An electrical fire early warning detection and device installation self-diagnosis method and system

Abstract

This invention relates to the field of electrical safety monitoring technology, and in particular to a method and system for electrical fire early warning detection and equipment installation self-diagnosis. The invention constructs an equipment safety diagnostic model and a learning dataset. After the electrical fire monitoring equipment is initially installed and powered on, multi-dimensional time-series monitoring data generated during the equipment's operation are collected and input into the equipment safety diagnostic model to obtain the safety assessment results output by the model. This invention proposes a self-diagnosis method for electrical fire monitoring equipment installation, enabling self-diagnosis of the installation status of the electrical fire monitoring system, ensuring the correct installation and operation of the system, and improving the accuracy, intelligence, and adaptability of electrical fire monitoring.

Description

Technical Field

[0001] This invention relates to the field of electrical safety monitoring technology, and in particular to a method and system for electrical fire early warning detection and equipment installation self-diagnosis. Background Technology

[0002] In the field of electrical fire monitoring, the detection and early warning of anomalies in electrical circuits and equipment are crucial for preventing fire accidents. Electrical fires are one of the main causes of building fires, especially in older buildings where the risk is extremely high due to aging electrical wiring, substandard construction, and the use of inferior materials. Currently, most electrical fire monitoring systems rely on basic current and temperature monitoring, typically only monitoring abnormal changes in a single parameter and lacking integrated real-time analysis of multiple parameters. This can lead to a failure to identify potential risks in complex situations. Furthermore, traditional electrical fire monitoring systems are prone to problems such as incorrect wiring and improper installation during installation, resulting in reduced monitoring accuracy or even false alarms, affecting the system's reliability and effectiveness.

[0003] US Patent US20140028253A1 proposes an electrical fire monitoring method based on residual current. This method detects circuit abnormalities by utilizing abnormal changes in current, but it is limited to monitoring a single current parameter and fails to comprehensively analyze the changing patterns of multiple electrical parameters, making it difficult to accurately identify more complex abnormalities. Additionally, Chinese Patent CN102345678B describes an electrical fire alarm device based on temperature monitoring. This device triggers an alarm by detecting abnormal increases in circuit temperature. However, in practical applications, relying solely on temperature parameters is insufficient to distinguish different fault types, such as overload and poor contact.

[0004] In summary, existing electrical fire early warning technologies have the following shortcomings:

[0005] 1. Insufficient monitoring accuracy: Existing electrical fire monitoring systems mostly rely on single parameters (such as current or temperature) for detection, lacking the ability to comprehensively analyze multiple parameters. This method is prone to misjudgment in complex situations, making it difficult to accurately distinguish between various fault types such as overload, poor contact, and circuit damage, resulting in inaccurate alarm effects and the risk of false alarms and missed alarms.

[0006] 2. Error-prone installation process: Traditional electrical fire monitoring systems are prone to connection errors during installation, such as reversed neutral wire, missing neutral wire, or insufficient live wire. This can lead to frequent false alarms or even system failure during operation. Current technology lacks automatic diagnostic capabilities for the installation process, relying on manual inspection, which introduces the risk of human error and affects the reliability of the monitoring system.

[0007] 3. Poor adaptability: Because the monitoring algorithms of existing systems are usually relatively simple, they are difficult to adapt to the complexities of different environments and changes in electrical load. For example, under high load or aging lines, the single-parameter alarm threshold of traditional systems may not meet the needs of accurate monitoring, thus affecting the applicability of the system.

[0008] 4. Low level of intelligence: Most current monitoring systems lack intelligent analysis and self-learning capabilities, and cannot adaptively adjust based on the on-site environment or historical data, making it difficult to achieve real-time and efficient monitoring of electrical fires. The systems often underperform when facing new abnormal scenarios and struggle to adapt to the complex and ever-changing characteristics of electrical fire risks. Summary of the Invention

[0009] To overcome the shortcomings of existing electrical fire monitoring systems in terms of monitoring accuracy, installation process, adaptability, and intelligence level, this invention proposes a self-diagnosis method for the installation of electrical fire monitoring equipment. This method enables self-diagnosis of the installation status of the electrical fire monitoring system, ensuring its correct installation and operation, and improving the accuracy, intelligence level, and adaptability of electrical fire monitoring.

