A fault analysis method for a heating pipe network in a limited space scenario and a related device

Abstract

The application provides a heat supply pipe network limited space scene fault analysis method and related device, comprising: preprocessing original time sequence data to obtain a standardized multi-parameter time sequence matrix; inputting an LSTM-Attention model to obtain a fault mode probability distribution; extracting device surface temperature, convective heat transfer coefficient and environmental temperature as boundary conditions, and inputting a physical mechanism driven model to obtain a temperature deviation value; inputting the probability distribution, the temperature deviation value and the original time sequence data into a D-S evidence theory model to obtain a comprehensive fault probability and confidence. The application realizes early accurate identification and reliability improvement of heat supply limited space faults through data-driven and physical mechanism fusion diagnosis, combined with adaptive wireless communication and edge computing, and is suitable for limited space scenes such as heat wells and inspection wells.

Description

Technical Field

[0001] This invention belongs to the field of heating, specifically relating to a fault analysis method and related devices for heating pipeline networks in confined space scenarios. Background Technology

[0002] Heating networks are a crucial component of urban energy supply systems, bearing the core responsibility of centralized heating during winter. Their operational stability directly impacts residents' quality of life and public safety. Within heating systems, confined spaces such as heating wells and inspection wells, due to their strong enclosure, poor ventilation, small size, high humidity, and tendency to accumulate toxic and harmful gases (such as carbon monoxide (CO), hydrogen sulfide (H2S), and methane (CH4), become weak links and high-risk areas for malfunctions. These environmental factors not only accelerate the corrosion and aging of pipes, valves, and auxiliary equipment but also pose a serious threat to the personal safety of maintenance personnel.

[0003] For a long time, the operation monitoring of heating pipe networks has mainly relied on regular manual inspections. This method has many drawbacks: the inspection cycle is long, making continuous monitoring difficult; the detection methods are limited, usually relying only on portable gas detectors or visual judgment, lacking systematic data support; the response is delayed, and faults are often only discovered after leaks, pipe bursts, or poisoning incidents have occurred; at the same time, manual entry into confined spaces is extremely risky and can easily lead to safety accidents. In addition, although traditional automated monitoring systems have been partially applied, they are mostly limited to the acquisition of single parameters (such as temperature or pressure), lacking the ability to collaboratively perceive and intelligently analyze multi-dimensional status information.

[0004] In recent years, with the development of Internet of Things (IoT), Wireless Sensor Networks (WSN), and Low Power Wide Area Network (LPWAN) technologies, intelligent monitoring systems have been gradually applied to urban infrastructure. While communication technologies such as LoRa and NB-IoT have made remote data transmission possible, wireless signals face severe attenuation and multipath interference in enclosed metal structures like heating well chambers, leading to unstable communication links, high data loss rates, and seriously affecting system reliability. Furthermore, existing systems generally adopt a simple "sensor-data upload-cloud alarm" architecture, lacking real-time processing capabilities at the edge, thus failing to achieve rapid response and local decision-making.

[0005] More critically, most current systems still rely on threshold-triggered alarms, meaning an alarm is only triggered when a parameter exceeds a set threshold. This approach fails to identify progressive faults or complex anomalies, hindering early warning systems. For example, a slow temperature rise accompanied by a small gas leak might indicate insulation damage or minor pipe seepage, but the system cannot identify the risk if a single parameter does not reach its threshold. Furthermore, fault diagnosis heavily relies on human experience and lacks scientific model support, resulting in low fault location efficiency and a high misjudgment rate.

