Method and system for mining operation data and evaluating and optimizing performance of environmental protection equipment

Abstract

The invention provides an environmental protection equipment operation data mining and performance evaluation optimization method and system, and relates to the technical field of environmental protection equipment, and the method comprises the steps: collecting real-time operation data, constructing a nonlinear annealing iteration model, carrying out the high-dimensional state coding processing, building a component state correlation characteristic spectrum, and calculating a fault propagation link and a degradation trend curve. A fault risk assessment matrix is constructed, equipment state evolution is simulated in a digital twin space, a maintenance strategy is generated and optimized, and intelligent maintenance plan making and model online optimization are realized. The operation reliability of the environment-friendly equipment can be improved, the service life of the equipment is prolonged, and maintenance cost is reduced.

Description

Technical Field

[0001] The present invention relates to environmental protection equipment technology, and in particular to a method and system for mining environmental protection equipment operation data and optimizing performance evaluation. Background Art

[0002] With increasing environmental protection requirements, various types of environmental protection equipment are playing an increasingly important role in industrial production and urban management. The stable operation of environmental protection equipment, such as wastewater treatment equipment, exhaust gas purification devices, and solid waste treatment systems, is directly related to the achievement of environmental protection goals. Environmental protection equipment typically consists of multiple component systems, with complex interactions and dependencies between components. Over the long-term operation of these devices, components will experience varying degrees of degradation and failure, affecting overall performance and environmental effectiveness. Traditional environmental protection equipment maintenance relies primarily on a passive model of regular inspections and post-failure repairs, which is difficult to address the coordinated degradation of multiple components in complex systems.

[0003] With the development of IoT technology and big data analysis methods, the collection and analysis of environmental protection equipment operating data has become increasingly important, and data-driven equipment performance evaluation and optimization methods are gaining increasing attention. Existing technologies typically use statistical analysis and machine learning to monitor and predict equipment operating status, and formulate maintenance strategies based on this information. The rise of digital twin technology provides a new technical means for virtual simulation and optimization of environmental protection equipment, enabling the evaluation of the effectiveness of different maintenance strategies in a virtual environment.

[0004] Most existing technologies use linear models or simple machine learning algorithms to process environmental protection equipment operation data, which cannot effectively capture the nonlinear correlation between components and complex fault propagation mechanisms, resulting in insufficient understanding of the overall system degradation process and the inability to accurately identify key influencing factors and root causes of faults.

[0005] Traditional environmental equipment performance evaluation methods usually only focus on the status monitoring of individual components, lack a systematic analysis of the fault propagation links between components, and are unable to accurately assess the chain reaction of component failures on the overall system performance, making it difficult to determine maintenance priorities and the optimal time for intervention.

[0006] Existing maintenance strategy optimization methods are often static and difficult to adapt to the dynamic changes of environmental protection equipment under different operating conditions. They lack an adaptive optimization mechanism based on real-time feedback, resulting in irrational allocation of maintenance resources and the inability to achieve the optimal balance between environmental protection equipment performance and operating costs. Summary of the Invention

[0007] The embodiments of the present invention provide a method and system for mining environmental protection equipment operation data and optimizing performance evaluation, which can solve the problems in the prior art.

[0008] A first aspect of an embodiment of the present invention provides a method for mining environmental protection equipment operation data and optimizing performance evaluation, comprising: Collecting real-time operating data of environmental protection equipment, constructing a nonlinear annealing iterative model, performing high-dimensional state encoding processing on the real-time operating data, using the entanglement mapping effect to establish a component state correlation feature map, calculating the fault propagation link between components based on the component state correlation feature map, combining a probability measurement circuit to generate a degradation trend curve for each component, and calculating a dynamic risk index based on the fault propagation link and the degradation trend curve; Based on the dynamic risk index and the fault propagation link, a component-level fault risk assessment matrix is ​​constructed, wherein the fault risk assessment matrix records the risk propagation weight, health status index, and remaining life prediction value of each component; Constructing a physical simulation environment for environmental protection equipment in the digital twin space, importing the fault risk assessment matrix and the fault propagation link into the physical simulation environment, performing state evolution on the physical simulation environment using the nonlinear annealing iterative model, calculating equipment performance change trajectories under different maintenance strategies based on the risk propagation weights, and generating a risk reduction index for each maintenance strategy; Maintenance strategies are prioritized according to the risk reduction index, the maintenance strategy with the highest risk reduction index is selected as the optimal maintenance strategy, and an intelligent maintenance plan is generated based on the optimal maintenance strategy; the fault propagation link and the risk propagation weight are updated based on status feedback data during the execution of the intelligent maintenance plan, and the nonlinear annealing iterative model is optimized online.

