Smart park infrastructure intelligent monitoring system

Abstract

The invention discloses a smart park infrastructure intelligent monitoring system, which relates to the technical field of intelligent monitoring, comprises a fusion module, a prediction module and an optimization module, and is characterized in that state related parameters are set, feature data are obtained through fusion, a state prediction model is obtained through training, and the state prediction model is optimized. The fusion module fuses the feature values of the multiple state related parameters, reduces the complexity and delay of data processing, improves the monitoring efficiency through automatic data processing and model optimization, fuses the feature values of the multiple related parameters through the fusion module, avoids the limitation of a single parameter, and improves the monitoring accuracy. The state prediction model is optimized, model parameters can be dynamically adjusted according to actual operation data, the prediction precision of the model is further improved, potential fault risks can be warned in advance, measures can be taken in time for maintenance and repair, the fault occurrence probability is reduced, and the reliability and stability of power supply facilities are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent monitoring, and particularly to an intelligent monitoring system for the infrastructure of a smart park. Background Art

[0002] In recent years, with the continuous progress of technology, the intelligent monitoring technology for the infrastructure of a smart park will develop towards a more intelligent direction, and the Internet of Things technology plays an important role in the monitoring of the infrastructure of a smart park. By deploying a large number of intelligent sensors, comprehensive perception and real-time monitoring of the park environment, equipment, etc. can be achieved. Edge computing nodes can process and analyze the data collected by front-end devices in real time and quickly respond to various demands in the park, such as security monitoring, energy management, etc.

[0003] Currently, in the Chinese invention patent with the publication number CN115457471A, a monitoring method and system for a smart park are disclosed. This method analyzes the video data, the facility operation data after classification processing, and the environmental data of each area in the park environment information, and divides the fire protection levels of the park fire protection facilities, can know the changes in emergencies in the park in real time, brings a certain degree of danger, improves safety, and is more intelligent. However, in the related technology, the operating state of the power supply network is not analyzed and predicted based on historical data, lacking the scientific nature of data utilization and the persuasiveness of prediction, not fusing data of different dimensions to obtain characteristic data, which is not conducive to the simplicity of operation, and at the same time is not conducive to the rapid responsiveness of system prediction, and does not update the prediction model according to the evaluation criteria calculated in real time, which is not conducive to the dynamic responsiveness of monitoring prediction. Summary of the Invention

[0004] The technical problem solved by the present invention is that in the related technology, the operating state of the power supply network is not analyzed and predicted based on historical data, lacking the scientific nature of data utilization and the persuasiveness of prediction, not fusing data of different dimensions to obtain characteristic data, which is not conducive to the simplicity of operation, and at the same time is not conducive to the rapid responsiveness of system prediction, and does not update the prediction model according to the evaluation criteria calculated in real time, which is not conducive to the dynamic responsiveness of monitoring prediction.

[0005] To solve the above technical problems, the present invention provides the following technical solution: An intelligent monitoring system for the infrastructure of a smart park, including a fusion module, a prediction module, and an optimization module;

[0006] The fusion module sets state-related parameters, fuses the characteristic values of the state-related parameters, and obtains characteristic data;

[0007] The prediction module trains a decision tree based on the characteristic data to obtain a state prediction model;

[0008] The optimization module optimizes the state prediction model, and predicts the state of the power supply facilities according to the optimized state prediction model.

[0009] As a preferred solution of the intelligent monitoring system for the infrastructure of a smart park according to the present invention, wherein: the historical related data of the power supply facilities includes historical voltage, historical current, historical temperature of power switch tubes, historical temperature of transformers, historical total harmonic distortion rate of voltage, historical total harmonic distortion rate of current, and historical state of power supply facilities;

[0010] The historical state of the power supply facilities includes a normal state and an abnormal state.

[0011] Count the number of times of historical voltage fluctuation exceeding the limit according to the historical voltage.

