A power equipment monitoring and early warning system based on intelligent patrol
By using intelligent inspection systems to acquire data, predict status, and optimize scheduling, the problems of data heterogeneity and reliance on manual operation and maintenance in power equipment monitoring systems have been solved. This enables dynamic simulation of equipment status and early warning of faults, thereby improving the safety and efficiency of power grid operation and maintenance.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- ANHUI LONGYUAN FENGLI POWER GENERATION CO LTD
- Filing Date
- 2026-05-14
- Publication Date
- 2026-07-31
AI Technical Summary
In existing power equipment monitoring systems, heterogeneous data sources make collaborative processing impossible, dynamic simulation and early warning are lacking, and operation and maintenance scheduling rely on manual experience, making it difficult to achieve full-cycle monitoring and pre-control of equipment status.
A power equipment monitoring and early warning system based on intelligent inspection is adopted, including a data acquisition module, an equipment condition deterioration prediction module, a scheduling optimization module, and a compliance collaborative intervention strategy module. Data is acquired through Modbus, TCP-IP, and OPC-UA protocols, a power equipment deterioration chain is constructed, and the optimal collaborative intervention strategy is generated.
It enables dynamic simulation of power equipment status and early fault identification, avoids conflicts of operation and maintenance resources, improves monitoring and fault prevention capabilities, and enhances the safety of power grid operation.
Smart Images
Figure CN122495694A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of power equipment monitoring technology, and in particular relates to a power equipment monitoring and early warning system based on intelligent inspection. Background Technology
[0002] The operating status of power equipment is directly related to the reliability, safety and economy of power grid supply, so the operation monitoring of power equipment is of utmost importance. However, there are still many problems in the current monitoring of power equipment: (1) The monitoring data of power equipment mainly comes from intelligent inspection equipment such as fixed inspection equipment, mobile inspection equipment and fixed sensor network. The data acquisition standards, transmission protocols and storage formats of various intelligent inspection equipment are quite different, which makes it impossible to achieve collaborative processing of power equipment monitoring data from different sources, and limits the linkage analysis and in-depth mining between data. (2) Traditional power equipment monitoring and early warning lacks dynamic simulation of equipment operating status, and cannot deduce the path, probability and stable equilibrium point of equipment status evolution from current characteristics to different potential fault modes, making it difficult to identify early hidden faults and early warning of status deterioration trends of equipment. (3) The operation and maintenance scheduling of power equipment currently relies on manual scheduling, lacking comprehensive consideration of global factors such as resources, time windows and geographical location, which easily leads to problems such as operation and maintenance resource conflicts and overlapping maintenance or troubleshooting time. At the same time, operation and maintenance intervention strategies are often formulated by operation and maintenance personnel based on personal experience, lacking scientific basis and selection of the best of multiple options.
[0003] As a result of the above-mentioned problems, the current power equipment operation and maintenance model lacks the ability to predict and anticipate equipment failures or deterioration, making it impossible to monitor and control the equipment status in a full-cycle and forward-looking manner. It is also difficult to eliminate potential equipment failure hazards in advance, and thus cannot meet the high-quality development needs of the power system for refined and proactive equipment operation and maintenance. Summary of the Invention
[0004] To address the aforementioned problems in existing technologies, this invention proposes a power equipment monitoring and early warning system based on intelligent inspection, which can significantly improve the monitoring and fault prevention capabilities of power equipment, operation and maintenance efficiency, and power grid operation safety.
[0005] To achieve the above objectives, the present invention specifically adopts the following technical solution: a power equipment monitoring and early warning system based on intelligent inspection, comprising a data acquisition module, an equipment condition deterioration prediction module, a scheduling optimization module, and a compliance collaborative intervention strategy module. The data acquisition module is connected to the equipment condition deterioration prediction module, the equipment condition deterioration prediction module is connected to the scheduling optimization module, and the scheduling optimization module is connected to the compliance collaborative intervention strategy module. The data acquisition module is used to acquire power data of power equipment monitored by fixed inspection equipment, mobile inspection equipment, and fixed sensor networks based on at least one protocol among Modbus, TCP-IP, and OPC-UA, and construct a two-dimensional data matrix with time-series points and power data as the time-series power data of the power equipment. The equipment condition deterioration prediction module is used to output a power equipment deterioration chain based on the time-series power data of the power equipment. The scheduling optimization module is used to output a scheduling scheme based on the power equipment deterioration chain. The compliance collaborative intervention strategy module is used to generate multiple sets of collaborative intervention strategies based on the scheduling scheme and output the optimal collaborative intervention strategy.
[0006] Furthermore, the equipment condition deterioration prediction module calculates the initial health of the power equipment based on the two-dimensional data matrix. The specific process is as follows: The mean of the two-dimensional data matrix is calculated column-wise, and then averaged to obtain a first mean; the standard deviation of the two-dimensional data matrix is calculated column-wise, and then averaged to obtain a first standard deviation; the first difference of the two-dimensional data matrix is calculated column-wise, and then averaged to obtain a first difference mean; the first mean is divided by 5 and compared with 1, and the smaller value is taken as the standardized mean; the first standard deviation is divided by 3 and compared with 1, and the smaller value is taken as the standardized standard deviation; the first difference mean is compared with 0, and the larger value is taken as the standardized first difference mean; the average of the standardized mean, standardized standard deviation, and standardized first difference mean is calculated; the difference between 1 and the average is calculated and compared with 0.1, and the larger value is taken as the initial health of the power equipment.