[0009] Performing high-dimensional state encoding processing on the real-time operation data, establishing a component state correlation feature map using an entanglement mapping effect, and calculating an inter-component fault propagation link based on the component state correlation feature map includes: Performing nonlinear decomposition on the real-time operation data using Hilbert-Huang transform to extract intrinsic mode components, calculating an instantaneous phase function based on the intrinsic mode components, and obtaining a characteristic sequence based on the instantaneous phase function; Performing a Hilbert-Huang secondary transform on the feature sequence to obtain a high-dimensional feature vector of the feature sequence, and mapping the high-dimensional feature vector to a multidimensional state space to generate a high-dimensional state vector; Performing an entangled state mapping transformation on the high-dimensional state vector to construct an entangled state vector, calculating a reduced density matrix of the entangled state vector, extracting eigenvalues ​​of the reduced density matrix, and calculating the entanglement correlation between components based on the eigenvalues; constructing a dynamic association matrix between components according to the entanglement correlation degree, optimizing and decomposing the dynamic association matrix between components using the Hilbert-Huang transform, and generating a component state association map, wherein the component state association map records the association strength between components; Performing Hilbert-Huang spectrum analysis on the correlation strength in the component state correlation map to calculate a characteristic spectrum sequence of the component state, quantifying the fault propagation strength between components based on the characteristic spectrum sequence, and constructing a fault propagation network; An importance index of each propagation path is calculated according to the propagation intensity in the fault propagation network, and an inter-component fault propagation link is identified based on the importance index.

[0010] Constructing a dynamic association matrix between components according to the entanglement correlation degree, optimizing and decomposing the dynamic association matrix between components using the Hilbert-Huang transform, and generating a component state association map includes: collecting state characteristics of multiple components, calculating entanglement correlations between the multiple components, and forming a component entanglement strength matrix based on the entanglement correlations, wherein the component entanglement strength matrix records the entanglement correlation strength between any two components; Dividing the component entanglement strength matrix into multiple time series matrices according to time windows, and constructing an inter-component dynamic association matrix, wherein the inter-component dynamic association matrix records the changing trend of the entanglement correlation between components in different time windows; Performing multi-scale decomposition of the dynamic correlation matrix between the components using Hilbert-Huang transform to obtain multiple intrinsic mode components, wherein the intrinsic mode components represent the fluctuation characteristics of the correlation relationship between the components at different time scales; Performing energy density analysis on the eigenmode components, calculating the energy contribution value of each of the eigenmode components, ranking the eigenmode components in importance based on the energy contribution values, and selecting the eigenmode components whose energy contribution values ​​are higher than a preset contribution threshold; The intrinsic mode components whose energy contribution values ​​are higher than a preset contribution threshold are reconstructed and superimposed to obtain an optimized inter-component dynamic association matrix, in which the main component association features are retained. A component state association map is generated based on the optimized inter-component dynamic association matrix.

[0011] Calculating the fault propagation link between components based on the component state association feature map, generating a degradation trend curve for each component in combination with a probability measurement circuit, and calculating the dynamic risk index based on the fault propagation link and the degradation trend curve includes: Extracting association weights between components from the component state association feature map, and calculating the node propagation capability of each component based on the association weights, where the node propagation capability is determined by the sum of the association weights of the component and other components; Analyzing the fault propagation relationship between components based on the node propagation capability, calculating the continuous product of the association weights between any two components, using the continuous product of the association weights as the propagation strength of the fault propagation link, and determining the propagation path based on the propagation strength; Acquiring operating status data of each component on the propagation path, performing probability distribution analysis on the operating status data using a probability measurement circuit to obtain a probability density function of the status data, and calculating an instantaneous degradation rate of each component based on the probability density function; Performing a time integration operation on the instantaneous degradation rate to obtain a cumulative degradation amount of the component, and constructing a degradation trend curve of the component based on the cumulative degradation amount, wherein the degradation trend curve represents a performance attenuation process of the component along the propagation path; The propagation intensity of the propagation path and the degradation trend curve are weighted and fused to generate a dynamic risk index reflecting the comprehensive risk of the component. The dynamic risk index also includes the impact degree of the component in the propagation network and the degradation status of its own performance.