[0006] At the equipment level, short sensor lifespans (especially electrochemical gas sensors, which typically last only 1.5–2 years) and limited battery power systems (usually requiring replacement every 3–5 years) lead to high maintenance costs. Frequent equipment replacements not only increase manpower but may also introduce new safety hazards due to improper operation. Although the industry has attempted improvements through multi-sensor integration, communication protocol optimization, and the introduction of machine learning models, the following bottlenecks remain: The data from multiple sensors has not been effectively fused and is still mainly analyzed in parallel and independently, failing to form a collaborative diagnostic capability. Communication optimization is mostly based on general scenario design, lacking targeted modeling and adaptive adjustment for metal-enclosed environments; Machine learning models rely on a large amount of labeled data, but there are few fault samples in heating pipe networks, resulting in poor model generalization ability. The lack of systematic optimization design for the entire life cycle of the equipment has failed to improve reliability from multiple dimensions such as power supply, power consumption, and materials.

[0007] Therefore, there is an urgent need to build a comprehensive solution that integrates multi-source sensing, intelligent analysis, reliable communication and long-life design to achieve early identification, accurate location and efficient handling of faults in the confined space of heating pipe networks. Summary of the Invention

[0008] The purpose of this invention is to provide a fault analysis method and related device in the confined space scenario of heating pipe network, which solves the problem that faults cannot be identified early, accurately located and efficiently handled in the confined space scenario of existing heating pipe network (thermal well chamber, inspection well, etc.) due to the harsh environment, complex multi-parameter coupling, unstable communication and limited diagnostic methods.

[0009] To achieve the above objectives, the technical solution adopted by the present invention is as follows: In a first aspect, the present invention provides a fault analysis method for a heating pipe network in a confined space scenario, comprising the following steps: The original time series data corresponding to the limited space of the heating network are preprocessed to obtain a standardized multi-parameter time series matrix. The obtained standardized multi-parameter time series matrix is ​​used as the input of the LSTM-Attention model to obtain the probability distribution of the fault modes; The equipment surface temperature, convective heat transfer coefficient and ambient temperature are extracted from the obtained raw time series data to form boundary conditions; The obtained boundary conditions are used as input to the physical mechanism-driven model to obtain the temperature deviation value; The probability distribution of the obtained failure modes, temperature deviation values, and original time series data are used as inputs to the DS evidence theory model to obtain the comprehensive failure probability and confidence level. The faults are analyzed based on the obtained comprehensive fault probability and confidence level.

[0010] Preferably, the acquired raw time series data is preprocessed to obtain a standardized multi-parameter time series matrix. The specific method is as follows: The obtained raw time series data are sequentially subjected to sliding window filtering, outlier removal and normalization to obtain the processed time series data; The time-frequency domain features of the processed time-series data are extracted to obtain a standardized multi-parameter time series matrix.

[0011] Preferably, the method for obtaining the physical mechanism-driven model is as follows: Obtain historical raw time-series data corresponding to the limited space of the heating network and the static inherent parameters of the equipment; Historical boundary conditions are extracted based on historical raw time series data; By training the thermodynamic model using historical boundary conditions and static intrinsic parameters, a physical mechanism-driven model is obtained.

[0012] Preferably, the mathematical expression of the thermodynamic model is:

[0013] in, The surface temperature of the equipment; The convective heat transfer coefficient; Ambient temperature; The thermal conductivity of the equipment; The direction of the outward normal to the surface of the device.

[0014] Preferably, the LSTM-Attention model is a pre-built lightweight LSTM model used to output the probability distribution of each fault mode based on the input standardized multi-parameter time series matrix.

[0015] Preferably, the method for constructing the DS evidence theory model is as follows: The probability distribution of failure modes, temperature deviation values, and raw time series data are used as three independent evidence bodies. Each piece of evidence is assigned an initial weight, and the weights are adaptively adjusted based on the confidence level of each piece of evidence. Based on the adjusted weights, the overall failure probability and confidence level are calculated using the DS synthesis rule; The adaptive weight adjustment adopts a dynamic weighting strategy based on the degree of evidence conflict. When the degree of conflict between evidence is higher than a preset threshold, the weight of conflicting evidence is reduced and the weight of consistent evidence is increased.