[0007] Furthermore, the equipment state degradation prediction module is also used to initialize the state matrix TM and construct a transition matrix state for multiple degradation state modes. Specifically, based on a preset time step S and the number of degradation states N, the state matrix state is initialized. The initial health of the power equipment is assigned to the element in the first row and first column of the state matrix state. The absolute value of the initial health minus 1 is divided by (N-1) and assigned to the remaining elements in the first row of the state matrix state. Non-negativity constraints and normalization constraints are applied to the elements in the state matrix state to ensure that the state matrix state... The elements in the matrix satisfy the condition that the probability is non-negative and the sum of all elements is 1; construct an N-dimensional multi-deterioration state mode transition matrix TM based on the number of deterioration states N, assign the value 0.95 to the element in the first row and first column of the transition matrix TM, and assign the value 0.05 / (N-1) to all other elements in the first row of the transition matrix TM; construct a loop with index i=2 to N, and for each loop i, assign 0.98 to the element in the i-th row and i-th column of the transition matrix TM, assign 0.01 to the element in the i-th row and 1-th column of the transition matrix TM, and assign the value 0.01 to all other elements in the transition matrix TM corresponding to this loop to 0.01 / (N-1).
[0008] Furthermore, the equipment state degradation prediction module is also used to update the transition matrix TM and iterate the state matrix state. Specifically, the process is as follows: calculate the product of the normalized first-order difference mean and 5, and take the smaller value between this product and 0.2 as the adaptive degradation rate; subtract the adaptive degradation rate from the elements of the first row and first column of the transition matrix TM to obtain the elements of the first row and first column of the updated transition matrix; divide the adaptive degradation rate by (N-1) and add it to the other elements of the first row of the transition matrix TM one by one to obtain the other elements of the first row of the updated transition matrix; set all values less than 0 in the first row of the updated transition matrix to 0, and update the first row of the updated transition matrix... The data is normalized so that the sum of all elements in the first row of the updated transition matrix is equal to 1. A loop with index t=2 to S is constructed. For each loop t, the product of the previous state matrix state(t-1) and the updated transition matrix is calculated and assigned to the current state matrix state(t). Random noise is added to the current state matrix state(t), and all negative elements are assigned the value 0. Based on the sum of the elements in the current state matrix state(t), the current state matrix state(t) is normalized so that the sum of all elements is equal to 1, and finally the updated state matrix state is obtained.
[0009] Furthermore, the specific process by which the equipment state degradation prediction module outputs the power equipment degradation chain is as follows: the updated state matrix is used as the power equipment degradation evolution path matrix; the last row of the updated state matrix is used as the power equipment degradation probability distribution prob; the position with the highest probability is found from the power equipment degradation probability distribution prob and used as the power equipment degradation equilibrium point; the power equipment degradation evolution path matrix, the power equipment degradation probability distribution prob, and the power equipment degradation equilibrium point are combined as the power equipment degradation chain and output to the scheduling optimization module.
[0010] Further, the scheduling optimization module is used to obtain the power equipment degradation probability distribution prob, total resource amount res_total, total time amount time_total, and total number of geographical locations loc_num; determine the fault location best_fault corresponding to the maximum probability value from the power equipment degradation probability distribution prob; calculate the product of the probability value prob(best_fault) corresponding to the fault location best_fault and the total resource amount res_total and round it to obtain the preliminary resource number raw_alloc; take the larger value between the preliminary resource number raw_alloc and 1 and assign it to the resource allocation amount res_alloc, and ensure that the resource allocation amount res_alloc does not exceed the total resource amount res_total; wherein, the total resource amount res_total refers to the normalized total amount of all resources available for fault handling, the total time amount time_total refers to the normalized total amount of all time available for fault handling, and the total number of geographical locations loc_num refers to the total number of geographical locations available for fault handling.
[0011] Further, the specific process of the scheduling optimization module outputting the scheduling scheme includes: calculating the product of the probability value prob(best_fault) corresponding to the fault location best_fault and the preset second proportional coefficient as the dynamic time allocation ratio time_ratio; rounding down the product of the dynamic time allocation ratio time_ratio and the total time amount time_total to obtain the initial allocation time rawtime_alloc; taking the larger value of the initial allocation time rawtime_alloc and 1 and assigning it to the allocation time time_alloc, ensuring that the allocation time time_alloc does not exceed the total time amount time_total; calculating the remainder of (best_fault-1) and the total number of geographical locations loc_num, and then adding 1 to the remainder to obtain the geographical location allocation loc_alloc, so that the geographical location allocation loc_alloc is mapped to a reasonable position between 1 and loc_num; and combining the resource allocation amount res_alloc, the allocation time time_alloc, and the geographical location allocation loc_alloc to form the scheduling scheme, and outputting it to the compliance collaborative intervention strategy module.
