according to different set pressure and flow combinations, and perform tests under different pressures of domestic water, with different flow branches opened and other flow branches closed;

[0024] Cycle cycle steps: In the start-stop cycle mode of the domestic hot water function, the 9 cycles N1-N9 are cycled;

[0025] Abnormal handling steps: In the cycle mode of the start-stop cycle test, when the digital panel of the gas heating furnace displays an abnormality and an error code is displayed, the camera visual acquisition diagnosis subsystem takes real-time video, and the image recognition program running in the industrial computer informs the durability test application of the recognition result. The durability test application takes the action of closing the gas valve, domestic water valve or heating water supply valve according to the recognition result.

[0026] In a preferred embodiment, the image recognition program specifically includes the following steps for the image recognition process of the digital panel: define a character definition table of a 7-segment digital image; obtain the original image; obtain the digital panel area, i.e., ROI, through canny edge detection; convert the ROI area color image to a grayscale image; perform Gaussian blur processing on the ROI grayscale image; perform binarization processing on the ROI area image; perform morphological opening operation processing; perform contour filtering; perform image segmentation; extract the character ROI; divide the character ROI definition into 7 segments; calculate the width of each segment Degree and height; determine whether the non-zero pixels of each segment are greater than 50% of the total area of ​​the segment. If so, set 1 for the segment whose non-zero pixels are greater than 50% of the total area of ​​the segment, otherwise set 0 for the segment whose non-zero pixels are less than 50% of the total area of ​​the segment; write the segments set to 1 or 0 into the character list; determine whether the 7 segments have been processed. If so, look up the characters in the character lookup table, record the recognized characters, and then determine whether the 2 characters have been processed. If so, jump to the step of obtaining the original image. If not, jump to the step of extracting the character ROI. If the 7 segments have not been processed, jump to the step of extracting the character ROI.

[0027] In a preferred embodiment, during the test: the MEMS microphone voiceprint collection and diagnosis subsystem collects the sounds of components including but not limited to the working sounds of the gas heating furnace burner, water pump, fan, and proportional valve; auscultation diagnosis is performed through the voiceprint diagnosis algorithm running in the industrial computer, and the results of the auscultation diagnosis are sent to the durability test application through shared memory or internal network communication; the durability test application records the auscultation results and pauses or stops the test mode according to the degree;

[0028] The voiceprint recognition reasoning process includes:

[0029] (1) Sound is picked up by a MEMS microphone;

[0030] (2) Decomposing the sound signal into multiple layers and obtaining the high-frequency signal coefficients of each layer;

[0031] (3) Sort and evaluate the coefficients of each layer and calculate the threshold to remove noise;

[0032] (4) Convert the denoised sound signal into a spectrogram;

[0033] (5) Use CNN one-dimensional convolution to extract low-level features and reduce spatial dimensions;

[0034] (6) The dot product method is used to calculate the similarity, and the dot product result is converted into self-attention weight through the softmax function;

[0035] (7) Connect the outputs of steps (5) and (6) together; softmax classification;

[0036] (8) The speech features in the database are input into the similarity matching module to calculate the similarity score. The system determines the result based on the output similarity score and the predefined threshold.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] 1. The present invention realizes image recognition of digital panels through a specific process, and can recognize the digital numbers on the digital panel of the gas heating furnace in real time and accurately, providing a basis for the durability test application, realizing the full automation of the durability test, and releasing manpower;

[0039] 2. The present invention also adopts a sound pickup module and voiceprint diagnosis algorithm designed with MEMS microphone as the core, which can detect abnormal sounds with low frequency and small sound amplitude when the burner and other components are working as early as possible, establish a voiceprint library through positive sample learning, and realize intelligent marking of abnormal voiceprints; in terms of pipeline design, the three branches of the domestic water inlet pipeline meet the requirements of durability test for pressure and flow, and improve the test efficiency. The expansion water tank on the heating return water line can keep the system stable; at the same time, the electronic control system composed of industrial computer and PLC realizes the collection and control of various components;

[0040] 3. The present invention uses a camera to identify error codes and automatically processes them without manual supervision, and uses a MEMS microphone to collect sound to detect faults early and to evaluate the quality of the heating furnace. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a schematic diagram of the test pipeline of the system of the present invention;

[0042] Figure 2 A schematic diagram of the electrical connection of the system of the present invention;

[0043] Figure 3 A schematic diagram of the implementation and deployment of the system of the present invention;