[0012] Set the historical voltage, historical current, historical temperature of power switch tubes, historical temperature of transformers, historical total harmonic distortion rate of voltage, and historical total harmonic distortion rate of current as state-related parameters.

[0013] As a preferred solution of the intelligent monitoring system for the infrastructure of a smart park according to the present invention, wherein: the statistical logic of the number of times of historical voltage fluctuation exceeding the limit includes:

[0014] Calculate the first difference between any two adjacent historical voltages, select the absolute value of the first difference, set the first value as the voltage fluctuation threshold, compare the absolute value of the first difference with the first value. When the first value is less than or equal to the first value, jump to the next set of adjacent historical voltages, and repeat the steps of taking the difference and comparing. When the absolute value of the first difference is greater than the first value, perform a weighted calculation of the number of times of historical voltage fluctuation exceeding the limit and 1 to obtain a first sum value. The first sum value is used as the value for weighted calculation with 1 when the absolute value of the next first difference is greater than the first value, and update the first sum value according to the new sum value obtained by each weighting.

[0015] As a preferred solution of the intelligent monitoring system for the infrastructure of a smart park according to the present invention, wherein: the fusion module preprocesses the state-related parameters, and the preprocessing includes cleaning processing and outlier removal processing;

[0016] The cleaning processing is performed through a data cleaning program;

[0017] The logic of the outlier removal processing includes:

[0018] Select any type of state-related parameter, calculate the average value and standard deviation of the state-related parameter, calculate the first difference between each state-related parameter and the average value of the state-related parameter respectively, set three times the value of the standard deviation as the second value, compare the first difference with the second value, when the first difference is greater than the second value, set the state-related parameter as an outlier, remove the state-related parameter and jump to the next state-related parameter, when the first difference is less than or equal to the second value, set the state parameter as a normal value, retain the state-related parameter and jump to the next state-related parameter;

[0019] Set the average value of the state-related parameters after preprocessing as the corresponding eigenvalue.

[0020] As a preferred solution of an intelligent monitoring system for the infrastructure of a smart park according to the present invention, wherein: the eigenvalues are fused through a fusion function, and the expression of the fusion function is:

[0021] Y == w1x1 + w2x2 +... + w n x n ;

[0022] wherein, Y is the feature data, x n is the eigenvalue, w n is the fusion feature parameter coefficient, and the sum of each fusion feature parameter coefficient is 1;

[0023] Optimize and match the fusion feature parameter coefficients according to the particle swarm algorithm.

[0024] As a preferred solution of an intelligent monitoring system for the infrastructure of a smart park according to the present invention, wherein: the logic for optimizing and matching the fusion feature parameter coefficients includes:

[0025] Set the number of particle swarms, the range of particle movement speed, the number of iterations, the inertia coefficient, randomly set the optimal position, construct the objective fitness function, calculate the objective fitness function values of each particle, compare the objective fitness function value of the particle with the corresponding best position, when the objective fitness function value of the particle is less than the corresponding best position, set the objective fitness function value of the particle as the new best position, when the objective fitness function value of the particle is greater than or equal to the corresponding best position, the best position remains unchanged, until the number of iterations is satisfied, the optimization ends, and the optimal fusion feature parameter coefficients are obtained.

[0026] As a preferred solution of an intelligent monitoring system for the infrastructure of a smart park according to the present invention, wherein: the calculation expression of the objective fitness function is:

[0027]

[0028] where p is the value of the target fitness function, and are respectively the average value of the eigenvalue in the normal state and the average value of the eigenvalue in the abnormal state, and are respectively the standard deviation of the eigenvalue in the normal state and the standard deviation of the eigenvalue in the abnormal state. The smaller the p value, the higher the sensitivity of the fusion feature parameter to the state judgment.

