A fault pre-diagnosis system and method for a gas heating device

Abstract

This invention discloses a fault pre-diagnosis method and system for gas-fired heating equipment. The system collects equipment operating status data via a 4G module and pushes it to a big data platform via an MQTT server. The platform performs data standardization processing based on a physical model and stores the data in a time-series database. It then employs a fault prediction model that combines LSTM neural networks and random forests for intelligent analysis, enabling early fault warnings and automatic work order generation. The system supports remote control command issuance and dynamic adjustment of combustion parameters, and features altitude-adaptive compensation. A super table statistically analyzes regional fault distribution, providing data support for equipment optimization. This invention achieves real-time monitoring, intelligent diagnosis, and precise control of equipment status, significantly improving equipment reliability and reducing maintenance costs.

Description

Technical Field

[0001] This invention relates to the field of gas heating equipment technology, specifically to a fault pre-diagnosis system and method for gas heating equipment. Background Technology

[0002] Traditional after-sales service for gas heating equipment (such as wall-hung boilers and dual-purpose boilers for heating and hot water) heavily relies on reactive response and on-site diagnosis. The standard process is typically "equipment malfunction → user notices and reports repair → initial communication with customer service → arranging for an engineer to visit → on-site troubleshooting and problem location." This linear process has significant drawbacks. First, it suffers from high response delays and enormous time costs: from the occurrence of a malfunction to the engineer's arrival, there is an average waiting period of 24 to 72 hours. During this time, users may face the predicament of no heating or hot water, resulting in a very poor experience. Second, diagnosis and repair are inefficient: engineers have limited understanding of the malfunction before visiting and may need to make multiple visits due to a lack of suitable spare parts, increasing the time and manpower costs for both parties. Furthermore, the multiple rounds of communication between customer service and users, customer service and engineers, and engineers and users also consume considerable resources. This reactive and delayed maintenance model can no longer meet the modern user's demand for immediate and reliable service. Summary of the Invention

[0003] To address the low maintenance efficiency of existing gas-fired heating equipment, this application provides a fault pre-diagnosis system and method for gas-fired heating equipment. The specific technical solution of this application is as follows: A method for fault pre-diagnosis of gas heating equipment includes the following steps: A 4G module of the gas heating equipment sends equipment status commands to a controller based on a polling mechanism, causing the controller to control a sensor module to collect the current status information of the gas heating equipment and send it to the 4G module; the 4G module sends the received current status information to an MQTT server, which pushes the received current status information to a big data platform and simultaneously to the corresponding user terminal of the gas heating equipment; the big data platform performs data abstraction and data constraint processing on the current status information based on the physical model of the current gas heating equipment, and then writes the processed current status information into the corresponding time-series database; the big data platform performs fault analysis on the current status information in the time-series database based on a fault prediction model, and then generates a fault work order for the current gas heating equipment based on the analysis results and sends it to the corresponding user terminal.

[0004] Furthermore, the 4G module of the gas heating equipment sends equipment status commands to the controller based on a polling mechanism, enabling the controller to control the sensor module to collect the current status information of the gas heating equipment. This includes the following steps: the 4G module sends a setting status command to the controller at set intervals, causing the controller to send an information collection command to the sensor module; after receiving the information collection command, the sensor module collects sensor information from different sensors of the gas heating equipment; the sensor module arranges and summarizes the collected sensor information, and then directly sends the summarized sensor information to the 4G module; the sensor information includes gas valve pressure, heating outlet water temperature, heating return water temperature, domestic hot water temperature, burner temperature, flue gas temperature, ambient temperature, system water pressure, gas pressure, pressure difference before and after the water pump, hot water flow rate, heating water flow rate, gas flow rate, voltage, current, power, fan speed, ignition current, NTC resistance, sensing needle current, carbon monoxide concentration, and Hall sensor signal.

[0005] Furthermore, after receiving the current status information, the 4G module compares the currently received current status information with the sensor information in the current status information received in the previous polling mechanism. Based on the comparison results, the 4G module sends the changed sensor information in the currently received current status information to the MQTT server in the form of uplink attribute topics.