[0016] Secondly, the present invention provides a fault analysis system for confined space scenarios in heating pipe networks, comprising: The data processing unit is used to preprocess the original time series data corresponding to the limited space of the heating network to obtain a standardized multi-parameter time series matrix. The fault probability acquisition unit is used to take the obtained standardized multi-parameter time series matrix as the input of the LSTM-Attention model to obtain the probability distribution of fault modes. The boundary condition setting unit is used to extract the equipment surface temperature, convective heat transfer coefficient and ambient temperature from the obtained raw time series data to form boundary conditions. The temperature deviation value acquisition unit is used to take the obtained boundary conditions as input to the physical mechanism driven model and obtain the temperature deviation value. The parameter acquisition unit is used to take the probability distribution of the obtained fault modes, temperature deviation values ​​and raw time series data as inputs to the DS evidence theory model to obtain the comprehensive fault probability and confidence level. The fault analysis unit is used to extract time-frequency domain features from the processed time-series data to obtain a standardized multi-parameter time series matrix.

[0017] Thirdly, the present invention provides an electronic device including a processor and a memory, wherein the memory stores computer instructions, and when the computer instructions are executed by the processor, the electronic device performs the method described thereon.

[0018] Fourthly, the present invention provides a computer program product, the computer program product including computer-executable instructions, which, when executed, implement the method described.

[0019] Fifthly, the present invention provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the method described herein.

[0020] Compared with the prior art, the beneficial effects of the present invention are: This invention provides a fault analysis method for heating pipe networks in confined space scenarios. First, it utilizes sliding window filtering, outlier removal, and normalization preprocessing to transform multi-source heterogeneous raw time-series data into a standardized multi-parameter time-series matrix. This effectively eliminates noise and outliers introduced by sensors in the harsh environment of heating well chambers due to factors such as humidity, vibration, and signal interference, fundamentally ensuring the quality and reliability of the input data for subsequent diagnostic models. Then, the standardized matrix is ​​input into an LSTM-Attention model, leveraging the powerful ability of long short-term memory networks to capture temporal dependencies to extract multiple parameters such as temperature, pressure, and gas concentration. The system analyzes the evolution of parameters over time and introduces an attention mechanism to automatically identify key time periods and parameter channels strongly correlated with fault modes, thereby outputting the probability distribution of each fault mode. This achieves data-driven early warning that is independent of a single threshold and can perceive progressive degradation trends. Simultaneously, it extracts equipment surface temperature, convective heat transfer coefficient, and ambient temperature from the raw time-series data as boundary conditions. These are input into a physical mechanism-driven model based on thermodynamic equations. The deviation between theoretical and actual temperatures is calculated using a mathematical expression coupled with Fourier's law of heat conduction and Newton's law of cooling. This deviation directly reflects... Physical anomalies such as insulation layer damage, pipe micro-leakage, and deteriorated heat exchange provide independent verification evidence with clear physical meaning and unaffected by data sample scarcity for the diagnostic results. Ultimately, the aforementioned failure mode probability distributions, temperature deviation values, and original time-series data are treated as three complementary pieces of evidence with different information source characteristics. These are jointly input into the DS evidence theory model. Through the basic probability allocation function, evidence synthesis rules, and adaptive weight adjustment mechanism within this theoretical framework, the potential conflicts and uncertainties among multi-source evidence are effectively resolved. This organically integrates data-driven probabilistic inference with deterministic calculation of physical mechanisms. By combining these methods, a comprehensive fault probability and confidence level with high reliability are output, which significantly reduces the risk of misjudgment that may easily occur when relying on a single data-driven model under conditions of scarce fault samples. At the same time, it also makes up for the limitation of pure physical mechanism models in handling complex multi-parameter coupled scenarios. This forms a three-level progressive fault confirmation chain of "data discovery of anomalies → physical verification of anomalies → fusion confirmation of anomalies", which greatly improves the accuracy, robustness and early warning capability of fault diagnosis in the confined space scenario of heating pipe network. It also provides a solid quantitative decision-making foundation for subsequent risk classification and disposal based on confidence level, root cause location and tracing and closed-loop model iterative optimization. Attached Figure Description