[0029] As a preferred solution of the intelligent monitoring system for the infrastructure of a smart park according to the present invention, wherein: the training logic of the state prediction model includes:

[0030] Set the normal state to 1, set the abnormal state to 0, normalize the feature data so that the feature data is distributed between 0 and 1, divide the normalized feature data into a test set and a training set according to the first ratio, and select a decision tree as the basic model. The decision tree includes a normal state decision tree and an abnormal state decision tree;

[0031] Set the decision tree parameters. The decision tree parameters include the learning rate, the maximum depth of the tree, the number of trees, and the minimum number of samples for node splitting. Train the tree model one by one according to the set number of trees. In each iteration, calculate the residual, which is expressed as the negative gradient of the loss function and the current predicted value. Train a new decision tree according to the current pseudo-residual, weight the predicted value of the newly trained tree with the initial estimate, and update the initial estimate. Repeat the above training process until all decision trees are trained, and then stop the iteration;

[0032] Set the superposition combination of the normal state decision tree model and the abnormal state decision tree after the iteration as the state prediction model, and obtain the state of the power supply facility by inputting the feature data into the state prediction model.

[0033] As a preferred solution of the intelligent monitoring system for the infrastructure of a smart park according to the present invention, wherein: the logic for optimizing the state prediction model includes:

[0034] Set the situation where the prediction is correct in the normal state as TP, the situation where the prediction is incorrect in the normal state as FP, the situation where the prediction is correct in the abnormal state as TN, and the situation where the prediction is correct in the abnormal state as FN. Calculate the precision, recall, and overall accuracy of the state prediction model for the normal state and the abnormal state respectively. Set the third value, the fourth value, and the fifth value as the precision threshold, the recall threshold, and the overall accuracy threshold respectively. Compare the precision, recall, and overall accuracy with the third value, the fourth value, and the fifth value respectively. When the precision is less than or equal to the third value, or the recall is less than or equal to the fourth value, or the overall accuracy threshold is less than or equal to the fifth value, reset the decision tree parameters and retrain the decision tree to obtain an optimized state prediction model. When the precision is greater than the third value, the recall is greater than the fourth value, and the overall accuracy threshold is greater than the fifth value, do not optimize the state prediction model.

[0035] As a preferred solution of the intelligent monitoring system for the infrastructure of the smart park described in the present invention, wherein: the calculation expression of the precision is:

[0036] Precision=TP / TP+FP;

[0037] The calculation expression of the recall is:

[0038] Recall=TP / TP+FN;

[0039] The calculation expression of the overall accuracy is:

[0040] Accruracy=TP+TN / TP+TN+FP+FN;

[0041] Among them, Precision, Recall, and Accruracy represent precision, recall, and overall accuracy respectively.

[0042] Advantages of the present invention: The fusion module fuses the feature values of multiple status-related parameters, enabling rapid integration of information from different sensors or data sources, reducing the complexity and latency of data processing, thereby achieving real-time monitoring of the status of power supply facilities. Through automated data processing and model optimization, the need for manual intervention is reduced, the monitoring efficiency is improved, and the labor cost is lowered. By fusing the feature values of multiple relevant parameters through the fusion module, the status of power supply facilities can be more comprehensively reflected, avoiding the limitations of a single parameter, thereby improving the accuracy of prediction. Optimizing the status prediction model can dynamically adjust the model parameters according to actual operation data, further enhancing the prediction accuracy of the model. Through the accurate status prediction model, the system can give early warnings of potential fault risks, take timely measures for maintenance and repair, reduce the probability of faults occurring, and improve the reliability and stability of power supply facilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 FIG. is a schematic diagram of the basic process of an intelligent monitoring system for the infrastructure of a smart park provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0044] In order to make the above objects, features, and advantages of the present invention more obvious and understandable, the following detailed description of the specific embodiments of the present invention will be given in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of them.

[0045] Embodiment, referring to Figure 1 , which is an embodiment of the present invention, provides an intelligent monitoring system for the infrastructure of a smart park, including a fusion module, a prediction module, and an optimization module;

[0046] The fusion module sets status-related parameters and fuses the feature values of the status-related parameters to obtain feature data;

[0047] The prediction module trains a decision tree based on the feature data to obtain a status prediction model;

[0048] The optimization module optimizes the status prediction model and predicts the status of the power supply facilities according to the optimized status prediction model.