[0006] Furthermore, the big data platform writes the processed current state information into the corresponding time-series database, including the following steps: When the big data platform uses the object model to constrain the received changing data, it records the data in the changing data that exceeds the threshold of the object model constraint; the big data platform writes the changed data processed by the object model and the timestamp into the time-series database of the corresponding gas heating equipment; if a certain data in the current state information exceeds the threshold of the object model constraint N times consecutively, the big data platform will mark the data; where N is a natural number greater than 1.

[0007] Furthermore, the big data platform constructs a fault prediction model using LSTM neural networks and random forests, including the following steps: The big data platform processes sensor information through data cleaning, normalization / standardization, time window construction, and data partitioning to generate training, validation, and test sets of sensor information; The big data platform constructs and trains an LSTM model using the training, validation, and test sets of sensor information to obtain the LSTM's prediction values; Based on the LSTM's prediction values, the big data platform calculates the residuals, and then concatenates the original features and temporal features to generate the input feature set for the random forest; The big data platform uses the concatenated input feature set to train the random forest model for fault classification or regression and evaluates the importance of features; The big data platform combines the outputs of the LSTM model and the random forest for final prediction and uses accuracy and RMSE to evaluate model performance; The big data platform optimizes the fault prediction model and deploys it for online prediction through hyperparameter tuning, model compression, and real-time data processing.

[0008] Furthermore, the big data platform dynamically adjusts the combustion parameters of the gas heating equipment, including the following steps: the controller acquires the current location of the gas heating equipment and the gas pressure of that area, and then acquires the ion signal fed back by the ignition needle of the gas heating equipment; the controller detects the conductivity of the flame through the ion signal, and then obtains the combustion state through the current intensity of the flame ions; the controller dynamically adjusts the ratio of gas valve opening and fan speed according to the combustion state, and then superimposes altitude feedforward compensation on the ratio of gas valve opening and fan speed.

[0009] Furthermore, the big data platform statistically analyzes fault types based on the time-series database, including the following steps: The big data platform statistically analyzes the fault types of all gas heating equipment in the time-series database, and then obtains the model ratio and regional ratio of the fault types based on the model and region of the gas heating equipment.

[0010] Furthermore, after the big data platform obtains the model proportion and regional proportion of fault types, it will use a super table of the time series database to summarize the data of the same model of gas heating equipment; the data in the super table includes the province, city, district, equipment ID, equipment serial number, and whether the equipment is online.

[0011] Furthermore, the big data platform controls the gas heating equipment based on control commands sent by the user terminal, including the following steps: The big data platform receives the control commands sent by the user terminal and then writes the control commands into the time-series database of the corresponding gas heating equipment; The big data platform sends the control commands to the MQTT server in the form of a downlink attribute topic, so that the MQTT server pushes the control commands to the 4G module; The 4G module sends the control commands to the controller through the communication protocol, so that the controller adjusts the gas heating equipment according to the control commands.

[0012] A fault pre-diagnosis system for gas heating equipment is disclosed. The fault pre-diagnosis system performs fault pre-diagnosis using the aforementioned fault pre-diagnosis method for gas heating equipment. The fault pre-diagnosis system includes a big data platform, gas heating equipment, and an MQTT server. The gas heating equipment includes a controller, a sensor module, and a 4G module.

[0013] Compared with existing technologies, this application constructs a proactive pre-diagnosis system based on the Internet of Things and big data, which completely subverts the traditional after-sales maintenance model and achieves the following significant technical effects: Shifting from reactive to proactive, enabling early warning of faults: The core value of this application lies in its ability to continuously analyze equipment operating status through a fault prediction model, predicting potential faults up to 72 hours in advance. This shifts maintenance actions from "remedial action" to "pre-emptive intervention," allowing maintenance to begin before users notice problems or the equipment is shut down, greatly ensuring the continuity and comfort of user experience.

[0014] Significantly improves service efficiency and reduces overall costs: The system's automatically generated fault tickets already contain precise fault location information. This allows engineers to identify the problem and prepare the corresponding parts before visiting the site, achieving "one-stop solution" and significantly reducing the number of repeat visits due to unclear diagnosis or incomplete spare parts. At the same time, the automated information flow (from equipment to platform to engineer) reduces the cost of repeated communication by about 80% in the traditional model, improving the operational efficiency of the entire service chain.