[0021] Figure 1 This is a flowchart of a method involved in an embodiment of the present invention. Detailed Implementation

[0022] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0023] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0024] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0025] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0026] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0027] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0028] Example 1 This embodiment provides a fault analysis method for heating pipe networks in confined space scenarios, including the following steps: Obtain raw time-series data corresponding to the limited space of the heating pipe network, the raw time-series data including multi-channel parameters such as temperature, pressure, gas concentration, humidity and liquid level; The obtained raw time series data is preprocessed to obtain a standardized multi-parameter time series matrix; The obtained standardized multi-parameter time series matrix is ​​used as the input of the LSTM-Attention model to obtain the probability distribution of the fault modes; The equipment surface temperature, convective heat transfer coefficient and ambient temperature are extracted from the obtained raw time series data to form boundary conditions; The obtained boundary conditions are used as input to the physical mechanism-driven model to obtain the temperature deviation value; The probability distribution of the obtained failure modes, temperature deviation values, and original time series data are used as inputs to the DS evidence theory model to obtain the comprehensive failure probability and confidence level.

[0029] Example 2 This embodiment provides a fault analysis method for a confined space scenario in a heating pipe network, including the following steps: Step 1: Real-time acquisition of multi-source heterogeneous data Intelligent monitoring terminals are deployed in the limited space of the heating network (heating wells, inspection wells) to collect raw time-series data every 10 seconds. The raw time-series data includes gas parameters, thermal parameters and environmental parameters.

[0030] In this embodiment, the gas parameters include the concentration of carbon monoxide (CO) and the concentration of hydrogen sulfide (H2S).

[0031] The thermodynamic parameters include pipe temperature and medium pressure.

[0032] The environmental parameters include ambient humidity, water level, and ambient temperature.

[0033] Each sensor uploads data to the edge computing gateway via RS485 bus and Modbus protocol to complete analog-to-digital conversion and data packaging.

[0034] Step 2: Edge-side data preprocessing and feature extraction The original time series data is subjected to sliding window filtering, outlier removal and normalization in sequence to obtain the processed time series data; The time-frequency domain features of the processed time-series data are extracted to obtain a standardized multi-parameter time series matrix.

[0035] In this embodiment, the extracted time-frequency domain features include mean, variance, peak value, rate of change, and spectral energy.

[0036] The surface temperature of the heating pipe, the convective heat transfer coefficient, and the ambient temperature are extracted from the processed time-series data to form boundary conditions. Step 3: Adaptive wireless communication transmission in a metal-enclosed environment Based on the finite element signal attenuation model, the path loss in the cast iron / metal well chamber environment is evaluated in real time. Dynamically adjust communication strategies: Good signal: high-order modulation (16-QAM), low transmit power; Signal attenuation: Switch to low-order modulation (BPSK), enable FEC forward error correction, and increase transmit power (10–20 dBm). Ensure stable data uploads to the cloud, and prioritize the transmission of abnormal data.

[0037] Input: Edge layer feature data, real-time channel quality; Output: Data packets uploaded to the cloud with high reliability.

[0038] Step 4: Mechanism - Data Fusion Fault Diagnosis The obtained standardized multi-parameter time series matrix is ​​used as the input of the LSTM-Attention model to obtain the probability distribution of the fault modes; The obtained boundary conditions are used as input to the physical mechanism-driven model to obtain the temperature deviation value; The probability distribution of the obtained failure modes, temperature deviation values, and original time series data are used as inputs to the DS evidence theory model to obtain the comprehensive failure probability and confidence level.

[0039] In this embodiment, the LSTM-Attention model is a pre-built lightweight LSTM model used to output the probability distribution of each fault mode based on the input standardized multi-parameter time series matrix.