[0044] Figure 4 The figure is a flow chart of the image recognition of the digital panel according to the present invention. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] See also Figure 1-4The present invention provides a technical solution: a gas heating furnace durability performance test system with audio-visual collaborative diagnosis, a set of test pipelines are designed and pneumatic valves, flow, pressure, temperature sensors and other components are set on the pipelines, the inlet of the pipelines are connected to the water and gas circuits of the factory, and the outlet of the pipelines is connected to the heating furnace. An automated durability performance test system composed of an industrial computer and a PLC is used to collect and control the electrical components on the pipeline to achieve automation. A MEMS microphone is used to collect the operating sound of the heating furnace in real time, and an RGB camera is used to monitor the display panel of the heating furnace, and the electronic function code numbers of the panel are identified in real time. In the industrial computer, an algorithm is run in the industrial computer to analyze abnormal sounds and identify the function code numbers of the panel in real time. The abnormal sound is used as one of the bases for evaluating the quality of the heating furnace, and the function code numbers are used as an important basis for the durability performance test system to handle abnormal compliance of the heating furnace.

[0047] Specifically include:

[0048] System test pipelines, including gas test pipelines, domestic water pipelines, heating water supply branches, heating return water branches and bathroom outlet branches. Each pipeline is connected to the gas heating furnace through a corresponding interface to achieve gas supply, water circulation and drainage functions, and the pipeline is equipped with instruments and valves for monitoring and regulating fluid parameters;

[0049] The electrical control system collects and controls the digital flowmeter, temperature sensor, pressure sensor, pneumatic valve, gas pressure regulating valve, and water flow proportional valve installed on the test pipeline through the PLC. The PLC communicates data with the industrial computer through Ethernet, and the industrial computer runs programs for data processing and system control.

[0050] The camera visual acquisition diagnosis subsystem includes a camera installed on the test bench station with its lens facing the digital panel of the gas heating furnace. The camera is powered by POE and connected to the industrial computer through the network. The industrial computer runs an image recognition program to collect image information of the digital panel of the gas heating furnace and perform recognition, and send the recognition result to the durability test application program.

[0051] The MEMS microphone voiceprint acquisition and diagnosis subsystem includes a sound pickup module with a MEMS microphone as the core, which is magnetically attached to the gas heating furnace. The sound pickup module is connected to the industrial computer through a collection gateway. The collection gateway realizes analog-to-digital conversion and transmits real-time audio information to the industrial computer. The voiceprint diagnosis algorithm runs in the industrial computer to collect sound information when the gas heating furnace is working and perform auscultation diagnosis, and send the diagnosis results to the durability test application; the voiceprint diagnosis algorithm includes sound detection, sound preprocessing, acoustic feature extraction, voiceprint feature extraction model and feature similarity matching module, and its workflow includes training, registration and reasoning stages.

[0052] During specific implementation, the inlet of the gas test pipeline is connected to the factory gas port, and a ball valve 11 and a pneumatic valve 12 are arranged on the gas test pipeline in sequence. At least one gas branch is led out from the pneumatic valve 12, and a manual ball valve 101, a gas pressure regulating valve 102, a gas pressure gauge 103, a gas flow meter 104, a pneumatic valve 105, a digital gas pressure gauge 106, and a gas temperature sensor 107 are arranged on the gas branch in sequence. The gas branch outlet is connected to the gas inlet of the heating furnace, and the gas pressure regulating valve 102 adjusts the valve through a stepper motor, the stepper motor is connected to the motor driver, and the motor driver is connected to the PLC through the serial port.

[0053] During specific implementation, the inlet of the domestic water pipeline is connected to the cold water outlet of the factory. A ball valve 21 and a pneumatic valve 22 are arranged on the domestic water pipeline in sequence. At least one domestic water branch is led out from the pneumatic valve 22 to provide cold water for the heating furnace. A manual ball valve 201 and a filter 202 are installed on the domestic water branch. Three pressure regulating branches are separated after the filter, namely, high-pressure, normal-pressure and low-pressure branches. The high-pressure branch is installed with a pressure gauge 2011, a pneumatic valve 2012 and a booster pump 2013 in sequence, the normal-pressure branch is installed with a pressure gauge 2021, a pneumatic valve 2022 and a water pressure regulating valve 2023 in sequence, the low-pressure branch is installed with a pressure gauge 2031, a pneumatic valve 2032 and a water pressure regulating valve 2033 in sequence. The three pressure regulating branches are connected to the domestic water branch where a 203 digital flow meter is installed. The flow meter is divided into four flow branches, namely high flow, normal flow, low flow branch and heat flow exchange branch. Pneumatic valves and proportional valves are installed on the high flow, normal flow and low flow branches respectively, and then they are connected to the domestic water branch where a 204 pressure sensor and a 205 temperature sensor are installed respectively. The domestic water branch terminal outlet is connected to the domestic water inlet of the heating furnace. The heat flow exchange branch is connected to the plate heat exchanger after a pneumatic valve is installed. The domestic water is discharged through the plate heat exchanger to achieve heat exchange. The drainage branch is led out from the plate heat exchanger and connected to the factory drain.