[0015] Building a data-driven precision operation and maintenance system: Real-time collection and stable transmission of equipment status information are achieved through 4G modules and MQTT servers. The data is standardized and stored using object models and time-series databases, providing a high-quality data foundation for fault prediction. This not only serves single repairs but also continuously optimizes prediction algorithms through accumulated historical data, forming an intelligent operation and maintenance closed loop that becomes increasingly accurate with use. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating a method for fault pre-diagnosis of gas heating equipment in one embodiment of this application. Detailed Implementation

[0017] The embodiments of this application are described in detail below. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout.

[0018] In the description of this application, it should be noted that the directional terms such as "center", "lateral", "longitudinal", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", and "counterclockwise" indicate the orientation and positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. They should not be construed as limiting the specific protection scope of this application.

[0019] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features. Thus, the use of "first" or "second" to define a feature may explicitly or implicitly include one or more of that feature, and in the description of this application, "at least" means one or more, unless otherwise explicitly specified.

[0020] In this application, unless otherwise expressly specified and limited, the terms "assembly," "connection," and "joining" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can also refer to a mechanical connection; they can refer to a direct connection or a connection through an intermediate medium; or they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0021] In the application, unless otherwise specified and limited, "above" or "below" the second feature can include direct contact between the first and second features, or contact between the first and second features through another feature between them. Furthermore, "above," "below," and "over" the second feature includes the first feature being directly above or diagonally above the second feature, or simply indicating that the first feature is at a higher horizontal level than the second feature. "Above," "below," and "below" the second feature includes the first feature being directly below or diagonally below the second feature, or simply indicating that the first feature is at a lower horizontal level than the second feature.

[0022] The following description, in conjunction with the accompanying drawings, further illustrates specific embodiments of this application, making the technical solution and its beneficial effects clearer and more explicit. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, but should not be construed as limiting it.

[0023] like Figure 1As shown, a fault pre-diagnosis method for gas heating equipment includes the following steps: The 4G communication module of the gas heating equipment periodically sends equipment status query commands to the controller inside the equipment based on a polling mechanism. This causes the controller to control the sensor module to collect the current real-time operating status information of the gas heating equipment and send it to the 4G module through the internal communication interface. The 4G module adopts a differential data transmission mechanism. During the 2023 heating season test, the average monthly data traffic per unit decreased from 35MB to 12MB, a reduction of 65.7%. In specific implementation, the module has a built-in status buffer, and data is only uploaded when the temperature change is ≥1℃, the pressure change is ≥0.1bar, or the flow rate change is ≥0.5L / min. In the deployment of 500 units in Hebei Province, this mechanism effectively reduced network congestion, and the data transmission success rate reached 99.2%. The 4G module sends the received current status information to an MQTT server deployed in the cloud via the mobile network. The MQTT server, based on its publish / subscribe model, pushes the received current status information to the big data analysis platform in real time, and also pushes it to the user's mobile application corresponding to the gas heating equipment. The user interface is an app, serving as the user's interaction point with the wall-hung boiler. It primarily enables remote control of the boiler, such as power on / off, winter / summer switching, and temperature settings. Additionally, it can connect to a big data platform to query historical water, electricity, and gas usage data. The big data platform abstracts and constrains the received current state information based on a pre-defined physical model of the gas heating equipment. This standardized current state information is then persistently stored in a corresponding time-series database. The physical model filters the current state information, effectively reducing the amount of data that needs to be stored in the time-series database. Users can then use this time series data for statistical analysis. The big data platform performs multi-dimensional fault analysis on the current state information entering the time-series database using a pre-trained fault prediction model. Based on the analysis results, it automatically generates fault work orders for the current gas heating equipment and sends them to the corresponding user terminal via push notifications, providing early warning of faults. Since the equipment's attributes change in real time, the big data platform uses the domestically developed TDEngine as the time-series database to store historical data.

[0024] The big data platform uses the TensorFlow framework to build an LSTM-random forest hybrid model. Training data comes from over 8 million historical operation records across 20,000 devices nationwide. Model input features include: Temporal characteristics: temperature change gradient and pressure fluctuation frequency over the past 2 hours.

[0025] Static characteristics: equipment installation altitude, regional average temperature.

[0026] Real-time characteristics: ion current intensity, fan PWM duty cycle.