[0010] The present invention proposes a self-diagnostic method for electrical fire monitoring equipment installation, comprising the following steps:

[0011] S1. Construct an equipment safety diagnostic model and a learning dataset; train the equipment safety diagnostic model on the learning dataset until convergence; the input of the equipment safety diagnostic model is multivariate time-series monitoring data collected during the operation of electrical equipment, and the output of the equipment safety diagnostic model is the safety assessment result; the safety assessment result is correlated with the risk level of electrical fire.

[0012] The equipment safety diagnostic model is an improvement on the TFT model. The improvement method is to connect a weighted attention module and a multi-scale attention module in parallel between the input layer and the time distribution embedding layer, and replace the multi-head attention module with a context embedding module and a multi-head sparse attention module connected in front and behind.

[0013] The weighted attention module introduces a multi-parameter weighted attention mechanism to generate dynamic weights for each monitored object; the monitored objects correspond one-to-one with the data categories of multivariate time-series monitoring data; the multi-scale attention module captures the complex interaction between short-term abnormal fluctuations and long-term trend changes in multivariate time-series monitoring data.

[0014] The context embedding module embeds and fuses the feature representations of the acquired multivariate time-series monitoring data with environmental data to generate feature representations that adapt to the needs of the scenario; the multi-head sparse attention module integrates sparse attention mechanism and multi-head attention mechanism to process the feature representations output by the context embedding module and then outputs them.

[0015] S2. After the electrical fire monitoring equipment is installed and powered on for the first time, collect multi-dimensional time-series monitoring data generated during the operation of the electrical fire monitoring equipment and input it into the equipment safety diagnosis model to obtain the safety assessment results output by the equipment safety diagnosis model.

[0016] Preferably, the weighted attention module includes: a fully connected layer, a feature weighting layer, and a first output layer; the input of the fully connected layer is connected to the input layer to obtain multivariate time-series monitoring data, the fully connected layer generates dynamic weights for various types of data, the input of the feature weighting layer performs weighted calculations on various types of data in the multivariate time-series monitoring data obtained by the input layer according to the dynamic weights output by the fully connected layer; the first output layer performs convolution processing on the weighted features output by the feature weighting layer.

[0017] Preferably, the multi-scale attention module includes a short-term window module, a first attention module, a second attention module, a feature aggregation module, and a second output layer. The multivariate time-series monitoring data obtained from the input layer is divided into short-window sequences by the short-term window module, and then the first attention module extracts data attention features of approximately instantaneous features and performs weighted operations before outputting them. The multivariate time-series monitoring data is also directly input into the second attention module to extract attention features of long sequences and perform weighted operations before outputting them. The weighted features output by the first attention module and the weighted features output by the second attention module are dimensionally superimposed in the feature aggregation layer and then input into the second output layer for convolution processing, and then used as the output data of the multi-scale attention module.

[0018] Preferably, the context embedding module includes a first feature encoding layer, an environmental information input layer, a second feature encoding layer, a feature concatenation layer, and a third output layer; the first feature encoding layer encodes the feature representation output by the time distribution embedding layer; the environmental information input layer extracts and outputs environmental features based on the obtained environmental data, including environmental conditions and load type; the second feature encoding layer encodes the environmental features output by the environmental information input layer; the feature concatenation layer concatenates the feature codes output by the first feature encoding layer and the feature codes output by the second feature encoding layer in terms of dimensions; and the third output layer performs convolution processing on the concatenated features output by the feature concatenation layer.