[0054] During specific implementation, the heating water supply branch is connected to the heating water supply outlet of the gas heating furnace. The heating water supply branch is sequentially provided with a 401 temperature sensor, a 402 water replenishment branch tee, a 403 pressure relief test point, a 404 drainage branch, a 405 water flow proportional valve and a 406 pneumatic valve. The end of the heating water supply branch is connected to the heating water outlet of the plate heat exchanger. The heating return water branch is led out from the heating return water outlet of the plate heat exchanger, and is sequentially provided with a 501 water blowing branch tee, a 502 expansion water tank, a 503 digital flow meter, a 504 pressure sensor and a 505 temperature sensor. A compressed air water blowing branch is led out at the tee 501 of the water blowing branch. The road is connected and pneumatic valves are installed for on and off. The water blowing branch is mainly used for draining the residual water in the domestic water pipe after the durability performance test system of the gas heating furnace is completed to prevent the pipe from freezing in winter. A pressure relief point is introduced next to the 502 expansion water tank connection port, and the end of the heating return water branch is connected to the heating return water outlet of the gas heating furnace; the bathroom water outlet branch is connected to the bathroom water outlet of the gas heating furnace, and the bathroom water outlet branch is installed with 601 temperature sensor, 602 pressure gauge, 603 pneumatic valve, 604 drainage tee in sequence. The end of the bathroom water outlet branch is connected to the factory drain, and the 604 drainage tee is connected to the 404 drainage branch.

[0055] In specific implementation, in the electrical control system, the program running on the industrial computer includes a human-computer interaction interface program, a database and a model. The human-computer interaction interface program is used to realize the interactive operation between the user and the system, the database is used to store the collected data and diagnosis results, and the model is used to analyze and process the collected data;

[0056] Further, the electrical control system connection characteristics are described as follows:

[0057] The rotating motor on the 102 gas pressure regulating valve on the gas branch is connected to the motor driver, and the motor driver signal line is connected to the PWM port of the PLC. When the user clicks on the human-machine interactive interface running on the industrial computer to increase or decrease the value, a signal will be sent to the PLC, and the PLC will eventually rotate the motor forward or reverse to achieve the purpose of regulating the pressure of the 102 gas pressure regulating valve. The 104 gas flow meter on the gas branch is connected to the analog input port of the PLC; the 105 pneumatic valve is connected to the control gas path of the solenoid valve, and the IO control line of the solenoid valve is connected to the switch output port of the PLC; the analog signal lines of the 106 digital gas pressure gauge and the 107 gas temperature sensor are connected to the analog input port of the PLC.

[0058] The solenoid valves connected to the control gas circuits of the 2012 pneumatic valves, 2022 pneumatic valves and 2032 pneumatic valves of the three domestic water pressure branches are connected to the switch output ports of the PLC respectively; the control line of the 2013 booster pump on the high-pressure branch is connected to the control output port of the frequency converter, and the RS485 line of the frequency converter is connected to the RS485 port of the PLC. The booster pump boost setting is set by the user by clicking on the human-machine interactive interface running on the industrial computer, and the PLC issues instructions to control the booster pump speed to achieve the purpose of boosting. The 2023 water pressure regulating valve and 2033 water pressure regulating valve on the normal pressure and low pressure branches are manually adjusted. The electrical signals of the solenoid valves and proportional valves connected to the control gas circuits of the three pneumatic valves for domestic water flow are connected to the switch output port and analog output port of the PLC respectively. The 203 digital flowmeter, 204 pressure sensor and 205 temperature sensor of the domestic water branch are connected to the analog input port of the PLC respectively. The control lines of the solenoid valves connected to the pneumatic valve control gas circuits installed on the heating circuit water replenishment branch, the heat flow exchange branch and the compressed air water blowing branch are respectively connected to the switching output ports of the PLC.

[0059] The 401 temperature sensor, 405 water flow proportional valve, 406 pneumatic valve on the heating water supply branch and the 503 digital flow meter, 504 pressure sensor, 505 temperature sensor on the heating return water branch are respectively connected to the PLC control interface.

[0060] In specific implementation, in the camera visual acquisition diagnosis subsystem, the camera is a high-definition camera that can collect high-definition image information of the digital panel of the gas heating furnace in real time. The image recognition program can identify the display content on the digital panel, including fault codes, operating status parameters, etc., and classify and store the recognition results.

