Building electrical safety intelligent evaluation method and system based on dynamic monitoring of Internet of Things

Abstract

The invention relates to the technical field of building electrical, and discloses a building electrical safety intelligent evaluation method and system based on Internet of Things dynamic monitoring, and the method comprises the steps: collecting multi-dimensional data in real time through an Internet of Things sensor network deployed in a building electrical system; carrying out wavelet decomposition on the collected multi-dimensional data, extracting time domain and frequency domain features, carrying out feature fusion and noise removal on multi-scale wavelet coefficients by using a convolutional neural network, and obtaining a denoised multi-dimensional feature vector; inputting the multi-dimensional feature vector into a dynamic evaluation model, and outputting a risk level, fault probability prediction and a key risk factor through the dynamic evaluation model; according to an output result of the dynamic evaluation model, performing graded early warning on the building electrical system; according to the invention, dynamic evaluation and early warning are carried out on the safety state of the building electrical system, evaluation is comprehensive and accurate, and the efficiency and level of building electrical safety management are improved.

Description

Technical Field

[0001] The present invention relates to the field of building electrical technology, and in particular to a building electrical safety intelligent assessment method and system based on dynamic monitoring of the Internet of Things. Background Art

[0002] With the acceleration of urbanization and the continuous expansion of building scale, building electrical systems are becoming increasingly complex. As an important component of building safety, building electrical safety is directly related to the safety of people's lives and property. However, traditional building electrical safety assessment methods mainly rely on manual inspections and regular testing. These problems include incomplete data collection, low assessment efficiency, and poor real-time performance, making it difficult to meet the needs of modern building electrical safety management. The development of Internet of Things technology has provided new ideas for building electrical safety assessment. By deploying a large number of sensors in building electrical systems, multi-dimensional information such as electrical equipment operating data and environmental data can be collected in real time. However, how to effectively utilize this massive amount of data for building electrical safety assessment remains an urgent problem. Summary of the Invention

[0003] The purpose of the present invention is to solve the above problems and design a building electrical safety intelligent assessment method and system based on dynamic monitoring of the Internet of Things.

[0004] A first aspect of the present invention provides a method for intelligent assessment of building electrical safety based on dynamic monitoring of the Internet of Things, the method comprising the following steps: Collect multi-dimensional data in real time through IoT sensor networks deployed in building electrical systems; Perform wavelet decomposition on the collected multi-dimensional data to extract time domain and frequency domain features. Use convolutional neural network to fuse the multi-scale wavelet coefficients and remove noise to obtain the denoised multi-dimensional feature vector. Input the multi-dimensional feature vector into the dynamic assessment model, and output the risk level, failure probability prediction and key risk factors through the dynamic assessment model; Based on the output results of the dynamic assessment model, graded early warning is given to the building electrical system.

[0005] Optionally, in a first implementation of the first aspect of the present invention, performing wavelet decomposition on the collected multi-dimensional data to extract time domain and frequency domain features includes: The multi-dimensional data is decomposed into three layers by Symlet wavelet to obtain wavelet coefficients of different frequency bands; The decomposed wavelet coefficients of each layer are subjected to feature extraction, including time domain features and frequency domain features, wherein the time domain features include at least mean, variance, peak, kurtosis and skewness, and the frequency domain features include at least frequency amplitude, energy distribution and frequency center of gravity.

[0006] Optionally, in a second implementation of the first aspect of the present invention, the use of a convolutional neural network to perform feature fusion on multi-scale wavelet coefficients to remove noise to obtain a denoised multi-dimensional feature vector includes: The multi-scale wavelet coefficients are received through the input layer of the convolutional neural network, and the wavelet coefficients of each layer are superimposed layer by layer to form a multi-dimensional feature matrix; The convolution layer sets multiple convolution kernels, extracts features from the input multi-dimensional feature matrix through convolution operations, and automatically learns features of different scales and positions; The pooling layer uses the maximum pooling method to reduce the dimension of the feature map output by the convolutional layer. The fully connected layer expands the pooled features into a one-dimensional vector, performs nonlinear transformation through the ReLU function, and the output layer outputs the denoised multi-dimensional feature vector.