[0019] Preferably, the multi-head sparse attention module includes a sparse pattern generation module, a multi-head attention module, and a fourth output layer; the inputs of both the sparse pattern generation module and the multi-head attention module are connected to the output of the context embedding module. The sparse pattern generation module introduces sparsity constraints and generates a sparse attention pattern as the output; the multi-head attention module uses a multi-head attention mechanism to process the input data features and generate multi-head attention features; the input of the fourth output layer is the matrix product of the sparse attention pattern and the multi-head attention features; the fourth output layer performs convolution processing on the input data.

[0020] Preferably, the sparse pattern generation module adopts a weight decay model.

[0021] Preferably, the safety assessment results include correct wiring, reversed neutral wire, no neutral wire, and insufficient live wire.

[0022] Preferably, the data categories of the multivariate time-series monitoring data include at least two of the following: current, voltage, residual current, and wire temperature.

[0023] This application proposes an electrical fire early warning and detection method, which includes the following steps:

[0024] The system establishes a correspondence between security assessment results and risk categories, with some risk categories associated with corresponding alarm procedures.

[0025] Install electrical fire monitoring equipment, execute the self-diagnosis method for electrical fire monitoring equipment installation, and obtain the trained equipment safety diagnosis model and the correctly installed electrical fire monitoring equipment.

[0026] By collecting operational data of the monitored equipment through electrical fire early warning equipment, constructing multi-dimensional time-series monitoring data and inputting it into the equipment safety diagnosis model, the safety assessment results of the monitored equipment and the corresponding risk categories output by the equipment safety diagnosis model are obtained.

[0027] Determine if the risk category is associated with an alarm procedure; if so, execute the corresponding alarm procedure.

[0028] The present invention proposes a system employing the aforementioned electrical fire early warning and detection method, comprising a memory and a processor. The memory is used to store a device safety diagnostic model and a computer program. When the computer program is executed, it is used to implement the aforementioned electrical fire early warning and detection method.

[0029] The advantages of this invention are:

[0030] (1) This application combines multi-dimensional time-series monitoring data generated during the initial operation of electrical fire monitoring equipment and uses a deep learning model to perform real-time analysis of various abnormal scenarios, thereby promptly determining whether the electrical fire monitoring equipment is installed correctly. The system has a self-diagnostic function during the installation process and can automatically detect installation abnormalities such as reversed neutral wire, missing neutral wire, and insufficient live wire, ensuring the accuracy of installation and the reliability of operation. This application adopts intelligent algorithms that can adaptively adjust monitoring parameters according to changes in the site environment and load, improving the system's adaptability and intelligence level, thereby achieving real-time, efficient, and accurate early warning and monitoring of electrical fire risks.

[0031] (2) Multi-dimensional time-series monitoring data includes multiple types of data and integrates various electrical parameter monitoring and anomaly diagnosis technologies. It can identify multiple faults in electrical lines and equipment in real time to ensure that electrical fire monitoring is safe and reliable during installation and operation, and realize accurate fault detection and timely alarm.

[0032] (3) The electrical fire early warning and detection method proposed in this application comprehensively analyzes data such as current, voltage, temperature, and residual current using multiple parameters, and constructs a multi-parameter correlation model using machine learning algorithms. This model can accurately identify various fault types such as line overload, poor contact, line damage, and shared neutral wire. Compared with traditional single-parameter monitoring methods, this application has higher accuracy in diagnosing complex electrical faults and effectively reduces false alarms and missed alarms.

[0033] (4) The equipment safety diagnostic model proposed in this application can automatically adjust the monitoring parameter thresholds according to the electrical load and environmental changes in different locations, enabling the system to adapt to various load changes, different building environments, and line aging conditions, thus meeting a wide range of application needs. Its high adaptability and scalability effectively broaden the application scope and are suitable for various electrical fire monitoring scenarios.