[0027] In tests conducted in Northeast China (ambient temperature -25℃), the model achieved a prediction accuracy of 87.3% for E01 ignition failure faults. Feature importance analysis showed that gas pressure (weight 0.32), ignition current (weight 0.28), and ambient temperature (weight 0.18) were the key influencing factors.

[0028] In one embodiment, the 4G module of the gas heating equipment sends equipment status commands to the controller based on a polling mechanism. This causes the controller to control the sensor module to collect the current status information of the gas heating equipment. The specific steps include: the 4G module automatically sends equipment status query commands to the controller at set time intervals (e.g., every 24 hours). Upon receiving the equipment status command, the controller immediately sends information collection commands to various sensors in the sensor module. The sensors collect key operating parameters of the gas heating equipment as current status information. The sensor module then sorts and summarizes the collected sensor information according to a user-defined order and sends this summarized sensor information directly to the 4G module, eliminating the need for the controller to process the data and preventing controller crashes. The collected sensor information comprehensively covers the equipment's operating status, including key parameters such as gas valve pressure, heating outlet water temperature, heating return water temperature, domestic hot water temperature, burner temperature, flue gas temperature, ambient temperature, system water pressure, gas pressure, pressure difference before and after the water pump, hot water flow rate, heating water flow rate, gas flow rate, power supply voltage, operating current, real-time power, fan speed, ignition current, NTC temperature sensor resistance, flame sensing needle current, carbon monoxide concentration, and Hall sensor signal. Based on this sensor information, 22 fault types and their corresponding codes can be obtained: bathroom sensor fault 0x11, heating sensor fault 0x12, over-temperature fault 0x13, excessively rapid temperature rise protection 0x14, ignition failure 0x16, flue fault 0x17, water circuit fault 0x18, residual flame fault 0x19, valve body drive fault 0x1A, water pressure protection 0x1B, extreme temperature protection 0x1C, freezing protection 0x1D, excessively slow temperature rise protection 0x1E, temperature not changing protection 0x1F, accidental flameout protection 0x20, abnormal heating temperature protection 0x21, abnormal bathroom temperature protection 0x22, poor condensate drainage protection 0x25, abnormal fan drive protection 0x30, abnormal heating flow protection 0x31, excessive flue gas temperature protection 0x32, and flue gas temperature sensor fault 0x33.

[0029] In one embodiment, after receiving the current status information, the 4G module performs data change detection: it compares the currently received status information with the sensor information in the previous polling mechanism's status information item by item. Based on the comparison results, the 4G module adopts a data compression transmission strategy, sending only the changed data items in the currently received status information to the MQTT server as uplink attribute topics, effectively reducing network traffic consumption. By comparing the previous device status, the 4G module detects changed attributes and publishes MQTT uplink topics. By comparing statuses, uplink data traffic can be reduced, alleviating the transmission and storage pressure on the server. When the sensor module is not in the information collection state of the polling mechanism, if some sensors detect changes in sensor information, they will actively send the sensor information to the 4G module, causing the 4G module to actively send the changed sensor information to the big data platform. The big data platform then records the data and sends it to the user. For example, a pressure sensor before the valve detects the gas pressure on the gas valve. If this pressure value decreases rapidly, it indicates a gas leak or no gas in the gas pipe. If the gas heating equipment fails to ignite, it's not a problem with the equipment itself, but rather a lack of gas. The big data platform can notify the user to check for gas problems. A carbon monoxide monitoring sensor detects the concentration of carbon monoxide in the environment. When a high concentration is detected, it alerts the user to backflow or diffusion of exhaust gas into the room, effectively protecting personal and property safety.

[0030] As one embodiment, the big data platform writes the processed current status information into the corresponding time-series database, including the following detailed steps: When the big data platform performs data constraint verification on the received change data using the parameter specifications defined in the object model, it specifically records data points in the change data that exceed the threshold range of the object model constraints. The big data platform writes the change data, after being verified by the object model, along with a precise timestamp, into the time-series database of the corresponding gas heating equipment. If a specific data point in the current status information exceeds the threshold of the object model constraints N times consecutively (e.g., 3 times consecutively), the big data platform will mark the data as an anomaly and designate it as a key monitoring target.