[0007] Optionally, in a third implementation of the first aspect of the present invention, inputting the multi-dimensional feature vector into a dynamic assessment model and outputting the risk level, failure probability prediction, and key risk factors through the dynamic assessment model includes: The multi-dimensional feature vector is input into a dynamic evaluation model, wherein the dynamic evaluation model integrates a fuzzy hierarchical analysis algorithm, an LSTM network, and an isolation forest algorithm; The fuzzy hierarchical analysis algorithm is used to generate a comprehensive risk-based score at the current moment. The LSTM network prediction automatically captures long-term dependencies and outputs risk trend prediction results. The isolation forest algorithm is used to check for abnormal data points. The weighted average method is used to integrate the output results of the fuzzy hierarchical analysis algorithm, LSTM network and isolation forest algorithm to obtain the final risk level, failure probability prediction and key risk factors.

[0008] Optionally, in a fourth implementation of the first aspect of the present invention, the generating of the comprehensive risk basic score at the current moment by using the fuzzy analytic hierarchy process algorithm includes: Decompose the evaluation target into multiple levels, including the target level, the criterion level, and the indicator level. Standardize the multi-dimensional feature vectors, and then match the standardized feature vectors with the indicator level and convert them into feature values. The fuzzy judgment matrix is ​​constructed using triangular fuzzy numbers, and the fuzzy weight of each index is calculated using the fuzzy eigenvector method. Fuzzy membership is assigned to each indicator according to the size of the eigenvalue, and the fuzzy membership of the eigenvalue is combined with the fuzzy weight to output the comprehensive risk-based score at the current moment through the weighted summation method.

[0009] Optionally, in a fifth implementation of the first aspect of the present invention, automatically capturing long-term dependencies and outputting risk trend prediction results through LSTM network prediction includes: Arrange the multi-dimensional feature vectors in time series to obtain a time series feature dataset with the current moment as the end point and intercept the first N time points as the input window; The LSTM network captures the long-term dependencies of data in multi-dimensional feature vectors and predicts changes in risk scores over the next 1 hour, 6 hours, and 24 hours. The predicted risk score changes are classified and multi-period risk trend prediction results are output, where the multi-period risk trend prediction results at least include specific score predictions and trend descriptions.

[0010] Optionally, in a sixth implementation of the first aspect of the present invention, the step of checking abnormal data points using an isolation forest algorithm includes: Randomly select 2-3 indicators from all indicators, randomly generate splitting planes in multidimensional space, and calculate the path length from the root node to the leaf node of each isolated tree in the multidimensional feature vector; The path lengths of all trees are averaged to obtain the anomaly score of the current data point, and data points exceeding the threshold are marked as anomalies, where the threshold is an anomaly score higher than 0.8; The characteristics of the anomaly points are analyzed to determine whether they are caused by equipment failure, external interference or data collection errors, and the marked anomaly points are fed back to the fuzzy hierarchical analysis algorithm and LSTM network.

[0011] A second aspect of the present invention provides a building electrical safety intelligent assessment system based on dynamic monitoring by the Internet of Things, a data acquisition module for collecting multi-dimensional data in real time through an Internet of Things sensor network deployed in the building electrical system; The feature processing module is used to perform wavelet decomposition on the collected multi-dimensional data, extract time domain and frequency domain features, use convolutional neural networks to perform feature fusion on multi-scale wavelet coefficients, remove noise, and obtain the denoised multi-dimensional feature vector; Dynamic assessment module, used to input multi-dimensional feature vectors into the dynamic assessment model, and output risk level, failure probability prediction and key risk factors through the dynamic assessment model; The graded warning module is used to provide graded warnings for the building electrical system based on the output results of the dynamic assessment model.

[0012] The third aspect of the present invention provides a building electrical safety intelligent assessment device based on dynamic monitoring of the Internet of Things, the building electrical safety intelligent assessment device based on dynamic monitoring of the Internet of Things includes a memory and at least one processor, the memory stores instructions; the at least one processor calls the instructions in the memory to enable the building electrical safety intelligent assessment device based on dynamic monitoring of the Internet of Things to perform each step of the building electrical safety intelligent assessment method based on dynamic monitoring of the Internet of Things as described in any one of the above items.

[0013] The fourth aspect of the present invention provides a computer-readable storage medium having instructions stored thereon, which, when executed by a processor, implement the various steps of the intelligent assessment method for building electrical safety based on dynamic monitoring of the Internet of Things as described in any of the above items.