[0034] (5) This application employs machine learning algorithms to perform real-time analysis of the collected data and possesses self-learning capabilities. It can continuously optimize the equipment safety diagnosis model based on historical data, thereby improving the accurate monitoring of electrical fire risks. The high level of intelligence in the system not only improves the accuracy of fault diagnosis but also simplifies manual operation, reduces the monitoring burden on staff, and improves operational efficiency.

[0035] (6) The equipment safety diagnostic model provided in this application establishes a correspondence between safety assessment results and risk levels. It improves classification robustness through a layered soft-label structure, achieving high-precision identification of multiple types of faults and providing comprehensive and efficient core technical support for electrical fire monitoring systems. Furthermore, it provides multi-level early warning prompts based on the urgency of different fault types, facilitating rapid response by management personnel in emergency situations. Attached Figure Description

[0036] Figure 1 A flowchart illustrating a self-diagnostic method for installing an electrical fire monitoring device;

[0037] Figure 2 This is a structural diagram of an improved equipment safety diagnostic model based on the TFT model.

[0038] Figure 3 This is a structural diagram of a TFT model;

[0039] Figure 4 This is a structural diagram of the weighted attention module;

[0040] Figure 5 For multi-scale attention module structure diagram;

[0041] Figure 6 Embed the module structure diagram for context;

[0042] Figure 7 This is a structural diagram of a multi-head sparse attention module;

[0043] Figure 8 Flowchart of an electrical fire early warning and detection method;

[0044] Figure 9 For model performance comparison;

[0045] Figure 10 For model delay comparison. Detailed Implementation

[0046] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0047] Reference Figure 1 This application proposes a self-diagnostic method for electrical fire monitoring equipment installation, comprising the following steps:

[0048] S1. Construct an equipment safety diagnostic model and a learning dataset; train the equipment safety diagnostic model on the learning dataset until convergence; the input of the equipment safety diagnostic model is multivariate time-series monitoring data collected during the operation of electrical equipment, and the output of the equipment safety diagnostic model is the safety assessment result, including correct wiring, reversed neutral wire, no neutral wire, and insufficient live wire.

[0049] Multivariate time-series monitoring data includes multiple parameters such as current, voltage, residual current, and wire temperature.

[0050] S2. After the electrical fire monitoring equipment is installed and powered on for the first time, collect multi-dimensional time-series monitoring data generated during the operation of the electrical fire monitoring equipment and input it into the equipment safety diagnosis model to obtain the safety assessment results output by the equipment safety diagnosis model.

[0051] In this embodiment, after the electrical fire monitoring equipment is installed, it first performs a self-diagnosis of its installation status, which effectively avoids incorrect detection results of the electrical system caused by incorrect installation of the electrical fire monitoring equipment, improves the reliability of the electrical fire monitoring system in monitoring other electrical equipment, ensures the correct installation of the electrical fire monitoring equipment in different environments, and improves its adaptability.

[0052] Reference Figure 2 The equipment safety diagnostic model includes an input layer, a weighted attention module, a multi-scale attention module, a temporal distribution embedding layer, a context embedding module, a multi-head sparse attention module, and a decoder. The input layer is used to input multivariate time-series monitoring data and feeds it to the weighted attention module and the multi-scale attention module respectively. The outputs of the weighted attention module and the multi-scale attention module are both connected to the input of the temporal distribution embedding layer. The input data of the temporal distribution embedding layer is the result of the dimensional superposition of the output data of the weighted attention module and the output data of the multi-scale attention module. The temporal distribution embedding layer, the context embedding module, the multi-head sparse attention module, and the decoder are connected sequentially, and the output of the decoder is the output of the equipment safety diagnostic model.

[0053] It is evident that the equipment safety diagnostic model proposed in this application is an improvement upon the TFT model, the structure of which is as follows: Figure 3 As shown. Compared to the TFT model, the device safety diagnostic model proposed in this application connects a weighted attention module and a multi-scale attention module in parallel between the input layer and the time distribution embedding layer, and replaces the multi-head attention module with a context embedding module and a multi-head sparse attention module connected in sequence.

