Intelligent power distribution method and system based on industrial internet

Abstract

The invention discloses an intelligent power distribution method and system based on the industrial Internet, and relates to the technical field of 123, and the method comprises the steps: collecting the operation state data of power distribution equipment, carrying out the preprocessing of the operation state data, and extracting key feature parameters based on the preprocessed operation state data; inputting the extracted key feature parameters into a pre-constructed power distribution equipment health state prediction model to obtain a power distribution equipment health state prediction value; and formulating a power distribution equipment maintenance and regulation strategy according to the power distribution equipment health state prediction value in combination with an industrial production plan. According to the method, the multi-mode sensor collects data, the missing value complementation formula and filtering processing are combined, the problems of data transmission packet loss and delay under severe working conditions are solved, and data integrity and accuracy are guaranteed. And the constructed health state prediction model is combined with a deep learning algorithm, so that the health state of the equipment can be accurately predicted, and a basis is provided for maintenance.

Description

Technical Field

[0001] This invention relates to the field of power distribution technology, specifically to an intelligent power distribution method and system based on the Industrial Internet. Background Technology

[0002] Driven by the Industry 4.0 wave, Industrial Internet technology has developed rapidly, and intelligent power distribution systems, as a key component of industrial power supply, are becoming increasingly important. Existing intelligent power distribution systems, by integrating advanced technologies such as sensing, measurement, communication, and automation, can achieve real-time monitoring and analysis of the operating status of power distribution equipment, thereby effectively improving power supply reliability and operational efficiency, and providing an important guarantee for the stable operation of modern industrial production. However, in the complex industrial environment, existing technologies still face pressing technical challenges. Especially under harsh conditions such as strong electromagnetic interference, high temperature, and high humidity, the data transmission reliability of power distribution data communication networks is significantly insufficient. For example, in industrial sites such as steel mills, the large number of harmonics generated during production can easily interfere with communication signals, leading to data packet loss, transmission delays, and other problems. These issues not only affect the accurate judgment and timely control of the operating status of power distribution equipment but may also further threaten the stability of the power supply system and the continuity of production. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides an intelligent power distribution method and system based on the Industrial Internet, which solves the problem of insufficient data transmission reliability in power distribution data communication networks under harsh conditions such as strong electromagnetic interference, high temperature and high humidity.

[0004] To achieve the above objectives, the present invention provides a smart power distribution method based on the Industrial Internet, the method comprising: S1. Collect the operating status data of the power distribution equipment, preprocess the operating status data, and extract key feature parameters based on the preprocessed operating status data; S2. Input the extracted key feature parameters into the pre-built power distribution equipment health status prediction model to obtain the predicted value of the power distribution equipment health status. S3. Based on the predicted health status of the power distribution equipment and in conjunction with the industrial production plan, formulate a maintenance and control strategy for the power distribution equipment. S4. Monitor the actual operating status of the power distribution equipment in real time and compare it with the predicted health status of the power distribution equipment. Optimize the maintenance and control strategy of the power distribution equipment based on the comparison results.

[0005] Preferably, extracting key feature parameters in step S1 includes: S101. Set the data acquisition time interval, acquire the operating status sequence of the power distribution equipment based on the multimodal sensor, and fill in the missing values ​​in the operating status sequence; S102. Filter the running state sequence; S103. Based on the filtered operating state sequence, extract the operating state features of the equipment using a feature extraction algorithm; S104. Filter the operating status characteristics of the equipment to obtain key feature parameters.

[0006] Preferably, the expression for extracting the device operating status features in step S103 is: ;in, Indicates the characteristics of equipment operating status. Indicates the length of the running state sequence. Indicates the first The running status value at each point in time. and Let represent the mean and standard deviation of the running state sequence, respectively. Indicates the first The ambient temperature value at each time point. and Let represent the mean and standard deviation of the temperature series, respectively. Indicates the weighting coefficient. The entropy value representing the sequence of running states, after normalization. in, For historical data The minimum value is usually 0; For historical data The maximum value is determined by the actual data distribution.