[0012] Furthermore, the compliance collaborative intervention strategy module generates four sets of collaborative intervention strategies based on the resource allocation amount res_alloc, allocation time time_alloc, and geographical location allocation loc_alloc in the obtained scheduling scheme: Strategy 1: [max(res_alloc-2,1), time_alloc-3, loc_alloc, IT1]; Strategy 2: [res_alloc-1, time_alloc-2, loc_alloc, IT2]; Strategy 3: [res_alloc, time_alloc-1, loc_alloc, IT3]; Strategy 4: [max(res_alloc-1,1), time_alloc, loc_alloc, IT4]; Among them, IT1 is the first intervention intensity; IT2 is the second intervention intensity; IT3 is the third intervention intensity; and IT4 is the fourth intervention intensity.
[0013] Furthermore, max(res_alloc-2, 1) is used to constrain the resource usage of Strategy 1 to be no less than 1, avoiding zero or negative resource usage; max(res_alloc-1, 1) is used to constrain the resource usage of Strategy 4 to be no less than 1, ensuring the effectiveness of resource allocation; IT1, IT2, IT3, and IT4 are used to characterize the priority execution level of the corresponding candidate collaborative intervention strategies. The larger the value, the higher the priority of the corresponding collaborative intervention strategy.
[0014] Furthermore, the specific process by which the compliance collaborative intervention strategy module outputs the optimal collaborative intervention strategy is as follows: obtain the intervention strategy matrix composed of the four sets of collaborative intervention strategies, the power equipment degradation evolution path matrix, and the power equipment degradation equilibrium point; initialize the score list score_list, the length of which is consistent with the number of collaborative intervention strategies; Each collaborative intervention strategy is evaluated individually: the intervention intensity of the current m-th collaborative intervention strategy is extracted, and the degradation inhibition rate (inhibit) of the m-th collaborative intervention strategy is calculated based on the formula inhibit = 0.3 + 0.6 × strategies(m, 4), where strategies(m, 4) is the intervention intensity of the m-th strategy; the resource usage and execution time of the current m-th collaborative intervention strategy are extracted, and the average of the resource utilization rate and time utilization rate is calculated to obtain the overall efficiency. The resource utilization rate is the ratio of the resource usage of the current m-th collaborative intervention strategy to the maximum resource usage among all candidate collaborative intervention strategies, and the time utilization rate is the execution time of the current m-th collaborative intervention strategy. The ratio of execution time to the maximum execution time among all candidate collaborative intervention strategies is used. The comprehensive score of the current m-th collaborative intervention strategy is calculated using a weighted summation formula: score_list(m) = inhibit × 0.7 + efficiency × 0.3, where the inhibition rate (inhibit) has a weight of 0.7 and the overall efficiency (efficiency) has a weight of 0.3. After evaluation, the highest score (best_score) and the corresponding strategy ID (best_idx) are extracted from the score_list. The intervention strategy corresponding to the optimal strategy ID (best_idx) is output as the optimal collaborative intervention strategy.
[0015] The beneficial technical effects of this invention compared with the prior art are as follows: 1. Based on the real-time operating conditions of power equipment, dynamically deduce the path, probability, and stable equilibrium point of the equipment status evolving into potential failure modes and output accurate degradation chains. This overcomes the shortcomings of traditional monitoring methods in that data cannot be processed collaboratively, and enables early identification of hidden faults and early warning of degradation trends. 2. Resource pre-allocation is completed with constraints of resources, time and geographical location to avoid scheduling conflicts and resource waste, thereby effectively forming a reasonable and feasible scheduling framework; 3. Automatically generate multiple sets of compliant collaborative intervention strategies, and construct a power equipment monitoring and operation and maintenance architecture that includes early warning and identification of power equipment degradation trends and optimization of collaborative intervention strategies. This can significantly improve the monitoring and fault prevention capabilities of power equipment, operation and maintenance efficiency, and power grid operation safety. Attached Figure Description
[0016] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary, and those skilled in the art can derive other embodiments based on the provided drawings without creative effort.
[0017] Figure 1 This is a structural block diagram of a power equipment monitoring and early warning system based on intelligent inspection, according to the present invention. Figure 2 This is a simplified flowchart of the workflow of a power equipment monitoring and early warning system based on intelligent inspection, according to the present invention. Detailed Implementation
[0018] 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.
[0019] Example Combination Figures 1-2 This invention proposes a power equipment monitoring and early warning system based on intelligent inspection, including a data acquisition module, an equipment condition deterioration prediction module, a scheduling optimization module, and a compliance collaborative intervention strategy module. These modules are linked and interconnected: the data acquisition module is connected to the equipment condition deterioration prediction module, the equipment condition deterioration prediction module is connected to the scheduling optimization module, and the scheduling optimization module is connected to the compliance collaborative intervention strategy module. This realizes a complete closed-loop management process from data collection, condition prediction, scheduling optimization to intervention strategy generation and screening, which can effectively predict the probability of power equipment failure and reduce the risk of power outage losses and equipment damage caused by failures.