[0054] The weighted attention module introduces a multi-parameter weighted attention mechanism to generate dynamic weights for each monitored object; the monitored objects correspond one-to-one with the data categories of multivariate time-series monitoring data, i.e., the monitored objects are current, voltage, residual current or wire temperature.

[0055] Reference Figure 4 The weighted attention module proposed in this application includes a fully connected layer, a feature weighting layer, and a first output layer. The input of the fully connected layer is connected to the input layer to obtain multivariate time-series monitoring data. The fully connected layer generates dynamic weights for various types of data. The input of the feature weighting layer performs weighted calculations on various types of data in the multivariate time-series monitoring data obtained from the input layer based on the dynamic weights output by the fully connected layer. The first output layer performs convolution processing on the weighted features output by the feature weighting layer and then outputs the result. In specific implementations, the first output layer can be a convolutional layer.

[0056] Specifically, let X be the multivariate time-series monitoring data obtained by the input layer;

[0057] X={X1;X2;X3;…;Xt;…;XT}

[0058] Xt={x(t,1);x(t,2);…;x(t,n);…;x(t,N)}

[0059] Where T represents the time step, 1≤t≤T; Xt represents the data collected at the t-th time point; x(t,n) represents the n-th type of data collected at the t-th time point; N is the total number of data types, 1≤n≤N; it can be seen that X1, X2, X3, and XT represent the data collected at the 1st, 2nd, 3rd, and T-th time points, respectively; x(t,1), x(t,2), and x(t,N) represent the 1st, 2nd, and Nth types of data collected at the t-th time point, respectively.

[0060] The fully connected layer generates dynamic weights for various data types, and the weight of the nth data type generated by it is denoted as a. n The feature-weighted layer then performs the following operations:

[0061] X'={a n x(t,n)|1≤t≤T;1≤n≤N}

[0062] X' represents the weighted feature output by the feature weighting layer.

[0063] Reference Figure 5 The multi-scale attention module captures the complex interaction between short-term abnormal fluctuations and long-term trend changes in multivariate time-series monitoring data. It uses different time windows to segment the input multivariate time-series monitoring data, performs hierarchical analysis of electrical parameters at different time scales, prioritizes the capture of sudden anomalies, and further improves the response speed to instantaneous faults.

[0064] Specifically, the multi-scale attention module includes a short-term window module, a first attention module, a second attention module, a feature aggregation module, and a second output layer. The multivariate temporal monitoring data obtained from the input layer is divided into short-window sequences by the short-term window module. The first attention module then extracts near-instantaneous data attention features, performs weighted operations, and outputs the results. The multivariate temporal monitoring data is also directly input into the second attention module to extract long-sequence attention features, which are then weighted and output. The weighted features output by the first and second attention modules are dimensionally superimposed in the feature aggregation layer and then input into the second output layer for convolution processing, ultimately serving as the output data for the multi-scale attention module.

[0065] Specifically, the feature aggregation layer superimposes the weighted features output by the first attention module and the weighted features output by the second attention module in the time dimension.

[0066] The input to the temporal distribution embedding layer is the result of superimposing the output data of the weighted attention module and the output data of the multi-scale attention module in the time dimension; the temporal distribution embedding layer performs time series embedding on the selected features to generate more expressive feature representations.

[0067] Reference Figure 6The context embedding module includes a first feature encoding layer, an environmental information input layer, a second feature encoding layer, a feature concatenation layer, and a third output layer. The first feature encoding layer encodes the feature representation output by the temporal distribution embedding layer; the environmental information input layer extracts environmental features from the obtained environmental data and outputs them, including environmental conditions such as temperature, humidity, and load type; the second feature encoding layer encodes the environmental features output by the environmental information input layer; the feature concatenation layer concatenates the feature codes output by the first and second feature encoding layers; and the third output layer performs convolution processing on the concatenated features output by the feature concatenation layer and outputs them as the context embedding module's output data.