[0014] In the technical solution provided by the present invention, multi-dimensional data is collected in real time through the Internet of Things sensor network deployed in the building electrical system; the collected multi-dimensional data is subjected to wavelet decomposition to extract time domain and frequency domain features, and the multi-scale wavelet coefficients are subjected to feature fusion using a convolutional neural network to remove noise, thereby obtaining a denoised multi-dimensional feature vector; the multi-dimensional feature vector is input into a dynamic assessment model, and the dynamic assessment model outputs risk levels, failure probability predictions, and key risk factors; based on the output results of the dynamic assessment model, a graded early warning is performed on the building electrical system; the present invention collects multi-dimensional data of the building electrical system in real time, solving the problems of incomplete data collection and poor real-time performance of traditional assessment methods; the dynamic assessment and early warning of the safety status of the building electrical system are comprehensive and accurate, providing scientific and effective decision-making support for building electrical safety management, and improving the efficiency and level of building electrical safety management. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Various other advantages and benefits will become apparent to those skilled in the art by reading the following detailed description of the preferred embodiment.The accompanying drawings are only for the purpose of illustrating the preferred embodiment and are not to be considered as limiting the present invention.

[0016] Figure 1 A flowchart of a method for intelligent building electrical safety assessment based on dynamic monitoring of the Internet of Things provided by an embodiment of the present invention; Figure 2 A schematic diagram of the structure of a building electrical safety intelligent assessment system based on dynamic monitoring of the Internet of Things provided by an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of a building electrical safety intelligent assessment device based on Internet of Things dynamic monitoring provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The terms "first," "second," "third," "fourth," and so forth (if any) in the description and claims of the present invention and in the accompanying drawings are used to distinguish similar items and are not necessarily used to describe a particular order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments described herein can be implemented in an order other than that shown or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, apparatus, product, or device that includes a series of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, product, or device.

[0018] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 The flow chart of the method for intelligent building electrical safety assessment based on dynamic monitoring of the Internet of Things provided by an embodiment of the present invention specifically includes the following steps: Step 101: Collect multi-dimensional data in real time through an IoT sensor network deployed in a building electrical system; In this embodiment, sensors are deployed on key equipment, important lines, and areas susceptible to environmental impacts of the building electrical system to ensure comprehensive and representative data collection. These include key equipment such as transformers, distribution cabinets, generators, switchgear, and line connectors; important lines such as trunk lines and branch lines; and areas susceptible to environmental impacts such as humid, dusty, and high-temperature areas. Sensor installation locations are reasonably determined based on the importance of the equipment, the operating environment, and the points where faults are prone to occur. For example, temperature sensors and vibration sensors are installed on the oil tank, bushing, and windings of the transformer; temperature sensors and current sensors are installed on the switches, busbars, and cable connections of the distribution cabinet; and temperature and humidity sensors, dust concentration sensors, and harmful gas sensors are installed in areas with complex environments. Electromagnetic or capacitive voltage sensors are used to collect data such as phase voltage and line voltage at each node in real time at a frequency of 100 Hz to monitor voltage stability and symmetry. Current transformers or Hall-effect current sensors are used to collect data such as phase current, line current, and zero-sequence current at a frequency of 100 Hz to monitor current magnitude and trends. Power transmitters collect data such as active power, reactive power, apparent power, and power factor at a frequency of 50 Hz to assess power consumption and efficiency of electrical systems. Power quality sensors detect power quality parameters such as voltage swells, sags, interruptions, harmonics, interharmonics, voltage fluctuations, and flicker at a frequency of 200 Hz to assess the impact of power quality on electrical equipment. Use thermocouples, thermistors, or infrared temperature sensors, installed on or inside the equipment, to collect temperature data at a frequency of 10Hz. For example, transformer winding and core temperatures, switch contact temperatures within the distribution cabinet, and busbar temperatures are used to monitor equipment heating and determine whether the equipment is overloaded or has poor contact. Use accelerometers or velocity sensors, installed on the equipment's housing or foundation, to collect vibration signals at a frequency of 1000Hz. By analyzing characteristics such as the frequency, amplitude, and phase of the vibration signal, determine whether the equipment has mechanical faults such as bearing wear, gear damage, or motor imbalance. Use insulation resistance testers or online insulation monitoring devices to collect insulation resistance data from electrical equipment periodically or in real time at a frequency of 1 time per minute. This is used to evaluate the insulation performance of the equipment and determine whether the insulation is aged, damp, or damaged. Use ultrasonic sensors, high-frequency current sensors, or fiber optic sensors to detect partial discharge signals within electrical equipment at a frequency of 100kHz. Partial discharge is an early sign of insulation damage in equipment. By monitoring partial discharge signals, potential equipment failures can be detected promptly. Temperature and humidity sensors are installed in areas such as electrical equipment rooms, cable trenches, and distribution cabinets to collect environmental temperature and humidity data. High temperature and high humidity environments will accelerate the aging of equipment insulation and reduce equipment reliability, so real-time monitoring of environmental temperature and humidity is required. Dust concentration sensors are deployed in dusty environments to collect dust concentration data in the air. Dust accumulation can lead to poor heat dissipation, reduced insulation, and even short circuit failures. For areas where harmful gases may exist, such as SF6, CO, NOx, etc., corresponding gas sensors are installed to collect harmful gas concentration data. Leakage of harmful gases may indicate equipment failure or environmental safety hazards.