[0007] Preferably, the expression for filling in missing values ​​in the running state sequence in step S101 is: ;in, Indicates the first padding values ​​at the positions of missing values Indicates the first The valid value before the position of the missing value Indicates the first The valid value after each missing value position This represents the predicted value calculated based on a pre-trained time series trend prediction function. and All represent weighting coefficients, and satisfy the following conditions: .

[0008] Preferably, constructing the health status prediction model of the power distribution equipment in step S2 includes: S201. Collect historical operating status data and auxiliary data, wherein the auxiliary data includes environmental data and industrial production load data; S202. The historical operating status data and auxiliary data are preprocessed and feature extracted sequentially to obtain historical equipment health status features; S203. Construct a loss function. Based on the historical equipment health status characteristics and the loss function, construct a power distribution equipment health status prediction model using a deep learning algorithm, and train the prediction model using a gradient descent algorithm until the preset training accuracy or number of iterations is reached.

[0009] Preferably, the formulation of the power distribution equipment maintenance and control strategy in step S3 includes: S301. Based on the predicted health status of the power distribution equipment, combined with the load demand of industrial production and the operating characteristics of the equipment, comprehensively assess the maintenance needs of the equipment and obtain the assessment results. S302. Based on the assessment results, determine the maintenance priorities for different time periods or different equipment, and formulate equipment maintenance plans and control schedules based on the priorities. S303. Adjust the operating parameters of the power distribution equipment according to the equipment maintenance plan and control schedule; S304. Periodically evaluate the effectiveness of the maintenance and control strategies, and optimize the maintenance and control strategies based on the periodic evaluation results.

[0010] Preferably, determining the maintenance priority for different time periods or different devices in step S302 includes: S3021. Based on historical operating status data, industrial production load fluctuations, and equipment operating characteristics, analyze the degree of dependence of different time periods and different equipment on power supply stability, and obtain the analysis results. S3022. Based on the analysis results, the maintenance needs of different time periods and different equipment are quantitatively scored using the priority analysis method. S3023. Based on the quantitative scoring results, formulate a priority list for equipment maintenance.

[0011] Preferably, the priority analysis method in step S3022 includes: S30221. Obtain the judgment matrix ,in Representation scheme Compared to the plan The degree of importance; S30222, Normalize the judgment matrix ,in , This represents the normalized judgment matrix. Indicates the order of the judgment matrix; S30223, Based on the normalized judgment matrix Calculate the weight vector for each option. ,in Representation scheme The weights; S30224, Utilizing Consistency Indicators and random consistency ratio Perform a consistency check, if If so, then the matrix is ​​considered to have consistency.

[0012] The present invention also provides an intelligent power distribution system based on the Industrial Internet, the system comprising the intelligent power distribution method based on the Industrial Internet as described above; The system also includes: The data acquisition module is used to collect operating status data of power distribution equipment; A communication module, connected to the data acquisition module, is used to achieve reliable data transmission; A preprocessing module, connected to the communication module, is used to preprocess the running status data; A feature extraction module, connected to the preprocessing module, is used to extract key feature parameters based on the preprocessed running status data. The prediction model module, connected to the feature extraction module, is used to input the extracted key feature parameters into the pre-built power distribution equipment health status prediction model to obtain the predicted value of the power distribution equipment health status. The strategy formulation module, connected to the prediction model module, is used to formulate maintenance and control strategies for the power distribution equipment based on the predicted health status of the power distribution equipment and in conjunction with the industrial production plan. The monitoring adaptive optimization module is connected to the strategy formulation module. It is used to monitor the actual operating status of the power distribution equipment in real time, compare it with the predicted health status of the power distribution equipment, and adaptively optimize the maintenance and control strategy of the power distribution equipment based on the comparison results.