[0020] Figure 2This paper presents a simplified workflow diagram of a power equipment monitoring and early warning system based on intelligent inspection, as described in this invention. The data acquisition module acquires power data from power equipment monitored by fixed inspection devices, mobile inspection devices, and fixed sensor networks using at least one protocol among Modbus, TCP-IP, and OPC-UA. This power data can be voltage, current, vibration, or temperature data of the power equipment. A two-dimensional data matrix is then constructed using time-series points and the power data, serving as the time-series power data for the power equipment. The equipment condition deterioration prediction module outputs a power equipment deterioration chain based on the time-series power data. The scheduling optimization module outputs a scheduling scheme based on the power equipment deterioration chain. The compliant collaborative intervention strategy module generates multiple sets of collaborative intervention strategies based on the scheduling scheme and outputs the optimal collaborative intervention strategy.
[0021] Furthermore, the equipment condition deterioration prediction module calculates the initial health of the power equipment based on a two-dimensional data matrix. The specific process is as follows: the mean of the two-dimensional data matrix is calculated column by column, and then the average is calculated to obtain the first mean. The first mean reflects the overall operating level of the power equipment monitoring indicators; the standard deviation of the two-dimensional data matrix is calculated column by column, and then the average is calculated to obtain the first standard deviation; the first difference of the two-dimensional data matrix is calculated column by column, and then the average is calculated to obtain the first difference mean to reflect the changing trend of the monitoring data. To eliminate differences in the dimensions of different indicators and avoid the influence of single indicator bias on the calculation results, the first mean, first standard deviation, and first-order difference mean are standardized. Specifically: the first mean is divided by 5 and compared with 1, and the smaller value is taken as the standardized mean. 5 is the normalization coefficient, and taking a smaller value avoids distortion of the standardized results caused by extreme data. The first standard deviation is divided by 3 and compared with 1, and the smaller value is taken as the standardized standard deviation, ensuring that the standardized data is in the 0-1 range, which conforms to the quantitative requirements of health. The first-order difference mean is compared with 0, and the larger value is taken as the standardized first-order difference mean. This filters out negative interference in data fluctuations and focuses on specific deterioration trends in equipment status. Finally, the arithmetic mean of the standardized mean, standardized standard deviation, and standardized first-order difference mean is calculated. The difference between 1 and this mean is calculated and compared with 0.1, and the larger value is taken as the initial health of the power equipment. Here, the difference between 1 and the mean transforms the data fluctuation trend into a quantifiable health indicator.
[0022] The aforementioned equipment condition degradation prediction module is also used to initialize the state matrix TM and construct a transition matrix state for multiple degradation state modes. Specifically, based on a preset time point S and the number of degradation states N, the state matrix state is initialized to record the probability distribution of each degradation state of the power equipment at different time points. The length of the aforementioned time point S can be 1000 points, used to simulate and predict the dynamic evolution of the power equipment's degradation within S time points. The number of degradation states N can be set to 4 levels, corresponding to the normal state, slightly degraded state, moderately degraded state, and severely degraded state of the power equipment, respectively.
[0023] The initial health of the aforementioned power equipment is assigned to the element in the first row and first column of the state matrix `state`. The absolute value of the initial health minus 1 is divided by (N-1) and then assigned to the remaining elements in the first row of the state matrix `state`. Non-negativity and normalization constraints are applied to the elements in the state matrix `state` to ensure that the elements in the state matrix `state` have non-negative probabilities and that the sum of all elements is 1. An N-dimensional transition matrix `TM` of multiple deterioration state modes is constructed based on the number of deterioration states `N` to characterize the transition probabilities between different deterioration states. The element in the first row and first column of the transition matrix TM is assigned the value 0.95, and the other elements in the first row of the transition matrix TM are all assigned the value 0.05 / (N-1), so as to evenly distribute the probability of transitioning from the initial normal state to each deteriorated state; construct a loop with the index i=2 to N. For each loop i, assign 0.98 to the element in the i-th row and i-th column of the transition matrix TM, assign 0.01 to the element in the i-th row and 1-th column of the transition matrix TM, and assign the other elements in the transition matrix TM corresponding to this loop to the value 0.01 / (N-1).
[0024] The aforementioned equipment condition degradation prediction module is also used to update the transition matrix TM and iterate the state matrix state. The specific process is as follows: Calculate the product of the standardized first-order difference mean and 5, and take the smaller value between this product and 0.2 as the adaptive degradation rate. The standardized first-order difference mean reflects the data fluctuation trend. Calculating the adaptive degradation rate based on this allows the degradation rate to match the actual operating fluctuations of the equipment, avoiding the limitations of a fixed degradation rate. Taking the smaller value between this product and 0.2 limits the upper limit of the degradation rate, preventing the degradation rate from being predicted too quickly. Subtract the adaptive degradation rate from the elements in the first row and first column of the transition matrix TM, and obtain the elements in the first row and first column of the updated transition matrix. Divide the adaptive degradation rate by (N-1), and add it one by one to the other elements in the first row of the transition matrix TM, obtaining the other elements in the first row of the updated transition matrix. All elements in the first row of the updated transition matrix that are less than 0... The value is set to 0, and the data in the first row of the updated transition matrix is normalized so that the sum of all elements in the first row of the updated transition matrix is equal to 1; a loop with index t=2 to S is constructed. For each loop t, the product of the previous state matrix state(t-1) and the updated transition matrix is calculated and assigned to the current state matrix state(t). Thus, the probability of deteriorated state under different time steps is iteratively updated through matrix multiplication; random noise, such as slight Gaussian noise, is added to the current state matrix state(t), and all negative elements are set to 0; based on the sum of the elements of the current state matrix state(t), the current state matrix state(t) is normalized so that the sum of all elements is equal to 1, and finally the updated state matrix state is obtained.