[0019] Data collected by sensors is transmitted to a data acquisition terminal via wireless or wired communication. After preliminary data processing, the terminal transmits the data to a cloud server or local server for storage via the IoT platform. Distributed or relational databases are used to store the data, ensuring its security, integrity, and scalability. Data backup and recovery mechanisms are also established to prevent data loss.

[0020] Step 102: performing wavelet decomposition on the collected multi-dimensional data to extract time domain and frequency domain features, and using a convolutional neural network to perform feature fusion on the multi-scale wavelet coefficients to remove noise and obtain a denoised multi-dimensional feature vector; In this embodiment, a three-layer wavelet decomposition is performed on the multi-dimensional data using Symlet wavelet to obtain wavelet coefficients of different frequency bands; feature extraction is performed on the decomposed wavelet coefficients of each layer, including time domain features and frequency domain features, wherein the time domain features include at least mean, variance, peak, kurtosis and skewness, and the frequency domain features include at least frequency amplitude, energy distribution and frequency center of gravity.

[0021] In this embodiment, according to the characteristics of the collected data and the analysis requirements, a suitable wavelet basis function is selected. For electrical parameter data, such as voltage and current signals, symlet wavelets are selected due to their periodicity and non-stationarity. These wavelets have good symmetry and orthogonality and can effectively extract the time domain and frequency domain features of the signal. The more decomposition layers there are, the richer the detail information that can be extracted, but the computational complexity will also increase. For electrical signals, decomposition of 3 layers can meet the requirements of feature extraction. Taking the voltage signal as an example, a 3-layer wavelet decomposition is performed to obtain wavelet coefficients of different frequency bands: the first layer decomposition obtains high-frequency detail coefficients and low-frequency approximation coefficients. The high-frequency detail coefficients reflect high-frequency noise and rapid change characteristics, and the low-frequency approximation coefficients reflect low-frequency trends and main features. The second layer decomposition further decomposes the low-frequency approximation coefficients to obtain a higher-level high-frequency detail coefficient and low-frequency approximation coefficient. The same applies to the third layer decomposition. Feature extraction is performed on the wavelet coefficients of each layer after decomposition, including time domain features and frequency domain features. Time domain features such as mean, variance, peak, kurtosis, skewness, etc., and frequency domain features such as frequency amplitude, energy distribution, frequency center of gravity, etc.; for example, the energy proportion of the detail coefficients of each layer is calculated as a feature reflecting the energy distribution of different frequency components of the signal; the mean and variance of the approximate coefficients are calculated as features reflecting the overall trend and fluctuation degree of the signal.

[0022] In this embodiment, multi-scale wavelet coefficients are received through the input layer of the convolutional neural network, and the wavelet coefficients of each layer are superimposed layer by layer to form a multi-dimensional feature matrix; the convolution layer sets multiple convolution kernels, and extracts features from the input multi-dimensional feature matrix through convolution operations, automatically learning features of different scales and positions, and the number and size of the convolution kernels are adjusted according to the data characteristics and model complexity. For example, the first convolution layer can set 32 ​​3×3 convolution kernels to extract preliminary local features; the second convolution layer sets 64 5×5 convolution kernels to extract more complex global features; the pooling layer uses the maximum pooling method to reduce the dimension of the feature map output by the convolution layer, reduce the amount of calculation, and retain important features. For example, a 2×2 maximum pooling layer is added after each convolution layer to halve the size of the feature map; the fully connected layer expands the pooled features into a one-dimensional vector, performs nonlinear transformation through the ReLU function to achieve feature fusion, and the output layer outputs the denoised multi-dimensional feature vector.