[0061] In the specific implementation, in the MEMS microphone voiceprint collection and diagnosis subsystem, the MEMS microphone has the characteristics of high sensitivity and low noise, and can accurately collect the sound information of the gas heating furnace when it is working. The voiceprint diagnosis algorithm can extract and analyze the features of the collected sound information, determine whether there is abnormal sound in the gas heating furnace, and record and alarm the diagnosis results.

[0062] In order to test the durability of the gas heating furnace, the present invention also proposes a gas heating furnace durability test method with audio-visual collaborative diagnosis. The durability test application running in the industrial computer supports three modes, namely, the start-stop cycle mode of the domestic hot water function, the start-stop cycle mode of the heating function, and the start-stop cycle mode of the domestic hot water function and the heating function conversion. Taking the start-stop cycle mode of the domestic hot water function as an example, the manual gas inlet ball valve is used, and the program opens the gas branch and monitors the gas flow and gas temperature.

[0063] First, start the first cycle, recorded as N1: manually open the domestic water inlet ball valve, the program starts counting, the program opens the high-pressure branch, and closes the normal pressure and low-pressure branches; the program opens the high-flow branch, and closes the normal flow and low-flow branches and the heat flow exchange branch; opens the bathroom outlet branch; monitors the flow, pressure, temperature of the domestic water branch and the temperature of the bathroom outlet branch; the gas heating furnace starts the domestic hot water function; after burning for 2 minutes (time is adjustable); closes the bathroom outlet branch and stops for 1 minute; automatically opens the bathroom outlet branch; starts the gas heating furnace to burn for 2 minutes; closes the bathroom outlet branch and stops for 1 minute; the program counts and records.

[0064] The N2 cycle is tested under high pressure of domestic water with the normal flow branch open and other flow branches closed;

[0065] The N3 cycle is tested under high pressure of domestic water with low flow branches open and other flow branches closed;

[0066] The N4 cycle is tested under normal pressure of domestic water, with the high flow branch open and other flow branches closed;

[0067] The N5 cycle is tested under normal pressure of domestic water, with the normal flow branch open and other flow branches closed;

[0068] The N6 cycle is tested under normal pressure of domestic water, with the low flow branch open and other flow branches closed;

[0069] The N7 cycle is tested under low pressure of domestic water, with the high flow branch opened and other flow branches closed;

[0070] The N8 cycle is tested under low pressure of domestic water with the normal flow branch open and other flow branches closed;

[0071] The N9 cycle is tested under low pressure of domestic water, with the low flow branch opened and other flow branches closed;

[0072] In the start-stop cycle mode of the domestic hot water function, 9 cycles N1-N9 are cycled.

[0073] In the cycle mode of the start-stop cycle test, the gas heating furnace is conducting a durability test of the domestic hot water function. During the test, under normal circumstances, the digital panel of the gas heating furnace will display the temperature value in real time, but when an abnormal situation occurs, an error code will be displayed on the digital panel. When seeing this error code, the human will take corresponding measures, such as closing the gas valve to stop gas supply or closing the pneumatic valve of the domestic water branch to stop water supply or closing the pneumatic valve of the heating water supply branch to stop heating. However, due to the long cycle of this durability test, it is very time-consuming and energy-consuming to rely on manual observation of abnormalities to take measures. The present invention uses a camera to take real-time video of the digital panel of the gas heating furnace, runs an image recognition program in an industrial computer, and informs the durability test application of the recognized results. The durability test application takes action (closing the gas valve or the domestic water valve or the heating water supply valve) according to the recognition results.

[0074] Furthermore, regarding the camera visual acquisition diagnosis subsystem, the camera is fixed on each station of the test bench through a movable bracket. Before starting the test, the user adjusts the camera lens to face the digital panel of each gas heating furnace through the movable bracket. Each camera is powered by POE, that is, power supply and data transmission are achieved through a network cable. The network cable of each camera is converged to the POE switch, and the network cable of the industrial computer is also converged to the POE switch, so that the industrial computer can access each camera. The image recognition program is run in the industrial computer for real-time recognition, and the recognition result is sent to the durability test application through shared memory or internal network communication. The durability test application takes measures according to the identification code. The identification code is "E1", "E2", "E6", "E9", "EU", "E7", and "EH", indicating a locked state. The test system needs to close the pneumatic valve of the domestic water supply branch and the pneumatic valve of the heating water supply branch, that is, suspend the start-stop cycle mode of the domestic hot water function or suspend the start-stop cycle mode of the heating function. The fault code is "ED". The test system not only needs to test the pneumatic valves of the domestic water branch and the heating water supply branch, but also needs to close the pneumatic valve of the gas branch, that is, stop the start-stop cycle mode of the domestic hot water function or suspend the start-stop cycle mode of the heating function. The durability performance test of the gas heating furnace is stopped and no longer waits for recovery.