[0040] Step 5: Multi-level alarm and intelligent decision output The overall failure probability is compared with a preset probability threshold, and the confidence level is compared with a preset confidence threshold. The current failure risk level is determined based on the comparison results.

[0041] In this embodiment, when the overall failure probability is ≥0.85 and the confidence level is ≥0.90, it is determined to be an emergency level, triggering automatic ventilation, linked valve shut-off and emergency shutdown commands; When the overall failure probability is between 0.60 and 0.85 and the confidence level is ≥0.75, it is judged as a general level, triggering a local audible and visual alarm and generating an inspection work order; When the overall failure probability is between 0.30 and 0.60 and the confidence level is ≥0.60, it is judged as a warning level, the abnormal trend is recorded and regular observation is recommended; When the overall failure probability is less than 0.30 or the confidence level is lower than the corresponding level threshold, it is judged as a normal or insufficient evidence state, and no alarm is triggered, but only an internal mark is made.

[0042] Step 6: Data Loop Closure Tracking and Model Iterative Optimization Maintenance personnel receive work orders via mobile devices, perform actions, and upload results. The system archives the entire process of alarm, handling, and verification results, forming a data closed loop; The LSTM-Attention and DS fusion model is iteratively optimized using newly labeled samples to improve diagnostic accuracy.

[0043] Inputs: Operation and maintenance handling feedback, fault verification data; Outputs: Closed-loop archived records, optimized diagnostic model Example 3 Based on Example 2, this example provides a fault analysis method for heating pipe networks in a confined space scenario. The method for obtaining the physical mechanism-driven model is as follows: Obtain historical boundary conditions and the corresponding static intrinsic parameters of the device; The thermodynamic model is trained using historical boundary conditions and the corresponding static intrinsic parameters of the device to obtain a physical mechanism-driven model.

[0044] In this embodiment, the mathematical expression of the thermodynamic model is:

[0045] in, The surface temperature of the equipment; The convective heat transfer coefficient; Ambient temperature; The thermal conductivity of the equipment; The direction of the outward normal to the surface of the device.

[0046] Example 4 Based on Example 2, this example provides a fault analysis method for heating pipe networks in a confined space scenario. The improved DS evidence theory model construction method is as follows: Constructing a DS evidence theory model; The probability distribution of failure modes, temperature deviation values, and raw time series data are used as three independent evidence bodies. Each piece of evidence is assigned an initial weight, and the weights are adaptively adjusted based on the confidence level of each piece of evidence. Based on the adjusted weights, the overall failure probability and confidence level are calculated using the DS synthesis rule; The adaptive weight adjustment adopts a dynamic weighting strategy based on the degree of evidence conflict. When the degree of conflict between evidence is higher than a preset threshold, the weight of conflicting evidence is reduced and the weight of consistent evidence is increased.

[0047] Example 5 This embodiment provides a fault analysis system for confined space scenarios in heating pipe networks, including: The data processing unit is used to preprocess the original time series data corresponding to the limited space of the heating network to obtain a standardized multi-parameter time series matrix. The fault probability acquisition unit is used to take the obtained standardized multi-parameter time series matrix as the input of the LSTM-Attention model to obtain the probability distribution of fault modes. The boundary condition setting unit is used to extract the equipment surface temperature, convective heat transfer coefficient and ambient temperature from the obtained raw time series data to form boundary conditions. The temperature deviation value acquisition unit is used to take the obtained boundary conditions as input to the physical mechanism driven model and obtain the temperature deviation value. The parameter acquisition unit is used to take the probability distribution of the obtained fault modes, temperature deviation values ​​and raw time series data as inputs to the DS evidence theory model to obtain the comprehensive fault probability and confidence level. The fault analysis unit is used to extract time-frequency domain features from the processed time-series data to obtain a standardized multi-parameter time series matrix.

[0048] Example 6 This embodiment also provides a computing device. The computing device includes a bus, a processor, a memory, and a communication interface. The processor, memory, and communication interface communicate with each other via the bus. The computing device can be a server or a terminal device. It should be understood that this application does not limit the number of processors and memory in the computing device.