[0023] Step 103: Input the multi-dimensional feature vector into a dynamic assessment model, and output the risk level, failure probability prediction, and key risk factors through the dynamic assessment model; In this embodiment, the multi-dimensional feature vector is input into the dynamic assessment model, wherein the dynamic assessment model integrates the fuzzy hierarchical analysis algorithm, the LSTM network and the isolation forest algorithm; the fuzzy hierarchical analysis algorithm is used to generate the comprehensive risk basic score at the current moment, the LSTM network prediction automatically captures the long-term dependency and outputs the risk trend prediction result, and the isolation forest algorithm is used to check abnormal data points; the weighted average method is used to integrate the output results of the fuzzy hierarchical analysis algorithm, the LSTM network and the isolation forest algorithm to obtain the final risk level, failure probability prediction and key risk factors.

[0024] In this embodiment, the assessment target is decomposed into multiple levels, including the target level, the criterion level and the indicator level. The multi-dimensional eigenvector is standardized, and the standardized eigenvector is matched with the indicator level and converted into eigenvalues. The fuzzy judgment matrix is ​​constructed using triangular fuzzy numbers, and the fuzzy eigenvector method is used to calculate the fuzzy weight of each indicator. Fuzzy membership is assigned to each indicator according to the size of the eigenvalue, and the fuzzy membership of the eigenvalue is combined with the fuzzy weight. The comprehensive risk basic score at the current moment is output through the weighted summation method. The building electrical safety assessment target is divided into the target level, the criterion level and the indicator level. The target level is For the comprehensive risk assessment of building electrical safety, the criterion layer includes three aspects: electrical parameters, environmental factors and equipment status; the indicator layer is the specific evaluation indicators, such as voltage stability, current anomaly rate, ambient temperature, equipment temperature, vibration amplitude, etc.; triangular fuzzy numbers such as very important, important, relatively important, general, less important, unimportant, and very unimportant; for example, the weight of electrical parameters is 0.5, the weight of environmental factors is 0.3, and the weight of equipment status is 0.2; in the electrical parameter indicator layer, the weight of voltage stability is 0.4, the weight of current anomaly rate is 0.3, the weight of power factor is 0.3, etc.

[0025] In this embodiment, the multi-dimensional feature vectors are arranged in time series to obtain a time series feature data set with the current moment as the end point, and the first N time points are intercepted as the input window; the LSTM network captures the long-term dependencies of the data in the multi-dimensional feature vectors and predicts the risk score changes in the next 1 hour, 6 hours, and 24 hours; the predicted risk score changes are classified, and the multi-period risk trend prediction results are output, where the multi-period risk trend prediction results at least include specific score predictions and trend descriptions.

[0026] In this embodiment, 2-3 indicators are randomly selected from all indicators, and a segmentation plane is randomly generated in a multidimensional space. The path length from the root node to the leaf node of the data point in each isolated tree in the multidimensional feature vector is calculated; the path lengths of all trees are averaged to obtain the anomaly score of the current data point, and data points exceeding the threshold are marked as outliers, where the threshold is an anomaly score higher than 0.8; the characteristics of the outliers are analyzed to determine whether they are caused by equipment failure, external interference, or data collection errors, and the marked outliers are fed back to the fuzzy hierarchical analysis algorithm and the LSTM network.

[0027] In this embodiment, the LSTM network includes an input layer, an LSTM layer, a fully connected layer, and an output layer. The number of neurons in the input layer is the number of input features, such as the comprehensive risk score, historical data of each key indicator, etc.; the LSTM layer is equipped with multiple LSTM units, such as 128 or 256, to capture the long-term dependencies of the data; the number of neurons in the fully connected layer is determined according to the prediction target, such as predicting the risk level in the next hour or day, and the number of neurons in the output layer is 1.