[0049] The fusion module of the present invention fuses the eigenvalues of multiple state-related parameters, can quickly integrate information from different sensors or data sources, reduce the complexity and latency of data processing, and thus achieve real-time monitoring of the status of power supply facilities. Through automated data processing and model optimization, the need for manual intervention is reduced, the monitoring efficiency is improved, and the labor cost is reduced. By fusing the eigenvalues of multiple relevant parameters through the fusion module, the status of power supply facilities can be more comprehensively reflected, avoiding the limitations of single parameters, and thus improving the accuracy of prediction. Optimizing the status prediction model can dynamically adjust the model parameters according to the actual operation data, further improving the prediction accuracy of the model. Through the accurate status prediction model, the system can give early warnings of potential fault risks, take measures for maintenance and repair in a timely manner, reduce the probability of faults occurring, and improve the reliability and stability of power supply facilities.

[0050] The historical relevant data of power supply facilities includes historical voltage, historical current, historical temperature of power switch tubes, historical temperature of transformers, historical total harmonic distortion rate of voltage, historical total harmonic distortion rate of current, and historical status of power supply facilities;

[0051] The historical status of power supply facilities includes normal status and abnormal status

[0052] Statistically count the number of times the historical voltage fluctuation exceeds the limit according to the historical voltage;

[0053] Set the historical voltage, historical current, historical temperature of power switch tubes, historical temperature of transformers, historical total harmonic distortion rate of voltage, and historical total harmonic distortion rate of current as state-related parameters.

[0054] The statistical logic of the number of times the historical voltage fluctuation exceeds the limit includes:

[0055] Calculate the first difference between any two adjacent historical voltages, select the absolute value of the first difference, set the first value as the voltage fluctuation threshold, compare the absolute value of the first difference with the first value. When the first value is less than or equal to the first value, jump to the next set of adjacent historical voltages and repeat the steps of taking the difference and comparing. When the absolute value of the first difference is greater than the first value, perform a weighted calculation of the number of times the historical voltage fluctuation exceeds the limit and 1 to obtain the first sum value. The first sum value is used as the value for weighted calculation with 1 when the absolute value of the next first difference is greater than the first value, and update the first sum value according to the new sum value obtained by each weighting.

[0056] In specific implementation, by calculating the first difference between any two adjacent historical voltages, the change of voltage can be accurately identified. The first value is set as the voltage fluctuation threshold, and the threshold can be flexibly set according to actual needs to adapt to different voltage fluctuation situations. The absolute value of the first difference is compared with the first value. When the absolute value of the first difference is less than or equal to the first value, it jumps to the next group of adjacent historical voltages, improving the statistical efficiency. The first sum value is updated according to the new sum value obtained by each weighting, ensuring the dynamics and accuracy of the statistical results. By effectively monitoring and recording voltage fluctuation over-limit events, voltage anomalies can be timely detected and processed, improving the stability and reliability of the system. By counting the number of historical voltage fluctuation over-limit times, the rules and trends of voltage fluctuations are analyzed, providing data support for system maintenance and optimization.

[0057] The fusion module preprocesses the state-related parameters, and the preprocessing includes cleaning processing and outlier removal processing;

[0058] The cleaning processing is carried out through a data cleaning program;

[0059] The logic of outlier removal processing includes:

[0060] Select any type of state-related parameter, calculate the average value and standard deviation of the state-related parameter, calculate the first difference between each state-related parameter and the average value of the state-related parameter respectively, set three times the value of the standard deviation as the second value, compare the first difference with the second value. When the first difference is greater than the second value, set the state-related parameter as an outlier, remove the state-related parameter and jump to the next state-related parameter. When the first difference is less than or equal to the second value, set the state parameter as a normal value, retain the state-related parameter and jump to the next state-related parameter;

[0061] Set the average value of the state-related parameters after preprocessing as the corresponding eigenvalue.