[0075] The key process of image recognition of digital panels is as follows:

[0076] Since the digital number is a 7-segment digital number, (1) first define a character definition table for the 7-segment digital number; (2) obtain the original image; (3) use canny edge detection to obtain the digital panel area, namely ROI; (4) convert the color image of the ROI area to a grayscale image; (5) perform Gaussian blur on the ROI grayscale image; (6) binarize the ROI area image; (7) perform morphological opening operation; (8) perform contour filtering; (9) perform image segmentation; (10) extract the character ROI; (11) define the character ROI as divided into 7 segments; (12) calculate the width and height of each segment; (13) determine whether the non-zero pixels in each segment are greater than the total area of ​​the segment. The product is 50%, if so, jump to (14), otherwise jump to (15); (14) set 1 for the segment whose non-zero pixels are greater than 50% of the total area of ​​the segment; (15) set 0 for the segment whose non-zero pixels are less than 50% of the total area of ​​the segment; (16) write the segments set to 1 or 0 into the character list; (17) determine whether the 7 segments have been processed, if so, jump to step (18), if not, jump to step (10); (18) search for the character in the character lookup table; (19) record the recognized character; (20) determine whether the 2 characters have been processed, if so, jump to step (2), if not, jump to step (10).

[0077] Through the above process, the digital numbers on the digital panel of the gas heating furnace can be identified in real time and accurately, providing a basis for processing measures for the durability test application, thereby achieving full automation and releasing manpower.

[0078] Furthermore, the camera visual acquisition diagnosis subsystem supports the automation of the durability test system, but it is unable to detect abnormal sounds with low frequency and small sound amplitude when the burners, water pumps, fans, proportional valves and other components are working at an early stage during the test. The present invention adopts a pickup module designed with MEMS microphone as the core. Before the test, the user magnetically adsorbs each pickup module on the gas heating furnace, and the wired cable led out of the pickup module is connected to the acquisition gateway. The acquisition gateway realizes analog-to-digital conversion and transmits real-time audio information to the industrial computer through the network. The voiceprint diagnosis algorithm runs in the industrial computer, and the results of the auscultation diagnosis are sent to the durability test application through shared memory or internal network communication. The durability test application records the auscultation results and pauses or stops the test mode according to the degree;

[0079] The MEMS microphone voiceprint acquisition and diagnosis subsystem also includes major modules such as sound detection, sound preprocessing, acoustic feature extraction, voiceprint feature extraction model and feature similarity matching. In terms of workflow, the voiceprint acquisition and diagnosis subsystem can be divided into training, registration and reasoning stages. Since there is no public voiceprint library for use in the gas heating furnace industry, the present invention adopts a positive sample learning method to collect sounds from gas heating furnaces that are working normally, and extracts features from the collected sounds and registers them into the library. In the voiceprint recognition and reasoning stage, if the recognized voiceprint is not similar to the normal voiceprint in the library, it is marked as an abnormality, and the original sound is saved. Later, a senior expert can manually play back this sound and mark it into the library. When the voiceprint acquisition and diagnosis subsystem recognizes this type of sound again in the future, the system automatically displays it as the abnormal type marked by the expert at the beginning.

[0080] The key process of voiceprint recognition reasoning is as follows:

[0081] The first is (1) picking up the sound through the MEMS microphone; (2) decomposing the sound signal into multiple layers and obtaining the high-frequency signal coefficients of each layer; (3) sorting and evaluating the coefficients of each layer, and calculating the threshold to remove noise; (4) converting the denoised sound signal into a spectrum; (5) using CNN one-dimensional convolution to extract low-level features and reduce the spatial dimension; (6) calculating the similarity using the dot product method and converting the dot product result into self-attention (SELF-ATTENTIVE) weights through the softmax function; (7) concatenating the outputs of (5) and (6); softmax classification; (8) inputting the speech features of the database into the similarity matching module to calculate the similarity score, and the system determines the result based on the output similarity score and the predefined threshold.

[0082] The domestic water inlet pipeline designed by the present invention is designed with five branches, of which three branches are designed to meet the tests of high pressure, normal pressure, low pressure and high flow, normal flow and low flow. A booster pump is set on the high-pressure branch to solve the problem of insufficient water pressure in the factory water circuit; a water pressure regulating valve is set on the normal pressure and low pressure branches; a water flow proportional valve is set on high flow, normal flow and low flow. The design of the three branches not only meets the requirements of the endurance test for pressure and flow, but also after the pressure and flow are set, there is no need for repeated adjustment during the cyclic traversal test, as long as the on and off of the three branches are controlled. This method is efficient and accurate.