[0028] In this embodiment, the fuzzy analytic hierarchy process (FAHP), LSTM network, and isolation forest algorithm each perform their respective functions and work together in a dynamic assessment model. FAHP scientifically generates a comprehensive risk score by quantifying the weights of various assessment indicators. LSTM predicts future electrical risk trends based on historical data, identifying potential risks in advance. Isolation forest detects abnormal data points in real time and quickly locates potential faults. The combination of these three methods not only assesses the current risk level but also predicts risk trends and identifies key risk factors. For example, if the system detects an abnormal temperature rise in a device, LSTM can predict future temperature trends, isolation forest can locate the abnormal device, and FAHP can determine the impact of the abnormality on the overall risk. This enables comprehensive, dynamic, and accurate prediction of electrical safety risks, preventing faults by hours or even days.

[0029] Step 104: Provide graded warnings for the building electrical system based on the output results of the dynamic assessment model.

[0030] In this embodiment, the risk level of the building electrical system is divided into four levels: Level 1 warning: The risk level score is 85 or higher, or a severely abnormal data point is detected, such as the device temperature exceeding 20% ​​of the rated temperature, the current suddenly increasing by more than 30%, and the LSTM predicts a sharp increase in the risk trend in the future period; Level 2 warning: 70 points ≤ Risk level score < 85 points, or moderately abnormal data points are detected, such as the device temperature exceeding the rated temperature by 10%-20%, the current fluctuating abnormally by 20%-30%, and the risk trend is increasing; Level 3 warning: 55 points ≤ Risk level score < 70 points, or a slightly abnormal data point is detected, such as the device temperature exceeding the rated temperature by 5%-10%, the current fluctuating abnormally by 10%-20%, and the risk trend is relatively stable or slightly increasing; Level 4 warning: The risk level score is <55 points, no abnormal data points are detected, and the risk trend is stable or decreasing.

[0031] In this embodiment, a visual early warning interface is used: on the large screen of the monitoring center or the client software, the risk level and key indicator data of the building electrical system are displayed in real time in the form of colors, icons, dashboards, etc. The first-level early warning is highlighted in red, the second-level early warning is displayed in orange, the third-level early warning is displayed in yellow, and the fourth-level early warning is displayed in green; SMS / email notification: when a first-level or second-level early warning occurs, the system automatically sends a SMS or email notification to the relevant management personnel, including the warning level, warning time, key risk factors and recommended measures; sound and light alarm: an sound and light alarm is set at the monitoring site. When a first-level or second-level early warning occurs, a strong sound and light signal is emitted to alert on-site personnel.

[0032] In this embodiment, the first-level warning response is to immediately trigger the emergency shutdown procedure, stop the operation of related electrical equipment, cut off the power supply, and organize professional maintenance personnel to conduct a comprehensive inspection and repair of the faulty equipment, analyze the cause of the fault, and develop a targeted solution. At the same time, a comprehensive inspection of the entire building electrical system is carried out to eliminate potential safety hazards. Level 2 early warning response: Notify relevant management and maintenance personnel to conduct key monitoring and inspections of early warning equipment, increase the frequency of inspections, analyze key risk factors, develop a maintenance plan, and arrange for repair or replacement of faulty parts as soon as possible. At the same time, closely monitor the operating status of other related equipment to prevent the fault from escalating. Level 3 Early Warning Response: Track and analyze early warning indicators to identify possible causes, such as slight equipment overload or abnormal ambient temperature and humidity. Measures such as adjusting operating parameters, increasing ventilation and heat dissipation, and cleaning equipment are taken to improve the equipment's operating environment. Warning conditions are recorded to provide a basis for subsequent equipment maintenance and system optimization. Level 4 early warning response: Continue to maintain normal monitoring and maintenance work, regularly inspect and maintain electrical equipment to ensure that the system is in a safe state.

[0033] See also Figure 2 , a schematic diagram of the structure of a building electrical safety intelligent assessment system based on Internet of Things dynamic monitoring provided by an embodiment of the present invention, the system includes: A data acquisition module for collecting multi-dimensional data in real time through an IoT sensor network deployed in the building electrical system; The feature processing module is used to perform wavelet decomposition on the collected multi-dimensional data, extract time domain and frequency domain features, use convolutional neural networks to perform feature fusion on multi-scale wavelet coefficients, remove noise, and obtain the denoised multi-dimensional feature vector; Dynamic assessment module, used to input multi-dimensional feature vectors into the dynamic assessment model, and output risk level, failure probability prediction and key risk factors through the dynamic assessment model; The graded warning module is used to provide graded warnings for the building electrical system based on the output results of the dynamic assessment model.