[0062] The eigenvalues are fused through a fusion function, and the expression of the fusion function is:

[0063] Y == w1x1 + w2x2 +... + w n x n ;

[0064] Among them, Y is the feature data, x n is the eigenvalue, w n is the fusion feature parameter coefficient, and the sum value of each fusion feature parameter coefficient is 1;

[0065] Optimize and match the fusion feature parameter coefficients according to the particle swarm algorithm.

[0066] In specific implementation, cleaning the status-related parameters through a data cleaning program can remove the noise and redundant information in the data, improve the neatness and consistency of the data. By calculating the average value and standard deviation of the status-related parameters, and identifying the values exceeding 3 times the standard deviation as outliers for elimination, it can effectively remove the extreme values and incorrect data in the data, reduce the impact of these data on subsequent analysis and modeling. The preprocessed data better conforms to the assumption conditions of the model, enables the model to have better generalization ability on new data, and reduces the risk of overfitting. High-quality data can provide more accurate results for system monitoring and early warning, reduce false alarms and missed alarms, and enhance users' trust in the system. Through automated outlier identification and elimination, it reflects the intelligent level of the system and can automatically adapt to different data environments.

[0067] The logic for optimizing and matching the coefficients of the fusion feature parameters includes:

[0068] Set the number of particle swarms, the range of particle movement speed, the number of iterations, the inertia coefficient, randomly set the optimal position, construct the objective fitness function, calculate the objective fitness function values of each particle, compare the objective fitness function value of the particle with the corresponding best position. When the objective fitness function value of the particle is less than the corresponding best position, set the objective fitness function value of the particle as the new best position. When the objective fitness function value of the particle is greater than or equal to the corresponding best position, the best position remains unchanged. When the number of iterations is satisfied, the optimization ends and the optimal coefficients of the fusion feature parameters are obtained.

[0069] The calculation expression of the objective fitness function is:

[0070]

[0071] where p is the value of the objective fitness function, and are the average values of the characteristic values in the normal state and the average values of the characteristic values in the abnormal state respectively, and are the standard deviations of the characteristic values in the normal state and the standard deviations of the characteristic values in the abnormal state respectively. The smaller the p value, the higher the sensitivity of the fusion feature parameters to state judgment.

[0072] In specific implementation, by setting an appropriate number of particle swarms and exploring multiple possible solution spaces simultaneously, the efficiency of parameter optimization is improved. By restricting the movement speed of particles, it is ensured that particles move effectively in the solution space, avoiding the influence of too fast or too slow search speeds on the optimization efficiency. By setting an appropriate number of iterations, it is ensured that the algorithm has enough time to converge to the optimal solution. By randomly initializing the optimal positions of particles, the diversity of solutions is increased, avoiding the algorithm falling into local optimal solutions. By adjusting the inertia coefficient, the global search and local search capabilities of particles are balanced, improving the convergence speed and accuracy of the algorithm. By constructing an objective fitness function, the solution quality of particles is quantified, guiding the particles to move towards the optimal solution direction. The particle swarm optimization algorithm automatically finds the optimal fusion feature parameter coefficients, reducing the complexity of manually setting parameters. The particle swarm optimization algorithm is an intelligent optimization method that can effectively find the optimal solution by simulating the foraging behavior of bird flocks, reflecting the intelligent level of the system.

[0073] The training logic of the state prediction model includes:

[0074] Set the normal state as 1, set the abnormal state as 0, normalize the feature data so that the feature data is distributed between 0 and 1, divide the normalized feature data into a test set and a training set according to the first ratio, and select a decision tree as the base model. The decision tree includes a normal state decision tree and an abnormal state decision tree;

[0075] Set the decision tree parameters. The decision tree parameters include the learning rate, the maximum depth of the tree, the number of trees, and the minimum number of samples for node splitting. Train the tree models one by one according to the set number of trees. In each iteration, calculate the residual, which is expressed as the negative gradient of the loss function and the current predicted value. Train a new decision tree according to the current pseudo-residual, weight the predicted value of the newly trained tree with the initial estimate, and update the initial estimate. Repeat the above training process until all decision trees are trained, and then stop the iteration;

[0076] Set the superposition combination of the normal state decision tree model and the abnormal state decision tree after iteration as the state prediction model. By inputting the feature data into the state prediction model, the state of the power supply facilities can be obtained.