[0013] Preferably, the system further includes: An alarm module, connected to the monitoring adaptive optimization module, is used to issue an alarm signal when the deviation between the actual operating state and the predicted value exceeds a preset threshold. The historical data storage submodule is used to store historical operating status data, auxiliary data, and optimized maintenance and control strategies; The user interface submodule provides an operation interface for displaying equipment operating status, health status predictions, and maintenance and control strategies. Beneficial effects

[0014] This invention utilizes multimodal sensors to collect data, combining missing value completion formulas and filtering to solve data transmission packet loss and latency issues under harsh operating conditions, ensuring data integrity and accuracy. The constructed health status prediction model, combined with deep learning algorithms, can accurately predict equipment health status, providing a basis for maintenance. Maintenance plans are formulated through priority analysis, combined with real-time monitoring and comparison, dynamically optimizing strategies to reduce equipment failures and production interruptions, improve power supply reliability and operational efficiency, and adapt to the intelligent needs of the Industrial Internet. Attached Figure Description

[0015] Figure 1 The system structure block diagram provided by the present invention. Detailed Implementation

[0016] 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 skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Please see Figure 1 The intelligent power distribution system based on the Industrial Internet provided by this invention includes a data acquisition module, a communication module, a preprocessing module, a feature extraction module, a prediction model module, a strategy formulation module, and a monitoring and adaptive optimization module. These modules work together to complete the entire process from data acquisition to maintenance strategy optimization. Specifically, the data acquisition module is responsible for acquiring the operating status data of the power distribution equipment. The adjusted operating parameters include, but are not limited to: equipment output voltage (within ±5%), operating current threshold, cooling fan start / stop temperature, load distribution ratio, etc. For example, for equipment with a low predicted health status, its load distribution ratio can be temporarily reduced to 80% of the rated value to reduce losses, and this data is transmitted to the communication module. The communication module uses industrial Ethernet or 5G communication technology to ensure the reliability of data transmission under harsh conditions such as strong electromagnetic interference, high temperature and high humidity. Subsequently, the data is transmitted to the preprocessing module for cleaning and formatting, laying the foundation for subsequent feature extraction.

[0018] The preprocessing module first fills in missing values ​​in the collected data to improve data integrity. For example, if some data is lost due to sensor malfunction within a certain time period, it can be corrected using a formula. Complete the missing parts. Indicates the first padding values ​​at the positions of missing values and These represent the valid values ​​adjacent to the missing value. This is the predicted value calculated based on the time series trend prediction function, while and These are weighting coefficients, satisfying... The weighting coefficients α and β typically range from 0 to 0.5 and are adjusted based on the temporal correlation of the data. When adjacent data are highly correlated, α and β can be appropriately increased. In addition, all input data must first be standardized, i.e., normalized.

[0019] When the data fluctuations before and after the missing value are small, α=0.3 and β=0.4 can be chosen, in which case (1-α-β)=0.3, taking into account the influence of historical trend predictions. This method can effectively reduce the impact of missing data on subsequent analysis. Next, the preprocessing module filters the data to eliminate noise interference, thereby obtaining a cleaner running state sequence.

[0020] After receiving the preprocessed data, the feature extraction module uses a feature extraction algorithm to extract key feature parameters of the device's operating status. Specifically, this module uses a formula... The operational status characteristics of computing devices. Among them, entropy value... The uncertainty of the running state sequence is represented by the following formula: ; m is the number of different values ​​in the running state sequence. State value The probability of appearing in the sequence ( ), Indicates the characteristics of equipment operating status. The length of the running state sequence, Indicates the first The running status value at each point in time. and Let represent the mean and standard deviation of the running state sequence, respectively. Indicates the first The ambient temperature value at each time point. and Let represent the mean and standard deviation of the temperature series, respectively. The weighting coefficients are used to balance the importance of different features. This formula allows for a comprehensive consideration of the relationship between equipment operating status and environmental factors, thereby extracting more representative feature parameters.