[0034] Figure 3 This is a schematic diagram of the structure of an intelligent building electrical safety assessment device based on dynamic monitoring using the Internet of Things (IoT) according to an embodiment of the present invention. This device 300, which can vary significantly depending on configuration or performance, may include one or more central processing units (CPUs) 310 (e.g., one or more processors), memory 320, and one or more storage media 330 (e.g., one or more mass storage devices) storing application programs 333 or data 332. The memory 320 and storage medium 330 may be either transient or persistent storage. The program stored in the storage medium 330 may include one or more modules (not shown), each of which may include a series of instructions for operating on the intelligent building electrical safety assessment device 300 based on dynamic monitoring using the IoT. Furthermore, the processor 310 may be configured to communicate with the storage medium 330, allowing the intelligent building electrical safety assessment device 300 based on dynamic monitoring using the IoT to execute the series of instructions stored in the storage medium 330 to implement the method provided in the aforementioned embodiment.

[0035] The building electrical safety intelligent assessment device 300 based on dynamic monitoring of the Internet of Things may also include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input and output interfaces 360, and / or one or more operating devices 331, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. It will be understood by those skilled in the art that Figure 3 The structure of the building electrical safety intelligent assessment device based on dynamic monitoring of the Internet of Things shown does not constitute a limitation on the computer device provided by the present invention, and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0036] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions. When the instructions are executed on a computer, the computer executes the various steps of the intelligent building electrical safety assessment method based on dynamic monitoring of the Internet of Things provided in the above-mentioned embodiments.

[0037] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices, and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0038] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0039] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely preferred examples of the present invention and are not intended to limit the present invention. Various changes and improvements may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and improvements fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent assessment method for building electrical safety based on dynamic monitoring of the Internet of Things, characterized by: The method comprises the following steps: Collect multi-dimensional data in real time through IoT sensor networks deployed in building electrical systems; Perform wavelet decomposition on the collected multi-dimensional data to extract time domain and frequency domain features. Use convolutional neural network to fuse the multi-scale wavelet coefficients and remove noise to obtain the denoised multi-dimensional feature vector. Input the multi-dimensional feature vector into the dynamic assessment model, and output the risk level, failure probability prediction and key risk factors through the dynamic assessment model; Based on the output results of the dynamic assessment model, graded early warning is given to the building electrical system.

2. The intelligent building electrical safety assessment method based on dynamic monitoring of the Internet of Things according to claim 1 is characterized in that: The method of performing wavelet decomposition on the collected multi-dimensional data and extracting time domain and frequency domain features includes: The multi-dimensional data is decomposed into three layers by Symlet wavelet to obtain wavelet coefficients of different frequency bands; The decomposed wavelet coefficients of each layer are subjected to feature extraction, including time domain features and frequency domain features, wherein the time domain features include at least mean, variance, peak, kurtosis and skewness, and the frequency domain features include at least frequency amplitude, energy distribution and frequency center of gravity.

3. The intelligent building electrical safety assessment method based on dynamic monitoring of the Internet of Things according to claim 2 is characterized in that: The method uses a convolutional neural network to perform feature fusion on multi-scale wavelet coefficients to remove noise and obtain a denoised multi-dimensional feature vector, including: The multi-scale wavelet coefficients are received through the input layer of the convolutional neural network, and the wavelet coefficients of each layer are superimposed layer by layer to form a multi-dimensional feature matrix; The convolution layer sets multiple convolution kernels, extracts features from the input multi-dimensional feature matrix through convolution operations, and automatically learns features of different scales and positions; The pooling layer uses the maximum pooling method to reduce the dimension of the feature map output by the convolutional layer. The fully connected layer expands the pooled features into a one-dimensional vector, performs nonlinear transformation through the ReLU function, and the output layer outputs the denoised multi-dimensional feature vector.