[0068] Specifically, the first feature encoding layer, the second feature encoding layer, and the third output layer all use convolutional networks, while the feature concatenation layer uses a Cancat network.

[0069] Reference Figure 7 The multi-head sparse attention module comprises a sparse pattern generation module, a multi-head attention module, and a fourth output layer. The inputs to both the sparse pattern generation module and the multi-head attention module are connected to the output of the context embedding module. The sparse pattern generation module introduces sparsity constraints to reduce redundant parameters. It selects key locations of input data features using a lightweight network and generates sparse attention patterns as output. The multi-head attention module processes the input data features using a multi-head attention mechanism to generate multi-head attention features. The input to the fourth output layer is the matrix product of the sparse attention pattern and the multi-head attention features; the fourth output layer performs convolution processing on the input data and serves as the output of the multi-head sparse attention module.

[0070] Specifically, the sparse pattern generation module uses a weight decay model, and the fourth output layer uses a convolutional network.

[0071] The decoder decodes the data features output by the multi-head sparse attention module to obtain the security assessment results.

[0072] In this embodiment, the temporal distribution embedding layer, the multi-head attention module in the multi-head sparse attention module, and the decoder are obtained from the TFT model (Temporal Fusion Transformer).

[0073] The equipment safety diagnostic model provided in this application introduces a multi-parameter weighted attention mechanism through a weighted attention module. A learnable fully connected layer generates dynamic weights for each input feature, highlighting the impact of key features such as current and temperature on fault detection. A multi-scale attention mechanism captures the complex interaction between short-term abnormal fluctuations and long-term trend changes. Different time windows are used to segment the input data, and electrical parameters are analyzed hierarchically at different time scales, prioritizing the capture of sudden anomalies and further improving the response speed to instantaneous faults. A context-aware feature embedding module is introduced to integrate input features with contextual information such as environmental conditions and load types, generating feature representations adapted to specific scenario requirements, improving adaptability to complex application scenarios, and enhancing the accuracy and robustness of fault detection. The model integrates a sparse attention mechanism to reduce computational complexity and supports embedded hardware deployment through a lightweight design, ensuring the model's real-time performance and applicability, further optimizing system performance.

[0074] Reference Figure 8 The electrical fire early warning and detection method proposed in this application includes the following steps.

[0075] S1. Set the correspondence between security assessment results and risk categories, and associate some risk categories with corresponding alarm procedures;

[0076] S2. Install electrical fire monitoring equipment, perform the above-mentioned self-diagnosis method for electrical fire monitoring equipment installation, and obtain the trained equipment safety diagnosis model and the correctly installed electrical fire monitoring equipment.

[0077] S3. Collect the operating data of the monitored equipment through the electrical fire early warning equipment, construct multi-dimensional time-series monitoring data and input it into the equipment safety diagnosis model, and obtain the safety assessment results of the monitored equipment and the corresponding risk category output by the equipment safety diagnosis model;

[0078] S4. Determine if the risk category is associated with an alarm procedure. If so, execute the corresponding alarm procedure.

[0079] The electrical fire early warning and detection system proposed in this application includes an electrical fire monitoring device and an alarm feedback module; the electrical fire monitoring device includes a data acquisition module and a fault diagnosis module.

[0080] The fault diagnosis module is connected to the data acquisition module and the alarm feedback module respectively.

[0081] The data acquisition module monitors the operating data of electrical equipment in real time through sensors, especially key parameters such as current, voltage, temperature and residual current in the circuit;

[0082] The fault diagnosis module stores the trained equipment safety diagnosis model. The fault diagnosis module acquires the key parameters collected by the data acquisition module and organizes them into a multivariate data monitoring sequence. The fault diagnosis module inputs the multivariate data monitoring sequence into the equipment safety diagnosis model to obtain the safety assessment results.