[0021] The extracted key feature parameters are then input into the prediction model module, which constructs a power distribution equipment health status prediction model based on a deep learning algorithm. Specifically, the prediction model module first collects historical operating status data and auxiliary data (such as environmental data and industrial production load data), and preprocesses and extracts features from this data to obtain historical equipment health status features. Next, the module constructs a loss function and uses a gradient descent algorithm to train the prediction model until a preset training accuracy or number of iterations is reached. In this invention, the mean squared error (MSE) loss function is used, which is suitable for regression tasks in health status prediction, and its expression is: ;in, The number of training samples. This represents the true value of the device's health status. These are the model's predicted values. The model parameters are optimized by minimizing this loss function.

[0022] In practical applications, Long Short-Term Memory (LSTM) networks can be chosen as the core architecture of the prediction model. LSTM is chosen because its gating mechanism (input gate, forget gate, output gate) effectively solves the gradient vanishing problem of traditional Recurrent Neural Networks (RNNs), enabling it to capture long-term dependencies in the operating state sequence of power distribution equipment (such as the correlation between equipment aging trends and historical load fluctuations), thus improving long-term prediction accuracy. During model training, the learning rate is set to 0.001 (dynamically adjusted using the Adam optimizer), the batch size is 32, and the maximum number of iterations is 500. Training is stopped early when the loss function decreases by less than 1e-5 for 10 consecutive iterations to avoid overfitting. The trained prediction model can accurately predict the health status of the equipment based on the current operating state characteristics, providing a basis for subsequent maintenance and control strategies.

[0023] The strategy formulation module receives the health status prediction values ​​output by the prediction model and formulates maintenance and control strategies for power distribution equipment in conjunction with the industrial production plan. Specifically, this module first assesses the maintenance needs of the equipment based on the health status prediction values, comprehensively considers the load demand of industrial production and the operating characteristics of the equipment, and determines the maintenance priority for different time periods or different equipment. The priority determination adopts the priority analysis method, and the specific steps are as follows: First, obtain the judgment matrix. Let n be an n-order square matrix (where n is the number of schemes to be evaluated, such as different devices or different time periods), where Representation scheme Compared to the plan The degree of importance (values ​​are usually from 1 to 9 and their reciprocals, such as 1 indicating equal importance, 3 indicating slightly important, and 9 indicating extremely important); for example, comparing the maintenance priorities of 3 devices (A, B, C), the judgment matrix A can be represented as: ;in, This indicates that equipment A has a slightly higher maintenance priority than equipment B. This indicates that A is significantly higher than C.

[0024] Then, normalize the judgment matrix. First, the normalization method mentioned in this invention is used. Calculate the weight vector ( ).

[0025] Calculate the product of matrix A and weight vector W to obtain a new vector AW (an n-dimensional vector).

[0026] in Let n represent the normalized judgment matrix, where n represents the order of the judgment matrix; then, based on the normalized judgment matrix... Calculate the weight vector for each option. According to the definition of eigenvalue For each element of vector AW, calculate the ratio of the corresponding element to the weight vector W, and then take the average. The formula is: ;in Representation scheme The weights are determined; finally, the consistency index is used. and random consistency ratio Perform a consistency check; the specific calculations for the consistency check are as follows: Calculate the largest eigenvalue of judgment matrix A ; Consistency Indicators , where n is the order of the judgment matrix; Random Consistency Ratio RI is the average random consistency index, and the RI values ​​corresponding to different orders are shown in the table below.

[0027] like If the condition is met, the judgment matrix is ​​consistent; otherwise, the judgment matrix needs to be adjusted until the consistency requirement is met. This method allows for the quantification of maintenance needs for different time periods or different equipment, and the development of equipment maintenance plans and scheduling schedules accordingly.