[0031] As one embodiment, the big data platform employs a combination of LSTM neural networks and random forest algorithms to construct a high-precision fault prediction model, including the following training steps: First, the big data platform preprocesses historical sensor information through data cleaning, normalization / standardization, time window construction, and standard data partitioning procedures to generate training, validation, and test sets of sensor information. The big data platform uses the training, validation, and test sets of sensor information to construct and train an LSTM time-series model to obtain LSTM prediction values. Based on the LSTM's predicted output values, the big data platform calculates the residuals between the predicted and actual values, and then concatenates the original features and extracted time features to generate the input feature set for the random forest model. The big data platform uses the concatenated input feature set to train the random forest model for fault classification or regression prediction and evaluates the importance of each feature. The big data platform combines the outputs of the LSTM model and the random forest model to make the final prediction decision and uses metrics such as accuracy and root mean square error (RMSE) to evaluate model performance. The big data platform continuously improves the accuracy of the fault prediction model through hyperparameter tuning, model compression, and real-time data processing optimization, and finally deploys the fault prediction model for online prediction.

[0032] Fault prediction example: The pressure before the gas valve is within the normal range of 1000-4000pa (when a fault is reported, if this sensor is not within this range, it is judged as an abnormality in external gas, that is, there is no gas in the gas pipeline or there is a gas leak).

[0033] Gas flow rate is greater than 0.5 m³ / h (if the flow rate is less than this when the fault is reported, it is determined that the main board is damaged).

[0034] Check if the temperature has changed (rise > 2°C) (if the temperature does not rise when the fault is reported, it is determined that the gas is not ignited and there is a problem with the gas valve).

[0035] Motherboard voltage range (175-280V) (If the fault is outside this range, it indicates an external power supply problem).

[0036] As one embodiment, the big data platform dynamically adjusts the combustion parameters of gas heating equipment based on environmental factors, including the following intelligent control steps: The controller first acquires the location information of the gas heating equipment and the real-time gas pressure data of that area, and then simultaneously acquires the ion current signal fed back by the ignition needle of the gas heating equipment. The controller accurately detects the flame conductivity through the ion current signal, and then evaluates the real-time combustion state by analyzing the current intensity of the flame ions. Based on the combustion state evaluation results, the controller dynamically adjusts the optimal ratio of gas valve opening and fan speed, and then superimposes feedforward compensation based on altitude calculation on the ratio of gas valve opening and fan speed to ensure that optimal combustion efficiency is maintained in different altitude areas. The controller controls the opening of the gas valve through a stepper motor, thereby controlling the combustion parameters of the gas heating equipment. Electro-ion combustion control technology is a technology that uses the ion current generated during combustion to monitor and control the combustion state. Its core principle is: during combustion, hydrocarbons ionize at high temperatures, generating positive ions and free electrons. These charged particles form a weak current (ion current) under the action of an external electric field. By detecting the intensity and changes of this current, it is possible to determine in real time whether combustion has occurred, the intensity of combustion, and the stability of combustion, thereby achieving precise control of the combustion process.

[0037] In actual deployment in Lijiang, Yunnan (altitude 2400m), the control system detected a decrease in flame stability (ion current dropped from the normal value of 3.2μA to 1.8μA) via ionization signals. The controller automatically executed the following adjustment sequence: The gas valve opening was adjusted from 65% to 58%. The fan speed was increased from 2800 rpm to 3200 rpm; The superimposed altitude compensation coefficient is 0.89; After adjustment, the combustion efficiency recovered from 84.3% to 91.7%, and the CO emission concentration decreased from 180ppm to 85ppm, meeting the national first-class energy efficiency standard.

[0038] As one embodiment, the big data platform statistically analyzes the distribution of fault types based on historical data in the time series database, including the following analysis steps: The big data platform statistically analyzes the fault types recorded by all gas heating equipment in the time series database, and then performs model proportion statistics and regional proportion analysis of fault types based on the model and regional information of the gas heating equipment to identify common problems and regional characteristic problems.