[0049] A bus can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, a bus can include a path for transmitting information between various components of a computing device (e.g., memory, processor, communication interfaces).

[0050] The processor may include any one or more of the following: central processing unit (CPU), graphics processing unit (GPU), tensor processing unit (TPU), application specific integrated circuit (ASIC), field-programmable gate array (FPGA), microprocessor (MP), or digital signal processor (DSP).

[0051] Memory can include volatile memory, such as random access memory (RAM). Processors can also include non-volatile memory. volatile memory, such as read-only memory (ROM). ROM (memory only), flash memory, hard disk drive (HDD), or solid state drive (SSD).

[0052] The memory stores executable program code, which the processor executes to implement the functions of the aforementioned units, thereby achieving, for example, the method described in Embodiment 1. That is, the memory may store instructions for the methods and functions relating to the computing device in any of the above embodiments.

[0053] The communication interface uses transceiver modules such as, but not limited to, network interface cards and transceivers to enable communication between computing devices and other devices or communication networks.

[0054] Example 7 This embodiment also provides a computer-readable storage medium storing computer instructions that, when executed by a processor, cause the processor to perform the methods and functions of the computing device involved in any of the above embodiments.

[0055] Generally, the various embodiments of this disclosure can be implemented in hardware or dedicated circuitry, software, logic, or any combination thereof. Some aspects can be implemented in hardware, while others can be implemented in firmware or software, which can be executed by a controller, microprocessor, or other computing device. Although various aspects of the embodiments of this disclosure are shown and described as block diagrams, flowcharts, or represented using some other illustration, it should be understood that the blocks, apparatuses, systems, techniques, or methods described herein can be implemented as, as non-limiting examples, in hardware, software, firmware, dedicated circuitry or logic, general-purpose hardware or controllers or other computing devices, or some combination thereof.

[0056] Example 8 This embodiment provides at least one computer program product tangibly stored on a non-transitory computer-readable storage medium. The computer program product includes computer-executable instructions, such as instructions included in program modules, which execute in a device on a target real or virtual processor to perform the processes / methods as described above with reference to the accompanying drawings. Typically, program modules include routines, programs, libraries, objects, classes, components, data structures, etc., that perform specific tasks or implement specific abstract data types. In various embodiments, the functionality of program modules can be combined or divided among program modules as needed. The machine-executable instructions for the program modules can execute within a local or distributed device. In a distributed device, the program modules can reside in both local and remote storage media.

[0057] Computer program code used to implement the methods of this disclosure may be written in one or more programming languages. This computer program code may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus, such that when executed by the computer or other programmable data processing apparatus, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be performed. The program code may be executed entirely on a computer, partially on a computer, as a stand-alone software package, partially on a computer and partially on a remote computer, or entirely on a remote computer or server.

[0058] In the context of this disclosure, computer program code or related data may be carried on any suitable carrier to enable a device, apparatus, or processor to perform the various processes and operations described above. Examples of carriers include signals, computer-readable media, and so on. Examples of signals may include electrical, optical, radio, sound, or other forms of propagation signals, such as carrier waves, infrared signals, etc.

[0059] Computer-readable media can be any tangible medium that contains or stores programs for or relating to an instruction execution system, apparatus, or device, or a data storage device such as a data center containing one or more available media. Computer-readable media can be computer-readable signal media or computer-readable storage media. Computer-readable media can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. More detailed examples of computer-readable storage media include electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0060] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A fault analysis method for heating pipe networks in confined space scenarios, characterized in that, Includes the following steps: The original time series data corresponding to the limited space of the heating network are preprocessed to obtain a standardized multi-parameter time series matrix. The obtained standardized multi-parameter time series matrix is ​​used as the input of the LSTM-Attention model to obtain the probability distribution of the fault modes; The equipment surface temperature, convective heat transfer coefficient and ambient temperature are extracted from the obtained raw time series data to form boundary conditions; The obtained boundary conditions are used as input to the physical mechanism-driven model to obtain the temperature deviation value; The probability distribution of the obtained failure modes, temperature deviation values, and original time series data are used as inputs to the DS evidence theory model to obtain the comprehensive failure probability and confidence level. The faults are analyzed based on the obtained comprehensive fault probability and confidence level.