4. The intelligent building electrical safety assessment method based on dynamic monitoring of the Internet of Things according to claim 1 is characterized in that: The multi-dimensional feature vector is input into the dynamic assessment model, and the risk level, failure probability prediction and key risk factors are output through the dynamic assessment model, including: The multi-dimensional feature vector is input into a dynamic evaluation model, wherein the dynamic evaluation model integrates a fuzzy hierarchical analysis algorithm, an LSTM network, and an isolation forest algorithm; The fuzzy hierarchical analysis algorithm is used to generate a comprehensive risk-based score at the current moment. The LSTM network prediction automatically captures long-term dependencies and outputs risk trend prediction results. The isolation forest algorithm is used to check for abnormal data points. The weighted average method is used to integrate the output results of the fuzzy hierarchical analysis algorithm, LSTM network and isolation forest algorithm to obtain the final risk level, failure probability prediction and key risk factors.

5. The intelligent building electrical safety assessment method based on dynamic monitoring of the Internet of Things according to claim 4 is characterized in that: The fuzzy hierarchical analysis algorithm is used to generate the comprehensive risk basic score at the current moment, including: Decompose the evaluation target into multiple levels, including the target level, the criterion level, and the indicator level. Standardize the multi-dimensional feature vectors, and then match the standardized feature vectors with the indicator level and convert them into feature values. The fuzzy judgment matrix is ​​constructed using triangular fuzzy numbers, and the fuzzy weight of each index is calculated using the fuzzy eigenvector method. Fuzzy membership is assigned to each indicator according to the size of the eigenvalue, and the fuzzy membership of the eigenvalue is combined with the fuzzy weight to output the comprehensive risk-based score at the current moment through the weighted summation method.

6. The intelligent building electrical safety assessment method based on Internet of Things dynamic monitoring according to claim 4 is characterized in that: The LSTM network prediction automatically captures long-term dependencies and outputs risk trend prediction results, including: Arrange the multi-dimensional feature vectors in time series to obtain a time series feature dataset with the current moment as the end point and intercept the first N time points as the input window; The LSTM network captures the long-term dependencies of data in multi-dimensional feature vectors and predicts changes in risk scores over the next 1 hour, 6 hours, and 24 hours. The predicted risk score changes are classified and multi-period risk trend prediction results are output, where the multi-period risk trend prediction results at least include specific score predictions and trend descriptions.

7. The intelligent building electrical safety assessment method based on dynamic monitoring of the Internet of Things according to claim 4 is characterized in that: The use of the isolation forest algorithm to check abnormal data points includes: Randomly select 2-3 indicators from all indicators, randomly generate a split plane in the multidimensional space, and calculate the path length from the root node to the leaf node of each isolated tree in the multidimensional feature vector; The path lengths of all trees are averaged to obtain the anomaly score of the current data point, and data points exceeding the threshold are marked as anomalies, where the threshold is an anomaly score higher than 0.8; The characteristics of the anomaly points are analyzed to determine whether they are caused by equipment failure, external interference or data collection errors, and the marked anomaly points are fed back to the fuzzy hierarchical analysis algorithm and LSTM network.

8. The building electrical safety intelligent assessment system based on dynamic monitoring of the Internet of Things is characterized by: The system includes: A data acquisition module for collecting multi-dimensional data in real time through an IoT sensor network deployed in the building electrical system; The feature processing module is used to perform wavelet decomposition on the collected multi-dimensional data, extract time domain and frequency domain features, use convolutional neural networks to perform feature fusion on multi-scale wavelet coefficients, remove noise, and obtain the denoised multi-dimensional feature vector; Dynamic assessment module, used to input multi-dimensional feature vectors into the dynamic assessment model, and output risk level, failure probability prediction and key risk factors through the dynamic assessment model; The graded warning module is used to provide graded warnings for the building electrical system based on the output results of the dynamic assessment model.

9. An intelligent building electrical safety assessment device based on dynamic monitoring of the Internet of Things, characterized in that: The intelligent assessment device for building electrical safety based on dynamic monitoring of the Internet of Things includes a memory and at least one processor, wherein instructions are stored in the memory; the at least one processor calls the instructions in the memory so that the intelligent assessment device for building electrical safety based on dynamic monitoring of the Internet of Things executes each step of the intelligent assessment method for building electrical safety based on dynamic monitoring of the Internet of Things as described in any one of claims 1-7.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the various steps of the building electrical safety intelligent assessment method based on dynamic monitoring of the Internet of Things as described in any one of claims 1 to 7 are implemented.

Images