[0077] In specific implementation, the feature data is normalized so that the feature data is distributed between 0 and 1, which can eliminate the dimensional difference between different features and improve the training effect of the model. By calculating the residual (the negative gradient of the loss function and the current predicted value) and training a new decision tree based on the residual, the prediction error of the model can be gradually corrected, and the prediction accuracy of the model can be improved. By superimposing and combining the prediction results of multiple decision trees, the complex relationships in the data can be captured, and the generalization ability of the model can be improved. By reasonably setting the decision tree parameters (such as the learning rate, the maximum depth of the tree, the number of trees, and the minimum number of samples for node splitting), the complexity of the model can be controlled, and the robustness of the model can be improved. By using the state prediction model to monitor the state of the power supply facilities in real time, problems can be discovered and processed in a timely manner, and the user experience can be improved.

[0078] The logic for optimizing the state prediction model includes:

[0079] Set the situation where the prediction is correct in the normal state as TP, the situation where the prediction is incorrect in the normal state as FP, the situation where the prediction is correct in the abnormal state as TN, and the situation where the prediction is incorrect in the abnormal state as FN. Calculate the precision, recall, and overall accuracy of the state prediction model for the normal state and the abnormal state respectively. Set the third value, the fourth value, and the fifth value as the precision threshold, the recall threshold, and the overall accuracy threshold respectively. Compare the precision, recall, and overall accuracy with the third value, the fourth value, and the fifth value respectively. When the precision is less than or equal to the third value or the recall is less than or equal to the fourth value or the overall accuracy threshold is less than or equal to the fifth value, reset the decision tree parameters and retrain the decision tree to obtain the optimized state prediction model. When the precision is greater than the third value and the recall is greater than the fourth value and the overall accuracy threshold is greater than the fifth value, do not optimize the state prediction model.

[0080] The calculation expression for precision is:

[0081] Precision = TP / (TP + FP);

[0082] The calculation expression for recall is:

[0083] Recall = TP / (TP + FN);

[0084] The calculation expression for the overall accuracy is:

[0085] Accruracy = (TP + TN) / (TP + TN + FP + FN);

[0086] Among them, Precision, Recall, and Accruracy represent precision, recall, and overall accuracy respectively.

[0087] In specific implementation, by comparing the precision, recall rate, and overall accuracy with preset thresholds (the third value, the fourth value, and the fifth value), it is possible to dynamically determine whether the model needs to be optimized. When the performance indicators of the model are lower than the preset thresholds, the decision tree parameters (such as the learning rate, the maximum depth of the tree, the number of trees, and the minimum number of samples for node splitting) are reset, and the decision tree is retrained, which can effectively improve the performance of the model. This optimization logic can adaptively adjust the model parameters according to different data characteristics and application scenarios to ensure the best performance of the model under different conditions. By optimizing the model to ensure that the precision, recall rate, and overall accuracy are all higher than the preset thresholds, the prediction accuracy of the model for normal and abnormal states can be improved, false alarms and missed alarms can be reduced, and the reliability of the system can be enhanced. The optimization logic is designed and optimized independently of other modules, which is convenient for the expansion and upgrade of the system.