[0039] Analyzing data from the winter of 2023 using HyperTable revealed an unusually high failure rate for a certain model of equipment in Shanxi Province (12.3% vs. the national average of 5.7%). Further analysis showed: Taiyuan area: Water pressure failures accounted for 41% (caused by hard water quality); Datong area: Ignition failure rate 38% (difficulty starting at low temperatures); Jincheng area: Sensor failure rate is 21% (due to dust environment). Based on this analysis, the manufacturer made targeted improvements: adding water softeners in Taiyuan, optimizing the ignition program in Datong, and upgrading the sensor protection level in Jincheng, reducing the failure rate of this model in Shanxi to 6.8% in the next heating season.

[0040] As one implementation, after obtaining the model and regional proportions of fault types, the big data platform uses the super table function of a time-series database to efficiently aggregate data on gas heating equipment of the same model. The metadata fields in the super table include key information such as the province, city, region, unique ID, serial number, and real-time online status of the equipment. The big data platform can effectively observe the current status of gas heating equipment in different regions using the super table.

[0041] The metadata fields contained in the super table are as follows: As one embodiment, the big data platform supports remote control of gas heating equipment based on control commands sent by the user terminal, including the following control process: The big data platform receives the control commands sent by the user terminal, and then writes the control commands and related parameters into the control instruction area of ​​the time-series database of the corresponding gas heating equipment. The big data platform sends the control commands to the MQTT server in the form of a downlink attribute topic, causing the MQTT server to push the control commands to the 4G module of the target device. After receiving the control commands, the 4G module forwards the control commands to the device controller via Modbus or a custom communication protocol, enabling the controller to precisely adjust the operating parameters of the gas heating equipment according to the control commands.

[0042] The specific process of users sending temperature setting commands via the APP is as follows: The user adjusts the heating temperature from 20℃ to 23℃ in the APP; the big data platform records the operation log and generates a control command JSON packet; the MQTT server issues the command through the "device / BN2023-001 / control" topic; the 4G module receives and parses the command within 3 seconds and sends it to the controller through the RS485 interface; the controller adjusts the gas valve and water pump frequency to achieve a stable temperature transition within 8 minutes; tests show that the average remote control response delay is <5 seconds, and the command execution success rate is 99.5%.

[0043] A fault pre-diagnosis system for gas heating equipment is disclosed. This system performs intelligent fault pre-diagnosis using the aforementioned fault pre-diagnosis methods for gas heating equipment. The system includes a big data analysis platform, a group of gas heating equipment, and an MQTT message server. Each gas heating equipment comprises a core controller, sensor modules, and a 4G module, forming a complete IoT monitoring and early warning system. The 4G module supports three-network communication (China Mobile / China Telecom / China Unicom networks), with data encryption conforming to GB / T 37092-2018 standard, and user privacy protection certified by ISO / IEC 27001. The controller acts as the brain of the gas heating equipment, responsible for sensor data acquisition and operational control. The control board has an interface for communication with the 4G module, facilitating data upload and download after 4G module integration. The MQTT server implements a lightweight communication mechanism based on topic subscription. Devices subscribe to their relevant uplink and downlink data topics. The app and big data platform also act as clients, implementing data subscription and publishing functions. Access control logic between clients ensures topic-specific permission isolation, preventing data leakage. The big data platform manages users and devices, collects data from users and devices, performs subsequent analysis and statistics, and displays device status information from various dimensions.

[0044] Sensor modules include, but are not limited to, sensors used to detect signals such as gas valve pressure, heating outlet water temperature, heating return water temperature, domestic hot water temperature, burner temperature, flue gas temperature, ambient temperature, system water pressure, gas pressure, pressure difference across the water pump, hot water flow rate, heating water flow rate, gas flow rate, voltage, current, power, fan speed, ignition current, NTC resistance, sensing needle current, carbon monoxide concentration, and Hall sensor signals. For example, the SFD2 series vortex flow sensor is used to detect water flow. The SFD2 series is a water flow sensor based on the von Kármán vortex street principle. Under certain conditions, when a fluid of a certain volume flow rate flows past a turbulence column, double-row linear vortices with opposite rotation directions and regular arrangement will periodically form on both sides of the turbulence column. After nonlinear interaction, a von Kármán vortex street is formed. The frequency of vortex formation is proportional to the flow velocity. The sensor detects these vortices, thereby measuring the flow rate of the fluid in the pipe. The SR-456 dual-gas module is used for detecting carbon monoxide concentration. Utilizing an advanced semiconductor gas sensor and an electrochemical carbon monoxide sensor, it can simultaneously detect carbon monoxide produced by incomplete combustion of combustible gases, as well as potentially leaking natural gas or liquefied petroleum gas. It exhibits excellent consistency and extremely high sensitivity and stability. The SW15A series pressure sensors are used for detecting the temperature and pressure of water in pipelines. The SW15A high-precision silicon ceramic series are piezoresistive silicon pressure sensors, providing digital outputs of pressure readings over specified full-scale and temperature ranges. The SW15A series is fully calibrated and temperature-compensated for sensor offset, sensitivity, temperature effects, and nonlinearity using an onboard application-specific integrated circuit (ASIC). The SW15A pressure sensors are suitable for non-corrosive, non-ionic gases (such as air and other dry gases). Available options extend the performance of these sensors to include non-corrosive, non-ionic liquids.