[0083] The heating outlet water and heating return water pipelines are interconnected through a plate exchanger, and an expansion water tank is set on the heating return water pipeline to solve the problem that during the durability test system, the expansion water tank can release extra water due to contraction caused by temperature drop, and absorb extra water due to expansion caused by temperature rise, thereby maintaining the stability of the system, avoiding excessive pressure fluctuations in the system, and preventing system damage.

[0084] Design an electronic control system that combines an industrial computer and a PLC to collect and control components such as the water pressure regulating valve, pneumatic valve, digital pressure gauge, digital flow meter, water flow proportional valve, temperature sensor, booster pump, plate converter, etc. installed on the test pipeline, to achieve on-off, pressure, flow regulation of the heating furnace's gas, domestic water, heating outlet water, and heating return water, as well as monitoring of temperature, pressure, and flow, to meet the requirements of durability testing and achieve the level of automated testing.

[0085] During the durability test of the heating furnace, various faults may occur in the heating furnace, such as water pressure fault, accidental flameout, heating return water sensor fault, wind pressure switch disconnection or closure, over-temperature fault, bathroom water outlet sensor fault, antifreeze fault, communication fault, abnormal temperature rise, temperature unchanged fault, etc. After these faults occur, the corresponding fault code will be displayed on the digital display panel of the heating furnace. When the fault code is manually observed, the processing manual of the fault code is searched, and the test button of the durability test bench is manually clicked, such as closing the bathroom valve, plate change valve or gas valve, etc., which requires manual 24-hour duty, which is time-consuming and labor-intensive. The present invention installs a camera on each station of the test bench with a movable bracket, and before the test, the camera lens is directed to the digital display panel of the heating furnace to shoot and obtain a video stream, and the algorithm is run in the industrial computer to identify the error code in real time, and the identified error code is sent to the electronic control system software. The electronic control system software automatically looks up the table according to the error code, and automatically performs corresponding actions on the bathroom valve, plate change valve or gas valve of the test pipeline. This method has a high degree of automation and does not require human 24-hour duty.

[0086] During the durability test of the heating furnace, in the early stage of the failure, the burner, water pump, fan, proportional valve and other components will occasionally make abnormal sounds when they work abnormally. This phenomenon has a low frequency and a small sound amplitude, and it is very easy to be detected by humans. The present invention uses a MEMS microphone placed in an aluminum alloy shell and magnetically adsorbed on the heating furnace shell to collect the sound of the heating furnace in real time. The amplified and filtered signal of the MEMS microphone is transmitted to the industrial computer through a shielded cable. The industrial computer acquisition card converts its analog-to-digital conversion and analyzes and judges it in real time through the built-in voiceprint model. This method can detect heating furnace failures very early and is used to evaluate the quality of the heating furnace.

[0087] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A gas heating furnace durability performance test system with audio-visual collaborative diagnosis, characterized in that: Included are: System test pipeline, the system test pipeline includes a gas test pipeline, a domestic water pipeline, a heating water supply branch, a heating return water branch and a bathroom water outlet branch. Each pipeline is connected to the gas heating furnace through a corresponding interface to realize the gas supply, water circulation and drainage functions, and the pipeline is provided with instruments and valves for monitoring and regulating fluid parameters; The electrical control system collects and controls the digital flow meter, temperature sensor, pressure sensor, pneumatic valve, gas pressure regulating valve, and water flow proportional valve installed on the test pipeline of the system through PLC. The PLC communicates data with the industrial computer through Ethernet. The industrial computer runs a program for data processing and system control. The camera visual acquisition diagnosis subsystem includes a camera installed on the test bench station with its lens facing the digital panel of the gas heating furnace. The camera is powered by POE and connected to the industrial computer through the network. The industrial computer runs an image recognition program to collect image information of the digital panel of the gas heating furnace and perform recognition, and send the recognition result to the durability test application program. The MEMS microphone voiceprint acquisition and diagnosis subsystem includes a sound pickup module with a MEMS microphone as the core, which is magnetically attached to a gas heating furnace. The sound pickup module is connected to an industrial computer through an acquisition gateway. The acquisition gateway realizes analog-to-digital conversion and transmits real-time audio information to the industrial computer. The voiceprint diagnosis algorithm runs in the industrial computer to collect sound information when the gas heating furnace is working and perform auscultation diagnosis, and send the diagnosis results to a durability test application. The voiceprint diagnosis algorithm includes sound detection, sound preprocessing, acoustic feature extraction, voiceprint feature extraction model and feature similarity matching module, and its workflow includes training, registration and reasoning stages.