[0088] The fusion module of the present invention fuses the feature values of multiple state-related parameters, which can quickly integrate information from different sensors or data sources, reduce the complexity and latency of data processing, and thus achieve real-time monitoring of the state of power supply facilities. Through automated data processing and model optimization, the need for manual intervention is reduced, the monitoring efficiency is improved, and the labor cost is lowered. By fusing the feature values of multiple relevant parameters through the fusion module, the state of power supply facilities can be more comprehensively reflected, avoiding the limitations of a single parameter, thereby improving the prediction accuracy. Optimizing the state prediction model can dynamically adjust the model parameters according to the actual operation data, further enhancing the prediction accuracy of the model. Through the precise state prediction model, the system can give early warnings of potential fault risks, take timely measures for maintenance and repair, reduce the probability of faults occurring, and improve the reliability and stability of power supply facilities.

[0089] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. Among them, the storage medium is implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured article including an instruction device, and the instruction device implements the process Figure 1 one process or multiple processes and / or blocks Figure 1 the functions specified in one block or multiple blocks.

[0090] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that modifications or equivalent replacements of the technical solutions of the present invention, without departing from the spirit and scope of the technical solutions of the present invention, should be covered by the scope of the claims of the present invention.

Claims

1. An intelligent monitoring system for the infrastructure of an intelligent park, characterized in that, It includes a fusion module, a prediction module, and an optimization module; The fusion module sets state-related parameters, fuses the eigenvalue of the state-related parameters, and obtains feature data; The prediction module trains a decision tree based on the feature data to obtain a state prediction model; The optimization module optimizes the state prediction model and predicts the state of the power supply facility according to the optimized state prediction model.

2. The intelligent monitoring system for the infrastructure of a smart park according to claim 1, characterized in that: The historical-related data of the power supply facility includes historical voltage, historical current, historical power switch tube temperature, historical transformer temperature, historical voltage total harmonic distortion rate, historical current total harmonic distortion rate, and historical power supply facility state; The historical power supply facility state includes a normal state and an abnormal state Count the number of times the historical voltage fluctuates beyond the limit according to the historical voltage; Set the historical voltage, historical current, historical power switch tube temperature, historical transformer temperature, historical voltage total harmonic distortion rate, and historical current total harmonic distortion rate as state-related parameters.

3. The intelligent monitoring system for the infrastructure of an intelligent park according to claim 2, wherein: The statistical logic of the number of times the historical voltage fluctuates beyond the limit includes: Calculate the first difference between any two adjacent historical voltages, select the absolute value of the first difference, set the first value as the voltage fluctuation threshold, compare the absolute value of the first difference with the first value, when the first value is less than or equal to the first value, jump to the next set of adjacent historical voltages, and repeat the steps of taking the difference and comparing. When the absolute value of the first difference is greater than the first value, perform a weighted calculation of the number of times the historical voltage fluctuates beyond the limit and 1 to obtain a first sum value. The first sum value is used for the value to be weighted with 1 when the absolute value of the first difference is greater than the first value next time, and update the first sum value according to the new sum value obtained by each weighting.

4. The intelligent monitoring system for the infrastructure of an intelligent park according to claim 1, wherein: The fusion module preprocesses the state-related parameters, and the preprocessing includes cleaning processing and outlier removal processing; The cleaning processing is performed through a data cleaning program; The logic of the outlier removal processing includes: Select any type of state-related parameter, calculate the average value and standard deviation of the state-related parameter, calculate the first difference between each state-related parameter and the average value of the state-related parameter, set three times the value of the standard deviation as the second value, compare the first difference with the second value, when the first difference is greater than the second value, set the state-related parameter as an outlier, remove the state-related parameter and jump to the next state-related parameter, when the first difference is less than or equal to the second value, set the state parameter as a normal value, retain the state-related parameter and jump to the next state-related parameter; Set the average value of the state-related parameters after preprocessing as the corresponding eigenvalue.

5. An intelligent monitoring system for the infrastructure of a smart park according to claim 1, characterized in that: Fuse the eigenvalues through a fusion function, and the expression of the fusion function is: Y == w1x1 + w2x2 +... + w n x n ; Among them, Y is the characteristic data, and x n is the eigenvalue, and w n is the coefficient of the fusion feature parameter, and the sum of the coefficients of each fusion feature parameter is 1; Optimize and match the fusion feature parameter coefficients according to the particle swarm algorithm.