[0045] Compared with existing technologies, this application constructs a proactive pre-diagnosis system based on the Internet of Things and big data, which completely subverts the traditional after-sales maintenance model and achieves the following significant technical effects: Shifting from reactive to proactive, enabling early warning of faults: The core value of this application lies in its ability to continuously analyze equipment operating status through a fault prediction model, predicting potential faults up to 72 hours in advance. This shifts maintenance actions from "remedial action" to "pre-emptive intervention," allowing maintenance to begin before users notice problems or the equipment is shut down, greatly ensuring the continuity and comfort of user experience.

[0046] Significantly improves service efficiency and reduces overall costs: The system's automatically generated fault tickets already contain precise fault location information. This allows engineers to identify the problem and prepare the corresponding parts before visiting the site, achieving "one-stop solution" and significantly reducing the number of repeat visits due to unclear diagnosis or incomplete spare parts. At the same time, the automated information flow (from equipment to platform to engineer) reduces the cost of repeated communication by about 80% in the traditional model, improving the operational efficiency of the entire service chain.

[0047] Building a data-driven precision operation and maintenance system: Real-time collection and stable transmission of equipment status information are achieved through 4G modules and MQTT servers. The data is standardized and stored using object models and time-series databases, providing a high-quality data foundation for fault prediction. This not only serves single repairs but also continuously optimizes prediction algorithms through accumulated historical data, forming an intelligent operation and maintenance closed loop that becomes increasingly accurate with use.

[0048] In the description of this specification, the terms "in one embodiment," "preferred," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this application. The illustrative expressions of the above terms in this specification do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described can be combined in any suitable manner in one or more embodiments or examples. The connection methods linked in the description of this specification have significant effects and practical utility.

[0049] Based on the above description of the structure and principles, those skilled in the art should understand that this application is not limited to the specific embodiments described above. Improvements and substitutions made using techniques known in the art based on this application all fall within the protection scope of this application and should be defined by the claims.

Claims

1. A method for pre-diagnosing faults in a gas-fired heating system, characterized in that, The method includes the following steps: The 4G module of the gas heating equipment sends equipment status commands to the controller based on a polling mechanism, which in turn controls the sensor module to collect the current status information of the gas heating equipment and send it to the 4G module. The 4G module sends the received current status information to the MQTT server, which then pushes the received current status information to the big data platform and simultaneously to the user terminal corresponding to the gas heating equipment. The big data platform performs data abstraction and data constraint processing on the current state information based on the physical model of the current gas heating equipment, and then writes the processed current state information into the corresponding time series database. The big data platform performs fault analysis on the current status information entering the time series database based on the fault prediction model, and then generates a fault work order for the current gas heating equipment based on the analysis results and sends it to the corresponding user terminal. The big data platform statistically analyzes fault types based on time-series databases, including the following steps: The big data platform statistically analyzes the fault types of all gas heating equipment in the time-series database, and then obtains the model ratio and regional ratio of the fault types based on the model and region of the gas heating equipment. After the big data platform obtains the model proportion and regional proportion of fault types, it will use the super table of the time series database to summarize the data of the same model of gas heating equipment. The data in the super table includes the province, city, district, device ID, device serial number, and whether the device is online.