[0028] The monitoring and adaptive optimization module is responsible for monitoring the actual operating status of power distribution equipment in real time and comparing it with predicted health status values ​​to evaluate the effectiveness of maintenance and control strategies. Specifically, this module continuously collects actual operating status data of the equipment through multimodal sensors and transmits it to the communication module. Subsequently, the module optimizes the maintenance and control strategies based on the comparison results. For example, when there is a significant deviation between the actual operating status and the predicted value, the module automatically adjusts the equipment's operating parameters or reallocates maintenance resources to ensure that the equipment is always in optimal operating condition. In addition, the module also periodically evaluates the effectiveness of the maintenance and control strategies and further optimizes the strategies based on the evaluation results, thereby achieving closed-loop control.

[0029] In practical applications, the method and system provided by this invention can be widely used in power distribution management of high-energy-consuming enterprises such as steel plants and chemical plants. Taking a large steel plant as an example, the plant has multiple power distribution devices distributed in different workshops and production lines. By deploying the intelligent power distribution system provided by this invention, the plant has achieved comprehensive monitoring and accurate prediction of the operating status of the power distribution equipment. For example, in a high-temperature and high-humidity environment, the system detects an abnormal temperature rise in a transformer and determines through a predictive model that its health status may deteriorate within the next week. Based on this, the strategy formulation module quickly formulates a maintenance plan and arranges for technicians to inspect the transformer during the off-peak production period, thereby avoiding potential equipment failures and production interruptions. In addition, the monitoring adaptive optimization module continuously optimizes maintenance and control strategies by continuously monitoring the actual operating status, significantly improving the operating efficiency and reliability of the equipment.

[0030] In summary, this invention achieves accurate prediction of the health status of power distribution equipment and intelligent optimization of maintenance strategies through the organic combination of multimodal sensors, deep learning algorithms, and priority analysis methods. Furthermore, the system's modular design gives it strong versatility and scalability, enabling it to adapt to the needs of different industries and scenarios. Practical application verification shows that this invention not only solves the problem of insufficient data transmission reliability in existing technologies but also significantly improves the operating efficiency and safety of power distribution equipment, providing strong support for intelligent power distribution management in the context of the Industrial Internet.

[0031] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0032] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An intelligent power distribution method based on industrial internet, characterized in that, The method comprises: S1, collecting power distribution equipment operating state data, and preprocessing the operating state data, and extracting key feature parameters based on the preprocessed operating state data; S2, inputting the extracted key feature parameters into a pre-constructed power distribution equipment health state prediction model to obtain a power distribution equipment health state prediction value; S3, formulating a power distribution equipment maintenance and control strategy according to the power distribution equipment health state prediction value and combining an industrial production plan; S4, monitoring the actual operating state of the power distribution equipment in real time, comparing it with the power distribution equipment health state prediction value, and optimizing the power distribution equipment maintenance and control strategy according to the comparison result.

2. The industrial internet-based intelligent power distribution method of claim 1, wherein, The extraction of the key feature parameters in step S1 comprises: S101, setting a data collection time interval, obtaining a power distribution equipment operating state sequence based on a multi-modal sensor, and completing missing values of the operating state sequence; S102, filtering the operating state sequence; S103, extracting equipment operating state features from the filtered operating state sequence using a feature extraction algorithm; S104, screening the equipment operating state features to obtain key feature parameters. 3.The smart power distribution method based on industrial internet of claim 2, wherein, The expression of the extraction of the equipment operating state features in step S103 is: ; wherein, represents a device operating state feature, represents the length of the operating state sequence, represents the operating state value at the time point, and respectively represent the mean and standard deviation of the operating state sequence, represents the ambient temperature value at the time point, and respectively represent the mean and standard deviation of the temperature sequence, represents a weight coefficient, represents the entropy value of the operating state sequence, and the normalization processing is: wherein, is the minimum value of in the historical data; is the maximum value of in the historical data.