6. The intelligent monitoring system for the infrastructure of a smart park according to claim 5, characterized in that: The logic of optimizing and matching the fusion feature parameter coefficients includes: Set the number of particle swarms, the range of particle movement speed, the number of iterations, the inertia coefficient, randomly set the optimal position, construct the objective fitness function, calculate the objective fitness function values of each particle, compare the objective fitness function value of the particle with the corresponding best position. When the objective fitness function value of the particle is less than the corresponding best position, set the objective fitness function value of the particle as the new best position. When the objective fitness function value of the particle is greater than or equal to the corresponding best position, the best position remains unchanged. When the number of iterations is satisfied, the optimization ends, and the optimal fusion feature parameter coefficient is obtained.

7. The intelligent monitoring system for the infrastructure of an intelligent park as described in claim 1, characterized in that: The calculation expression of the said objective fitness function is: where p is the value of the target fitness function, and are the average values of the eigenvalue in the normal state and the average value of the eigenvalue in the abnormal state, respectively, and are the standard deviations of the eigenvalue in the normal state and the standard deviation of the eigenvalue in the abnormal state, respectively. The smaller the p value, the higher the sensitivity of the fusion feature parameter to the state judgment.

8. An intelligent monitoring system for the infrastructure of a smart park according to claim 1, characterized in that: The training logic of the state prediction model includes: Set the normal state as 1, set the abnormal state as 0, normalize the feature data so that the feature data is distributed between 0 and 1, divide the normalized feature data into a test set and a training set according to the first ratio, select the decision tree as the basic model, and the decision tree includes a normal state decision tree and an abnormal state decision tree; Set the decision tree parameters, the said decision tree parameters include the learning rate, the maximum depth of the tree, the number of trees, and the minimum number of samples for node splitting. Train the tree model one by one according to the set number of trees. In each iteration, calculate the residual, and the said residual is expressed as the negative gradient of the loss function and the current predicted value. Train a new decision tree according to the current pseudo-residual, weight the predicted value of the newly trained tree with the initial estimate, and update the initial estimate. Loop the above training process until all decision trees are trained, and then stop the iteration; Set the superposition combination of the normal state decision tree model and the abnormal state decision tree after the iteration as the state prediction model, and obtain the state of the power supply facilities by inputting the feature data into the state prediction model.

9. The intelligent monitoring system for the infrastructure of an intelligent park according to claim 8, characterized in that: The logic for optimizing the state prediction model includes: Set the situation where the prediction is correct in the normal state as TP, set the situation where the prediction is wrong in the normal state as FP, set the situation where the prediction is correct in the abnormal state as TN, set the situation where the prediction is correct in the abnormal state as FN, calculate the precision rate, recall rate, and overall accuracy rate of the state prediction model for the normal state and the abnormal state respectively, set the third value, the fourth value, and the fifth value as the precision rate threshold, recall rate threshold, and overall accuracy rate threshold respectively, compare the precision rate, recall rate, and overall accuracy rate with the third value, the fourth value, and the fifth value respectively. When the precision rate is less than or equal to the third value or the recall rate is less than or equal to the fourth value or the overall accuracy rate threshold is less than or equal to the fifth value, reset the decision tree parameters and retrain the decision tree to obtain the optimized state prediction model. When the precision rate is greater than the third value and the recall rate is greater than the fourth value and the overall accuracy rate threshold is greater than the fifth value, do not optimize the state prediction model.

10. The intelligent monitoring system for the infrastructure of a smart park according to claim 9, characterized in that: The calculation expression of the said precision rate is: Precision = TP / (TP + FP); The calculation expression of the said recall rate is: Recall = TP / (TP + FN); The calculation expression of the said overall accuracy rate is: Accruracy = TP + TN / TP + TN + FP + FN; Among them, Precision, Recall, and Accruracy represent precision rate, recall rate, and overall accuracy respectively.

Images