2. The fault pre-diagnosis method for gas heating equipment according to claim 1, characterized in that, The 4G module of the gas heating equipment sends equipment status commands to the controller based on a polling mechanism, enabling the controller to control the sensor module to collect the current status information of the gas heating equipment, including the following steps: The 4G module sends a setting status command to the controller at set intervals, which in turn sends an information acquisition command to the sensor module. After receiving the information acquisition command, the sensor module collects sensor information from the gas heating equipment through different sensors in the sensor module. The sensor module will arrange and summarize the collected sensor information, and then send the summarized sensor information directly to the 4G module. The sensor information includes gas valve pressure, heating outlet water temperature, heating return water temperature, domestic hot water temperature, burner temperature, flue gas temperature, ambient temperature, system water pressure, gas pressure, pressure difference before and after the water pump, hot water flow rate, heating water flow rate, gas flow rate, voltage, current, power, fan speed, ignition current, NTC resistance, sensing needle current, carbon monoxide concentration, and Hall sensor signal.

3. The fault pre-diagnosis method for gas heating equipment according to claim 2, characterized in that, After receiving the current status information, the 4G module compares the currently received current status information with the sensor information in the current status information received by the previous polling mechanism. Based on the comparison results, the 4G module sends the changed sensor information in the currently received current status information to the MQTT server in the form of uplink attribute topics.

4. The fault pre-diagnosis method for gas heating equipment according to claim 3, characterized in that, The big data platform writes the processed current status information into the corresponding time-series database, including the following steps: When a big data platform uses a physical model to constrain the received changing data, it records the data in the changing data that exceeds the threshold of the physical model constraint. The big data platform writes the changed data and timestamps after processing by the physical model into the time series database of the corresponding gas heating equipment; If a certain data in the current state information exceeds the threshold of the object model constraint N times consecutively, the big data platform will mark the data. Where N is a natural number greater than 1.

5. The fault pre-diagnosis method for gas heating equipment according to claim 4, characterized in that, The big data platform constructs a fault prediction model using LSTM neural networks and random forests, including the following steps: The big data platform processes sensor information through data cleaning, normalization / standardization, time window construction, and data partitioning to generate training sets, validation sets, and test sets of sensor information. The big data platform constructs and trains an LSTM model using the training, validation, and test sets of sensor information to obtain the LSTM's prediction values. The big data platform calculates residuals based on LSTM predictions, then concatenates the original features and temporal features to generate the input feature set for a random forest. Big data platforms use spliced ​​input feature sets to train random forest models for fault classification or regression and to evaluate feature importance. The big data platform combines the outputs of the LSTM model and the random forest to make the final prediction, and uses accuracy and RMSE to evaluate the model performance. The big data platform optimizes the fault prediction model and deploys it for online prediction through hyperparameter tuning, model compression, and real-time data processing.

6. The fault pre-diagnosis method for gas heating equipment according to claim 1, characterized in that, The big data platform dynamically adjusts the combustion parameters of gas heating equipment, including the following steps: The controller obtains the current location of the gas heating equipment and the gas pressure in that area, and then obtains the ion signal fed back by the ignition needle of the gas heating equipment; The controller detects the conductivity of the flame through ion signals, and then obtains the combustion state through the current intensity of the flame ions; The controller dynamically adjusts the ratio of gas valve opening and fan speed according to the combustion status, and then adds altitude feedforward compensation to the ratio of gas valve opening and fan speed.

7. The fault pre-diagnosis method for gas heating equipment according to claim 1, characterized in that, The big data platform controls the gas heating equipment based on control commands sent by the user, including the following steps: The big data platform receives control commands sent by the user terminal and then writes the control commands into the time-series database of the corresponding gas heating equipment. The big data platform sends control commands to the MQTT server in the form of downlink attribute topics, so that the MQTT server pushes the control commands to the 4G module; The 4G module sends control commands to the controller via a communication protocol, enabling the controller to adjust the gas heating equipment according to the control commands.

8. A fault pre-diagnosis system for gas heating equipment, characterized in that, The fault pre-diagnosis system performs fault pre-diagnosis using the fault pre-diagnosis method for gas heating equipment as described in any one of claims 1 to 7. The fault pre-diagnosis system includes a big data platform, gas heating equipment, and an MQTT server. The gas heating equipment includes a controller, a sensor module, and a 4G module.

Images