2. The gas heating furnace durability performance test system with audio-visual collaborative diagnosis according to claim 1 is characterized by: The inlet of the gas test pipeline is connected to the factory gas port, and at least one gas branch is led out from the gas test pipeline. A manual ball valve, a gas pressure regulating valve, a gas pressure gauge, a gas flow meter, a pneumatic valve, a digital gas pressure gauge, and a gas temperature sensor are sequentially arranged on the gas branch. The gas branch outlet is connected to the gas inlet of the heating furnace. The gas pressure regulating valve adjusts the valve through a stepper motor, the stepper motor is connected to the motor driver, and the motor driver is connected to the PLC through a serial port.

3. The gas heating furnace durability performance test system with audio-visual collaborative diagnosis according to claim 2 is characterized by: The inlet of the domestic water pipeline is connected to the cold water outlet of the factory. At least one domestic water branch is led out from the domestic water pipeline. The domestic water branch provides cold water for the heating furnace. A manual ball valve and a filter are installed on the domestic water branch. Three pressure regulating branches are separated after the filter, namely high pressure branch, normal pressure branch and low pressure branch. The high pressure branch is installed with a pressure gauge, a pneumatic valve and a booster pump in sequence. The normal pressure branch and the low pressure branch are respectively installed with a pressure gauge, a pneumatic valve and a water pressure regulating valve in sequence. The three pressure regulating branches converge into the domestic water branch. A digital flow meter is installed on it, and the flow meter is divided into four flow branches, namely high flow, normal flow, low flow branch and heat flow exchange branch. Pneumatic valves and proportional valves are installed on the high flow, normal flow and low flow branches respectively, and then they are combined with the domestic water branch and then pressure sensors and temperature sensors are installed in sequence. The terminal outlet of the domestic water branch is connected to the domestic water inlet of the heating furnace, the pneumatic valve is installed on the heat flow exchange branch and then connected to the plate heat exchanger, and the drainage branch is led out from the plate heat exchanger and connected to the factory drain outlet.

4. The gas heating furnace durability performance test system with audio-visual collaborative diagnosis according to claim 3 is characterized by: The heating water supply branch is connected to the heating water supply outlet of the gas heating furnace. The heating water supply branch is sequentially provided with a temperature sensor, a water replenishment branch tee, a pressure relief test point, a drainage branch, a water flow proportional valve and a pneumatic valve. The end of the heating water supply branch is connected to the heating water outlet of the plate heat exchanger. The heating return water branch is led out from the heating return water outlet of the plate heat exchanger, and is sequentially provided with a water blowing branch tee, an expansion water tank, a digital flow meter, a pressure sensor, and a temperature sensor. A compressed air water blowing branch is led out from the tee of the water branch and a pneumatic valve is installed, a pressure relief point is led out next to the expansion water tank connection port, and the end of the heating return water branch is connected to the heating return water outlet of the gas heating furnace; the bathroom water outlet branch is connected to the bathroom water outlet of the gas heating furnace, and the bathroom water outlet branch is successively installed with a temperature sensor, a pressure gauge, a pneumatic valve, and a drainage tee, and the end of the bathroom water outlet branch is connected to the factory drain, and the drainage tee is connected to the drainage branch.

5. A gas heating furnace durability performance test system with audio-visual collaborative diagnosis according to any one of claims 1 to 4, characterized in that: In the electrical control system, the program running on the industrial computer includes a human-computer interaction interface program, a database and a model. The human-computer interaction interface program is used to realize the interactive operation between the user and the system, the database is used to store the collected data and diagnostic results, and the model is used to analyze and process the collected data.

6. The gas heating furnace durability performance test system with audio-visual collaborative diagnosis according to claim 5 is characterized by: In the camera visual acquisition diagnosis subsystem, the camera is a high-definition camera that can collect high-definition image information of the digital panel of the gas heating furnace in real time. The image recognition program can recognize the display content on the digital panel, including fault codes, operating status parameters, etc., and classify and store the recognition results.

7. The gas heating furnace durability performance test system with audio-visual collaborative diagnosis according to claim 6 is characterized by: In the MEMS microphone voiceprint acquisition diagnosis subsystem, the MEMS microphone has high sensitivity and low noise characteristics, and can accurately collect sound information when the gas heating furnace is working. The voiceprint diagnosis algorithm can extract and analyze the features of the collected sound information, determine whether there is abnormal sound in the gas heating furnace, and record and alarm the diagnosis results.