2. The fault analysis method for a confined space scenario in a heating pipe network according to claim 1, characterized in that, The acquired raw time series data is preprocessed to obtain a standardized multi-parameter time series matrix. The specific method is as follows: The obtained raw time series data are sequentially subjected to sliding window filtering, outlier removal and normalization to obtain the processed time series data; The time-frequency domain features of the processed time-series data are extracted to obtain a standardized multi-parameter time series matrix.

3. The fault analysis method for a confined space scenario in a heating pipe network according to claim 1, characterized in that, The method for obtaining the physical mechanism-driven model is as follows: Obtain historical raw time-series data corresponding to the limited space of the heating network and the static inherent parameters of the equipment; Historical boundary conditions are extracted based on historical raw time series data; By training the thermodynamic model using historical boundary conditions and static intrinsic parameters, a physical mechanism-driven model is obtained.

4. The fault analysis method for a confined space scenario in a heating pipe network according to claim 3, characterized in that, The mathematical expression of the thermodynamic model is: in, The surface temperature of the equipment; The convective heat transfer coefficient; Ambient temperature; The thermal conductivity of the equipment; The direction of the outward normal to the surface of the device.

5. The fault analysis method for a heating pipeline network in a confined space scenario according to claim 1, characterized in that, The LSTM-Attention model is a pre-built lightweight LSTM model used to output the probability distribution of each fault mode based on the input standardized multi-parameter time series matrix.

6. The fault analysis method for a confined space scenario in a heating pipe network according to claim 1, characterized in that, The DS evidence theory model construction method is as follows: The probability distribution of failure modes, temperature deviation values, and raw time series data are used as three independent evidence bodies. Each piece of evidence is assigned an initial weight, and the weights are adaptively adjusted based on the confidence level of each piece of evidence. Based on the adjusted weights, the overall failure probability and confidence level are calculated using the DS synthesis rule; The adaptive weight adjustment adopts a dynamic weighting strategy based on the degree of evidence conflict. When the degree of conflict between evidence is higher than a preset threshold, the weight of conflicting evidence is reduced and the weight of consistent evidence is increased.

7. A fault analysis system for confined space scenarios in heating pipe networks, characterized in that, include: The data processing unit is used to preprocess the original time series data corresponding to the limited space of the heating network to obtain a standardized multi-parameter time series matrix. The fault probability acquisition unit is used to take the obtained standardized multi-parameter time series matrix as the input of the LSTM-Attention model to obtain the probability distribution of fault modes. The boundary condition setting unit is used to extract the equipment surface temperature, convective heat transfer coefficient and ambient temperature from the obtained raw time series data to form boundary conditions. The temperature deviation value acquisition unit is used to take the obtained boundary conditions as input to the physical mechanism driven model and obtain the temperature deviation value. The parameter acquisition unit is used to take the probability distribution of the obtained fault modes, temperature deviation values ​​and raw time series data as inputs to the DS evidence theory model to obtain the comprehensive fault probability and confidence level. The fault analysis unit is used to extract time-frequency domain features from the processed time-series data to obtain a standardized multi-parameter time series matrix.

8. An electronic device, characterized in that, It includes a processor and a memory, the memory storing computer instructions that, when executed by the processor, cause the electronic device to perform the method of any one of claims 1 to 6.

9. A computer program product, characterized in that, The computer program product includes computer-executable instructions that, when executed, implement the method of any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions that, when executed by a processor, implement the method of any one of claims 1 to 6.

Images