[0025] The specific process by which the above-mentioned equipment condition deterioration prediction module outputs the power equipment deterioration chain is as follows: The updated state matrix is used as the power equipment deterioration evolution path matrix. This matrix completely records the probability distribution changes of the N deterioration states of the equipment at S time points, clearly presenting the entire process of the equipment's deterioration evolution trajectory from the initial state to the current moment, which is convenient for subsequent deterioration prediction; The last row of the updated state matrix is used as the power equipment deterioration probability distribution prob, which corresponds to the deterioration state probability at the final time step, accurately reflecting the current and predicted short-term deterioration state distribution of the equipment; The position with the highest probability is found in the power equipment deterioration probability distribution prob, which is used as the power equipment deterioration equilibrium point. This equilibrium point is the predicted most likely deterioration state of the equipment; The power equipment deterioration evolution path matrix, the power equipment deterioration probability distribution prob, and the power equipment deterioration equilibrium point are used together as the power equipment deterioration chain and output to the scheduling optimization module.
[0026] The aforementioned scheduling optimization module is used to obtain the power equipment degradation probability distribution prob, total resource quantity res_total, total time quantity time_total, and total number of geographical locations loc_num; determine the fault location best_fault corresponding to the maximum probability value from the power equipment degradation probability distribution prob; calculate the product of the probability value prob(best_fault) corresponding to the fault location best_fault and the total resource quantity res_total and round it to obtain the preliminary resource quantity raw_alloc; take the larger value between the preliminary resource quantity raw_alloc and 1, assign it to the resource allocation quantity res_alloc, and ensure that the resource allocation quantity res_alloc does not exceed the total resource quantity res_total.
[0027] Among them, total resource quantity res_total refers to the normalized total amount of resources available for fault handling. For example, if the resources allocated for fault handling are 5 maintenance personnel, 2 portable testing devices, and 10 sets of spare parts, the fault handling resources are at a high level, and the normalized total resource quantity res_total can be 0.8. Total time quantity time_total refers to the normalized total amount of time available for fault handling. For example, assuming the system specifies a total available time of 24 hours for fault handling, the normalized total time quantity time_total can be 0.6, indicating that there is sufficient time available for fault handling to support the complete maintenance process of routine faults. Total number of geographical locations loc_num refers to the total number of geographical locations available for fault handling, corresponding to the regional locations of power equipment, providing range constraints for subsequent geographical location allocation.
[0028] The specific process of the scheduling optimization module outputting the scheduling scheme includes: calculating the product of the probability value prob(best_fault) corresponding to the fault location best_fault and a preset second proportional coefficient as the dynamic time allocation ratio time_ratio, where the value of the second proportional coefficient ranges from 0.5 to 0.9 and can be adjusted according to the urgency of the power equipment fault; the higher the fault risk, the larger the second proportional coefficient; rounding the product of the dynamic time allocation ratio time_ratio and the total time amount time_total to obtain the initial allocation time rawtime_alloc; and setting the initial allocation time rawtime_alloc... The larger of oc and 1 is assigned to the allocation time time_alloc, ensuring that the allocation time time_alloc does not exceed the total time amount time_total; the remainder between (best_fault-1) and the total number of geographical locations loc_num is calculated, and then 1 is added to the remainder to obtain the geographical location allocation loc_alloc, so that the geographical location allocation loc_alloc is mapped to a reasonable position between 1 and loc_num; the resource allocation amount res_alloc, the allocation time time_alloc, and the geographical location allocation loc_alloc are combined to form a standardized scheduling scheme, which is then output to the compliance collaborative intervention strategy module.
[0029] The aforementioned compliance collaborative intervention strategy module generates four sets of collaborative intervention strategies based on the resource allocation amount res_alloc, allocation time time_alloc, and geographical location allocation loc_alloc in the obtained scheduling scheme: Strategy 1: [max(res_alloc-2,1), time_alloc-3, loc_alloc, IT1]; Strategy 2: [res_alloc-1, time_alloc-2, loc_alloc, IT2]; Strategy 3: [res_alloc, time_alloc-1, loc_alloc, IT3]; Strategy 4: [max(res_alloc-1,1), time_alloc, loc_alloc, IT4]; Wherein, IT1 is the first intervention intensity; IT2 is the second intervention intensity; IT3 is the third intervention intensity; and IT4 is the fourth intervention intensity. IT1 is preferably 0.6, IT2 is preferably 0.75, IT3 is preferably 0.9, and IT4 is preferably 0.85. The above `max(res_alloc-2, 1)` is used to constrain the resource usage of Strategy 1 to be no less than 1, avoiding zero or negative resource usage. `max(res_alloc-1, 1)` is used to constrain the resource usage of Strategy 4 to be no less than 1, ensuring the effectiveness of resource allocation. The above `IT1`, `IT2`, `IT3`, and `IT4` are used to characterize the priority execution level of the corresponding candidate collaborative intervention strategies; the larger the value, the higher the priority of the corresponding collaborative intervention strategy.