[0083] The fault diagnosis module is also used to identify the risk category of the safety assessment results. Specifically, safety assessment results other than correct wiring are recorded as risk categories and different risk levels are marked according to experience. When the safety assessment result is a risk category, the fault diagnosis module will install the assessment result and risk level to notify the alarm feedback module. The alarm feedback module will execute the alarm procedure, such as issuing an audible and visual alarm and notifying the corresponding personnel.

[0084] Thus, when a fault or installation abnormality is detected, an audible and visual alarm can be immediately issued through the alarm feedback module. Simultaneously, the fault information is wirelessly transmitted to the back-end management system, facilitating timely response and protective measures by management personnel. Furthermore, the system provides a tiered early warning function, adjusting the alarm level according to the urgency of the fault to ensure rapid and targeted fault handling, thereby improving the safety and management efficiency of electrical fire monitoring.

[0085] The following specific embodiments verify the above-mentioned device safety diagnostic model. This embodiment uses four existing models as comparison models: LSTM (Long Short-Term Memory Network), Transformer model, TCN (Temporal Convolutional Network) model, and TFT (Temporal Fusion Transformer) model.

[0086] The learning dataset constructed in this application consists of {voltage, current, residual current, wire temperature; safety assessment results}; the learning dataset is divided into a training set and a test set.

[0087] In this embodiment, four comparison models and Figure 2 The device safety diagnostic models shown were all trained to convergence on the training set and then tested on the test set. The test results are shown in Table 1 below. Figure 9 , Figure 10 As shown in the figure. The comparison shows that the present application outperforms the comparative models in terms of accuracy, recall, and F1 score, and significantly reduces latency compared to the TFT model, thereby improving the diagnostic efficiency of electrical equipment.

[0088] Table 1: Model Performance Comparison

[0089]

[0090]

[0091] Of course, those skilled in the art will recognize that the present invention is not limited to the details of the exemplary embodiments described above, but also includes the same or similar structures that can be implemented in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered illustrative and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0092] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

[0093] The technologies, shapes, and structures not described in detail in this invention are all known technologies.

Claims

1. A self-diagnostic method for installing electrical fire monitoring equipment, characterized in that, Includes the following steps: S1. Construct a device safety diagnostic model and a learning dataset; The equipment safety diagnostic model is trained on the learning dataset until convergence; the input of the equipment safety diagnostic model is multivariate time-series monitoring data collected during the operation of electrical equipment, and the output of the equipment safety diagnostic model is the safety assessment result; the safety assessment result is correlated with the risk level of electrical fire. The equipment safety diagnostic model is an improvement on the TFT model. The improvement method is to connect a weighted attention module and a multi-scale attention module in parallel between the input layer and the time distribution embedding layer, and replace the multi-head attention module with a context embedding module and a multi-head sparse attention module connected in front and behind. The weighted attention module introduces a multi-parameter weighted attention mechanism to generate dynamic weights for each monitored object; The monitored objects correspond one-to-one with the data categories of multi-dimensional time-series monitoring data; The multi-scale attention module captures the complex interaction between short-term abnormal fluctuations and long-term trend changes in multi-dimensional time-series monitoring data; The context embedding module embeds and fuses the feature representations of the acquired multi-dimensional time-series monitoring data with environmental data to generate feature representations that adapt to the needs of the scenario. The multi-head sparse attention module integrates sparse attention and multi-head attention mechanisms to process the feature representation output by the context embedding module before outputting it. S2. After the electrical fire monitoring equipment is installed and powered on for the first time, collect multi-dimensional time-series monitoring data generated during the operation of the electrical fire monitoring equipment and input it into the equipment safety diagnosis model to obtain the safety assessment results output by the equipment safety diagnosis model. The context embedding module includes a first feature encoding layer, an environment information input layer, a second feature encoding layer, a feature concatenation layer, and a third output layer; The first feature encoding layer encodes the feature representation output by the time distribution embedding layer; the environmental information input layer extracts environmental features and outputs them by combining the obtained environmental data, which includes environmental conditions and load type; the second feature encoding layer encodes the environmental features output by the environmental information input layer. The feature concatenation layer concatenates the feature codes output from the first feature encoding layer and the feature codes output from the second feature encoding layer in terms of dimensions. The third output layer performs convolution processing on the concatenated features output from the feature concatenation layer.