8. A method for testing the durability of a gas heating furnace using audio and visual collaborative diagnosis, characterized in that: The method is applied to the gas heating furnace durability performance test system of audio-visual collaborative diagnosis as described in any one of claims 1 to 7, and the durability test application running in the industrial computer supports three modes, namely, the start-stop cycle mode of the domestic hot water function, the start-stop cycle mode of the heating function, and the start-stop cycle mode of the domestic hot water function and the heating function conversion, wherein the start-stop cycle mode of the domestic hot water function includes the following steps: Gas preparation steps: manually open the gas inlet ball valve, programmatically open the gas branch and monitor the gas flow and gas temperature; Test steps for the first cycle (N1): manually open the domestic water inlet ball valve, the program starts counting, the program opens the high-pressure branch, closes the normal pressure and low-pressure branches, the program opens the high-flow branch, closes the normal flow and low-flow branches and the heat flow exchange branch, opens the bathroom outlet branch, monitors the flow, pressure, temperature of the domestic water branch and the temperature of the bathroom outlet branch, the gas heating furnace starts the domestic hot water function, burns for the set time, closes the bathroom outlet branch and stops for the set time, automatically opens the bathroom outlet branch, starts the gas heating furnace to burn for the set time, closes the bathroom outlet branch and stops for the set time, and the program counts and records; Subsequent cycle test steps (N2-N9): Perform subsequent cycle N2-N9 tests according to different set pressure and flow combinations, and perform tests under different pressures of domestic water, with different flow branches opened and other flow branches closed; Cycle cycle steps: In the start-stop cycle mode of the domestic hot water function, the 9 cycles N1-N9 are cycled; Exception handling steps: In the cycle mode of the start-stop cycle test, when the digital panel of the gas heating furnace has an abnormal situation and displays an error code, the camera visual acquisition diagnosis subsystem uses real-time video, and the image recognition program running in the industrial computer informs the durability test application of the recognition result. The durability test application takes the action of closing the gas valve, domestic water valve, or heating water supply valve based on the recognition result.

9. The method for testing the durability of a gas heating furnace using audio-visual collaborative diagnosis according to claim 8 is characterized in that: The image recognition process of the digital panel by the image recognition program specifically includes the following steps: Define a character definition table for 7-segment digital characters; Get the original image; The digital panel area, i.e. ROI, is obtained through canny edge detection; Convert the color image of ROI area to grayscale image; Perform Gaussian blur processing on the ROI grayscale image; Binarize the ROI area image; Perform morphological opening operation; Perform contour filtering; Perform image segmentation; Extract character ROI; The character ROI definition is divided into 7 segments; Calculate the width and height of each segment; Determine whether the non-zero pixels of each segment are greater than 50% of the total area of ​​the segment. If so, set the segment whose non-zero pixels are greater than 50% of the total area of ​​the segment to 1; otherwise, set the segment whose non-zero pixels are less than 50% of the total area of ​​the segment to 0; Write the segments set to 1 or 0 into the character list; Determine whether 7 segments have been processed. If so, search the character in the character lookup table, record the recognized character, and then determine whether 2 characters have been processed. If so, jump to the step of obtaining the original image. If not, jump to the step of extracting the character ROI. If not, jump to the step of extracting the character ROI.

10. The method for testing the durability of a gas heating furnace using audio and visual collaborative diagnosis according to claim 9 is characterized in that: During testing: The MEMS microphone voiceprint collection and diagnosis subsystem collects components including but not limited to the sounds of gas heating furnace burners, water pumps, fans, and proportional valves at work; auscultation diagnosis is performed through the voiceprint diagnosis algorithm running in the industrial computer, and the results of auscultation diagnosis are sent to the durability test application through shared memory or internal network communication; The endurance test application records the auscultation results and pauses or stops the test mode depending on the degree; The voiceprint recognition reasoning process includes: (1) Sound is picked up by a MEMS microphone; (2) Decomposing the sound signal into multiple layers and obtaining the high-frequency signal coefficients of each layer; (3) Sort and evaluate the coefficients of each layer and calculate the threshold to remove noise; (4) Convert the denoised sound signal into a spectrogram; (5) Use CNN one-dimensional convolution to extract low-level features and reduce spatial dimensions; (6) The dot product method is used to calculate the similarity, and the dot product result is converted into self-attention weight through the softmax function; (7) Connect the outputs of steps (5) and (6) together; softmax classification; (8) The speech features in the database are input into the similarity matching module to calculate the similarity score. The system determines the result based on the output similarity score and the predefined threshold.

Images