4. The industrial internet-based intelligent power distribution method of claim 2, wherein, The expression of the completion of missing values of the operating state sequence in step S101 is: ; wherein, denotes a completed value at the th missing value position, denotes a valid value before the th missing value position, denotes a valid value after the th missing value position, denotes a predicted value calculated based on a pre-trained time series trend prediction function, and both denote a weighting coefficient, and satisfy . 5.The smart power distribution method based on industrial internet of claim 1, wherein, The construction of the power distribution equipment health state prediction model in step S2 comprises: S201, collecting historical operating state data and auxiliary data, wherein the auxiliary data comprises environmental data and industrial production load data; S202, preprocessing and feature extraction of the historical operating state data and auxiliary data in sequence to obtain historical equipment health state features; S203, constructing a loss function, constructing a power distribution equipment health state prediction model based on the historical equipment health state features and the loss function using a deep learning algorithm, and training the prediction model using a gradient descent algorithm until a preset training precision or iteration number is reached. 6.The industrial internet of things based smart power distribution method of claim 1, wherein, The formulation of the power distribution equipment maintenance and control strategy in step S3 comprises: S301, comprehensively evaluating the maintenance demand of the equipment according to the power distribution equipment health state prediction value, combining the load demand of industrial production and the equipment operating characteristics, and obtaining an evaluation result; S302, determining the maintenance priority of different time periods or different equipment based on the evaluation result, formulating an equipment maintenance plan and a control schedule based on the priority; S303, adjusting the operating parameters of the power distribution equipment according to the equipment maintenance plan and the control schedule; S304, periodically evaluating the execution effect of the maintenance and control strategy, and optimizing the maintenance and control strategy based on the periodic evaluation result. 7.The smart power distribution method based on industrial internet of claim 6, wherein, The determination of the maintenance priority of different time periods or different equipment in step S302 comprises: S3021, analyzing the degree of dependence of different time periods and different equipment on power supply stability according to historical operating state data, industrial production load fluctuation and equipment operating characteristics, and obtaining an analysis result; S3022, quantitatively scoring the maintenance demand of different time periods and different equipment based on the analysis result using a priority analysis method; S3023, based on the quantitative score results, a priority list of equipment maintenance is formulated. 8.The smart power distribution method based on industrial internet of claim 7, wherein, The priority analysis method in step S3022 includes: S30221, obtaining a judgment matrix wherein representing a scheme relative to the scheme degree of importance; S30222, normalizing the judgment matrix wherein , denotes the normalized judgment matrix, denotes the order of the judgment matrix; S30223, based on the normalized judgment matrix computing the weight vector of each scheme where denotes the weight of scheme . S30224, using the consistency index and the random consistency ratio perform a consistency check, if then determine that the matrix has consistency.

9. An intelligent power distribution system based on industrial internet, characterized in that, The system comprises the intelligent power distribution method based on industrial internet as claimed in any one of claims 1 to 8; The system further comprises: a data acquisition module for acquiring power distribution equipment operating state data; a communication module connected with the data acquisition module for realizing reliable transmission of data; a preprocessing module connected with the communication module for preprocessing the operating state data; a feature extraction module connected with the preprocessing module for extracting key feature parameters based on the preprocessed operating state data; a prediction model module connected with the feature extraction module for inputting the extracted key feature parameters into a pre-constructed power distribution equipment health state prediction model to obtain a power distribution equipment health state prediction value; a strategy formulation module connected with the prediction model module for formulating power distribution equipment maintenance and control strategies according to the power distribution equipment health state prediction value in combination with industrial production plans; a monitoring adaptive optimization module connected with the strategy formulation module for monitoring the actual operating state of the power distribution equipment in real time and comparing it with the power distribution equipment health state prediction value, and adaptively optimizing the power distribution equipment maintenance and control strategies based on the comparison results. 10.The industrial internet of things based smart power distribution system of claim 9, wherein, The system further comprises: an alarm module connected with the monitoring adaptive optimization module for issuing an alarm signal when the actual operating state deviates from the prediction value by more than a preset threshold; a historical data storage submodule for storing historical operating state data, auxiliary data and optimized maintenance and control strategies; a user interaction interface submodule for providing an operation interface for displaying equipment operating state, health state prediction value and maintenance and control strategies.

Images