[0030] The resource allocation amount `res_alloc` represents the normalized value of the resources pre-allocated for a certain strategy in the scheduling scheme. These pre-allocated resources include human resources (including the number of schedulable maintenance shifts, the total number of maintenance personnel, etc.) and material resources (including drones, maintenance equipment, remote communication equipment, etc.). The allocation time `time_alloc` represents the normalized value of the allowed execution time pre-allocated for a certain strategy in the scheduling scheme, which is the time window width for executing the intervention strategy. The geographic location allocation `loc_alloc` represents the normalized value of the geographic location spatial resources pre-allocated for a certain strategy in the scheduling scheme, including the area to be inspected, the total commuting distance, etc. Strategy 1 to Strategy 4 represent four different complete scheduling strategies. Taking Strategy 1 as an example, `res_alloc-2` allocates 2 units less resource than the pre-allocated amount, `time_alloc-3` shortens the time window by 3 units compared to the pre-allocated time window, `loc_alloc` keeps the pre-allocated geographic location spatial resources unchanged, and `IT1` indicates the priority level and intervention intensity of Strategy 1. Each strategy group includes four indicators: resource usage, execution time, geographical location, and intervention intensity. The geographical location allocation (loc_alloc) remains constant, while resource usage and execution time are adjusted in a gradient to adapt to fault handling needs at different levels of degradation: for mild degradation and low risk, a strategy with low resource usage and short execution time is selected to save resources and time; for severe degradation and high risk, a strategy with sufficient resource usage and long execution time is selected to ensure intervention effectiveness. The four intervention intensities characterize the execution priority and intervention strength of the corresponding strategy, and can be flexibly adjusted by those skilled in the art based on the actual equipment degradation and on-site execution conditions.
[0031] The specific process by which the above-mentioned compliance collaborative intervention strategy module outputs the optimal collaborative intervention strategy is as follows: First, obtain the intervention strategy matrix composed of four sets of collaborative intervention strategies, the power equipment degradation evolution path matrix output by the equipment condition degradation prediction module, and the power equipment degradation equilibrium point; then initialize the score list score_list, the length of which is consistent with the number of collaborative intervention strategy groups; then, use a step-by-step evaluation method to conduct a comprehensive and accurate quantitative evaluation of each set of collaborative intervention strategies, the specific steps of which are as follows: The first step is to extract the intervention intensity of the current m-th collaborative intervention strategy, where m ranges from 1 to 4, corresponding to the four strategies. strategies(m, 4) is the fourth dimension parameter of the m-th strategy, corresponding to IT1-IT4 above. Based on the formula inhibit=0.3+0.6×strategies(m, 4), the degradation inhibition rate inhibit of the m-th collaborative intervention strategy is calculated. In the above formula, 0.3 is the basic inhibition rate, and 0.6 is the weighting coefficient of the intervention intensity. The weighting coefficient is used to amplify the influence of the intervention intensity on the inhibition effect.
[0032] The second step involves extracting the resource usage and execution time of the current m-th collaborative intervention strategy, calculating the resource utilization rate and time utilization rate respectively, and then taking their arithmetic mean to obtain the overall efficiency. Specifically, the resource utilization rate is the ratio of the resource usage of the current m-th collaborative intervention strategy to the maximum resource usage among all candidate collaborative intervention strategies, and the time utilization rate is the ratio of the execution time of the current m-th collaborative intervention strategy to the maximum execution time among all candidate collaborative intervention strategies.
[0033] The third step involves calculating the comprehensive score of the current m-th group of collaborative intervention strategies using a weighted summation formula: score_list(m) = inhibit × 0.7 + efficiency × 0.3. After evaluating all group strategies in a loop, the highest score (best_score) and its corresponding strategy ID (best_idx) are extracted from the score_list. It's understood that if multiple highest scores exist, a priority can be preset to ensure the selected strategy ID is unique. Finally, the intervention strategy corresponding to the optimal strategy ID (best_idx) is output as the optimal collaborative intervention strategy. This ensures that the output strategy balances equipment degradation and rational resource utilization, achieving optimal intervention effect and execution efficiency.
[0034] The embodiments and / or implementation methods described above are merely preferred embodiments and / or implementation methods for implementing the technology of the present invention, and are not intended to limit the implementation methods of the technology of the present invention in any way. Any person skilled in the art can make some modifications or alterations to other equivalent embodiments without departing from the scope of the technical means disclosed in the content of the present invention, but they should still be regarded as the technology or embodiments that are substantially the same as the present invention.