2. The self-diagnostic method for electrical fire monitoring equipment as described in claim 1, characterized in that, The weighted attention module includes a fully connected layer, a feature weighting layer, and a first output layer. The input of the fully connected layer is connected to the input layer to obtain multivariate time-series monitoring data. The fully connected layer generates dynamic weights for various types of data. The input of the feature weighting layer performs weighted calculations on various types of data in the multivariate time-series monitoring data obtained from the input layer based on the dynamic weights output by the fully connected layer. The first output layer performs convolution processing on the weighted features output by the feature weighting layer.

3. The self-diagnostic method for electrical fire monitoring equipment as described in claim 1, characterized in that, The multi-scale attention module includes a short-term window module, a first attention module, a second attention module, a feature aggregation module, and a second output layer. The multivariate time-series monitoring data obtained from the input layer is divided into short-window sequences by the short-term window module, and then the first attention module extracts data attention features of approximately instantaneous features and performs weighted operations before outputting them. The multivariate time-series monitoring data is also directly input into the second attention module to extract attention features of long sequences and perform weighted operations before outputting them. The weighted features output by the first attention module and the weighted features output by the second attention module are dimensionally superimposed in the feature aggregation layer and then input into the second output layer for convolution processing, and then used as the output data of the multi-scale attention module.

4. The self-diagnostic method for electrical fire monitoring equipment installation as described in claim 1, characterized in that, The multi-head sparse attention module includes a sparse pattern generation module, a multi-head attention module, and a fourth output layer; the inputs of the sparse pattern generation module and the multi-head attention module are both connected to the output of the context embedding module. The sparse pattern generation module introduces sparsity constraints and generates sparse attention patterns as output; the multi-head attention module uses a multi-head attention mechanism to process the input data features and generate multi-head attention features; the input of the fourth output layer is the matrix product of the sparse attention pattern and the multi-head attention features; the fourth output layer performs convolution processing on the input data.

5. The self-diagnostic method for electrical fire monitoring equipment as described in claim 4, characterized in that, The sparse pattern generation module uses a weight decay model.

6. The self-diagnostic method for electrical fire monitoring equipment installation as described in any one of claims 1-5, characterized in that, The safety assessment results include whether the wiring is correct, the neutral wire is reversed, the neutral wire is not installed, and the live wire is insufficient.

7. The self-diagnostic method for installing electrical fire monitoring equipment as described in any one of claims 1-5, characterized in that, The data categories of multivariate time-series monitoring data include at least two of the following: current, voltage, residual current, and wire temperature.

8. An electrical fire early warning and detection method employing the self-diagnostic method for installing electrical fire monitoring equipment as described in any one of claims 1-7, characterized in that, Includes the following steps: The system establishes a correspondence between security assessment results and risk categories, with some risk categories associated with corresponding alarm procedures. Install electrical fire monitoring equipment, perform the electrical fire monitoring equipment installation self-diagnosis method as described in any one of claims 1-7, and obtain the trained equipment safety diagnosis model and the correctly installed electrical fire monitoring equipment; By collecting operational data of the monitored equipment through electrical fire early warning equipment, constructing multi-dimensional time-series monitoring data and inputting it into the equipment safety diagnosis model, the safety assessment results of the monitored equipment and the corresponding risk categories output by the equipment safety diagnosis model are obtained. Determine if the risk category is associated with an alarm procedure; if so, execute the corresponding alarm procedure.

9. A system employing the electrical fire early warning and detection method as described in claim 8, characterized in that, It includes a memory and a processor, the memory being used to store a device safety diagnostic model and a computer program, which, when executed, is used to implement the electrical fire early warning and detection method as described in claim 8.

Images