Claims
1. A power equipment monitoring and early warning system based on intelligent inspection, characterized in that, The system includes a data acquisition module, an equipment condition deterioration prediction module, a scheduling optimization module, and a compliance collaborative intervention strategy module. The data acquisition module is connected to the equipment condition deterioration prediction module, the equipment condition deterioration prediction module is connected to the scheduling optimization module, and the scheduling optimization module is connected to the compliance collaborative intervention strategy module. The data acquisition module is used to acquire power data of power equipment monitored by fixed inspection equipment, mobile inspection equipment, and fixed sensor networks based on at least one protocol among Modbus, TCP-IP, and OPC-UA, and to construct a two-dimensional data matrix with time-series points and power data as the time-series power data of the power equipment. The equipment condition deterioration prediction module is used to output the power equipment deterioration chain based on the time-series power data of the power equipment. The scheduling optimization module is used to output a scheduling scheme based on the power equipment degradation chain. The compliance collaborative intervention strategy module is used to generate multiple sets of collaborative intervention strategies based on the scheduling scheme and output the optimal collaborative intervention strategy.
2. The power equipment monitoring and early warning system based on intelligent inspection according to claim 1, characterized in that, The equipment condition deterioration prediction module calculates the initial health of the power equipment based on the two-dimensional data matrix. The specific process is as follows: The mean of the two-dimensional data matrix is calculated column-wise, and then averaged to obtain the first mean; the standard deviation of the two-dimensional data matrix is calculated column-wise, and then averaged to obtain the first standard deviation; the first difference of the two-dimensional data matrix is calculated column-wise, and then averaged to obtain the first difference mean; the first mean is divided by 5 and compared with 1, and the smaller value is taken as the standardized mean; the first standard deviation is divided by 3 and compared with 1, and the smaller value is taken as the standardized standard deviation; the first difference mean is compared with 0, and the larger value is taken as the standardized first difference mean; the average of the standardized mean, standardized standard deviation, and standardized first difference mean is calculated; the difference between 1 and the average is calculated and compared with 0.1, and the larger value is taken as the initial health of the power equipment.
3. A power equipment monitoring and early warning system based on intelligent inspection according to claim 2, characterized in that, The equipment condition degradation prediction module is also used to initialize the state matrix TM and construct a transition matrix state for multiple degradation state modes. The specific process is as follows: based on the preset time step S and the number of degradation states N, initialize the state matrix state, assign the initial health of the power equipment to the element in the first row and first column of the state matrix state, divide the absolute value of the initial health minus 1 by (N-1), and assign it to the remaining elements in the first row of the state matrix state; apply non-negativity constraints and normalization constraints to the elements in the state matrix state so that the elements in the state matrix state satisfy the conditions of non-negativity and the sum of all elements is 1. Construct an N-dimensional multi-deterioration state mode transition matrix TM based on the number of deterioration states N. Assign the value 0.95 to the element in the first row and first column of the transition matrix TM, and assign the value 0.05 / (N-1) to all other elements in the first row of the transition matrix TM. Construct a loop with index i=2 to N. For each loop i, assign the value 0.98 to the element in the i-th row and i-th column of the transition matrix TM, assign the value 0.01 to the element in the i-th row and 1-th column of the transition matrix TM, and assign the value 0.01 to all other elements in the transition matrix TM corresponding to this loop.
4. A power equipment monitoring and early warning system based on intelligent inspection according to claim 3, characterized in that, The equipment state degradation prediction module is also used to update the transition matrix TM and iterate the state matrix state. The specific process is as follows: calculate the product of the normalized first-order difference mean and 5, and take the smaller value of the product and 0.2 as the adaptive degradation rate; subtract the adaptive degradation rate from the element of the first row and first column of the transition matrix TM to obtain the element of the first row and first column of the updated transition matrix. Divide the adaptive degradation rate by (N-1) and add it to the other elements of the first row of the transition matrix TM one by one to obtain the other elements of the first row of the updated transition matrix. Set all values less than 0 in the first row of the updated transition matrix to 0, and normalize the data in the first row of the updated transition matrix so that the sum of all elements in the first row of the updated transition matrix equals 1; construct a loop with indices t=2 to S. For each loop t, calculate the product of the previous state matrix state(t-1) and the updated transition matrix, and assign it to the current state matrix state(t); add random noise to the current state matrix state(t) and set all negative elements to 0; perform a normalization operation on the current state matrix state(t) based on the sum of the elements of the current state matrix state(t) so that the sum of all elements equals 1, and finally obtain the updated state matrix state.
5. A power equipment monitoring and early warning system based on intelligent inspection according to claim 4, characterized in that, The specific process of the equipment condition deterioration prediction module outputting the power equipment deterioration chain is as follows: the updated state matrix is used as the power equipment deterioration evolution path matrix; the last row of the updated state matrix is used as the power equipment deterioration probability distribution prob. Find the position with the highest probability from the power equipment degradation probability distribution prob, and take it as the power equipment degradation equilibrium point; take the power equipment degradation evolution path matrix, the power equipment degradation probability distribution prob and the power equipment degradation equilibrium point together as the power equipment degradation chain, and output it to the scheduling optimization module.
6. A power equipment monitoring and early warning system based on intelligent inspection according to claim 5, characterized in that, The scheduling optimization module is used to obtain the power equipment degradation probability distribution prob, total resources res_total, total time time, and total number of geographical locations loc_num; determine the fault location best_fault corresponding to the maximum probability from the power equipment degradation probability distribution prob; calculate the product of the probability value prob(best_fault) corresponding to the fault location best_fault and the total resources res_total and round it to obtain the preliminary resource number raw_alloc; take the larger value between the preliminary resource number raw_alloc and 1 and assign it to the resource allocation amount res_alloc, ensuring that the resource allocation amount res_alloc does not exceed the total resources res_total; wherein, the total resources res_total refers to the normalized total amount of all resources available for fault handling, the total time time time refers to the normalized total amount of all time available for fault handling, and the total number of geographical locations loc_num refers to the total number of geographical locations available for fault handling.
7. A power equipment monitoring and early warning system based on intelligent inspection according to claim 6, characterized in that, The specific process by which the scheduling optimization module outputs the scheduling scheme includes: The product of the probability value prob(best_fault) corresponding to the fault location best_fault and the preset second proportional coefficient is calculated as the dynamic time allocation ratio time_ratio; the product of the dynamic time allocation ratio time_ratio and the total time amount time_total is rounded down to obtain the initial allocation time rawtime_alloc; the larger value of the initial allocation time rawtime_alloc and 1 is assigned to the allocation time time_alloc, ensuring that the allocation time time_alloc does not exceed the total time amount time_total; the remainder of (best_fault-1) and the total number of geographical locations loc_num is calculated, and then the remainder is added to 1 to obtain the geographical location allocation loc_alloc, so that the geographical location allocation loc_alloc is mapped to a reasonable position between 1 and the total number of geographical locations loc_num; the resource allocation amount res_alloc, the allocation time time_alloc, and the geographical location allocation loc_alloc are combined to form the scheduling scheme, and output to the compliance collaborative intervention strategy module.
8. A power equipment monitoring and early warning system based on intelligent inspection according to claim 7, characterized in that, The compliance collaborative intervention strategy module generates four sets of collaborative intervention strategies based on the resource allocation amount res_alloc, allocation time time_alloc, and geographical location allocation loc_alloc in the obtained scheduling scheme: Strategy 1: [max(res_alloc-2,1), time_alloc-3, loc_alloc, IT1]; Strategy 2: [res_alloc-1, time_alloc-2, loc_alloc, IT2]; Strategy 3: [res_alloc, time_alloc-1, loc_alloc, IT3]; Strategy 4: [max(res_alloc-1,1), time_alloc, loc_alloc, IT4]; Among them, IT1 is the first intervention intensity; IT2 is the second intervention intensity; IT3 is the third intervention intensity; and IT4 is the fourth intervention intensity.
9. A power equipment monitoring and early warning system based on intelligent inspection according to claim 8, characterized in that, The max(res_alloc-2, 1) is used to constrain the resource usage of strategy one to be no less than 1, so as to avoid the resource usage being zero or negative; the max(res_alloc-1, 1) is used to constrain the resource usage of strategy four to be no less than 1, so as to ensure the effectiveness of resource allocation. IT1, IT2, IT3, and IT4 are used to characterize the priority level of the corresponding candidate collaborative intervention strategies. The larger the value, the higher the priority of the corresponding collaborative intervention strategy.
10. A power equipment monitoring and early warning system based on intelligent inspection according to claim 8, characterized in that, The specific process by which the compliance collaborative intervention strategy module outputs the optimal collaborative intervention strategy is as follows: obtain the intervention strategy matrix composed of the four sets of collaborative intervention strategies, the power equipment degradation evolution path matrix, and the power equipment degradation equilibrium point; initialize the score list score_list, the length of which is consistent with the number of collaborative intervention strategies; Each collaborative intervention strategy is evaluated individually: the intervention intensity of the current m-th collaborative intervention strategy is extracted, and the degradation inhibition rate (inhibit) of the m-th collaborative intervention strategy is calculated based on the formula inhibit = 0.3 + 0.6 × strategies(m, 4), where strategies(m, 4) is the intervention intensity of the m-th collaborative intervention strategy; the resource usage and execution time of the current m-th collaborative intervention strategy are extracted, and the average of the resource utilization rate and time utilization rate is calculated to obtain the overall efficiency. The resource utilization rate is the ratio of the resource usage of the current m-th collaborative intervention strategy to the maximum resource usage among all candidate collaborative intervention strategies, and the time utilization rate is the ratio of the execution time of the current m-th collaborative intervention strategy to the maximum execution time among all candidate collaborative intervention strategies. The comprehensive score of the current m-th collaborative intervention strategy is calculated based on a weighted summation formula: score_list(m) = inhibit × 0.7 + efficiency × 0.3, where the inhibition rate has a weight of 0.7 and the comprehensive efficiency has a weight of 0.
3. After evaluation, the highest score best_score and the corresponding strategy number best_idx are extracted from the score_list. The intervention strategy corresponding to the optimal strategy number best_idx is output as the optimal collaborative intervention strategy.