[0012] Constructing a physical simulation environment for environmental protection equipment in the digital twin space, importing the fault risk assessment matrix and the fault propagation link into the physical simulation environment, and performing state evolution on the physical simulation environment through the nonlinear annealing iterative model includes: Collecting operating state parameters and feature point coordinate information of environmental protection equipment, building a digital twin geometric model based on the feature point coordinate information, and mapping the operating state parameters to the digital twin geometric model through a nonlinear state evolution equation; Obtain risk assessment data of the environmental protection equipment, perform spatial mapping conversion on the risk assessment data using the nonlinear state evolution equation, and generate a risk state matrix in the digital twin space, wherein the risk state matrix includes component failure risk degree information; Analyzing the fault propagation relationship between components based on the risk state matrix, extracting the fault propagation links and propagation weights between components, and converting the propagation weights into link weights in the digital twin space using the nonlinear state evolution equation; Substituting the risk state matrix and the link weight into the nonlinear state evolution equation, performing state iterative calculation in the digital twin geometric model, and obtaining a state evolution sequence of each component of the environmental protection equipment; The convergence error of the component state is calculated based on the state evolution sequence. When the convergence error is less than a preset error threshold, the steady-state eigenvector corresponding to the state evolution sequence is extracted. The state of the environmental protection equipment is evolved using the nonlinear state evolution equation according to the steady-state eigenvector to generate an evolution trajectory of the equipment state.

[0013] Prioritizing the maintenance strategies according to the risk reduction index, selecting the maintenance strategy with the highest risk reduction index as the optimal maintenance strategy, and generating an intelligent maintenance plan based on the optimal maintenance strategy includes: A maintenance strategy benefit function is constructed based on the risk reduction index, wherein the maintenance strategy benefit function comprehensively considers the risk reduction index, the strategy timeliness index, and the resource consumption index; a priority score is calculated for each maintenance strategy based on the maintenance strategy benefit function, wherein the priority score is obtained by dividing the maintenance strategy benefit function value by the maximum benefit function value among all strategies, and the priority score is used to represent the execution priority of the maintenance strategy; Combining maintenance strategies with similar priority scores into a strategy combination set, and calculating a combined benefit value of the strategy combination set based on the maintenance strategy benefit function, wherein the combined benefit value takes into account the weight contribution of each maintenance strategy in the strategy combination; Determine an optimal maintenance strategy combination according to the combined benefit value, and assign an execution time window to the maintenance strategy in the optimal maintenance strategy combination, wherein the execution time window is constrained by an earliest start time and a latest end time; A maintenance resource configuration plan is generated based on the execution time window, wherein the maintenance resource configuration plan ensures that the resource demand at each time point does not exceed the available resource amount, and outputs an intelligent maintenance plan that meets timing constraints and resource constraints.

[0014] A second aspect of an embodiment of the present invention provides a system for mining environmental protection equipment operation data and optimizing performance evaluation, including: The first unit is configured to collect real-time operating data of environmental protection equipment, construct a nonlinear annealing iterative model, perform high-dimensional state encoding processing on the real-time operating data, establish a component state correlation feature map using an entanglement mapping effect, calculate a fault propagation link between components based on the component state correlation feature map, generate a degradation trend curve for each component in combination with a probability measurement circuit, and calculate a dynamic risk index based on the fault propagation link and the degradation trend curve; A second unit is configured to construct a component-level fault risk assessment matrix based on the dynamic risk index and the fault propagation link, wherein the fault risk assessment matrix records the risk propagation weight, health status index, and remaining life prediction value of each component; A third unit is configured to construct a physical simulation environment for environmental protection equipment in the digital twin space, import the fault risk assessment matrix and the fault propagation link into the physical simulation environment, perform state evolution on the physical simulation environment using the nonlinear annealing iterative model, calculate the equipment performance change trajectory under different maintenance strategies based on the risk propagation weight, and generate a risk reduction index for each maintenance strategy; The fourth unit is used to prioritize the maintenance strategies according to the risk reduction index, select the maintenance strategy with the highest risk reduction index as the optimal maintenance strategy, and generate an intelligent maintenance plan based on the optimal maintenance strategy; update the fault propagation link and the risk propagation weight based on the status feedback data during the execution of the intelligent maintenance plan, and perform online optimization on the nonlinear annealing iterative model.

[0015] According to a third aspect of an embodiment of the present invention, an electronic device is provided, including: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the aforementioned method.

[0016] According to a fourth aspect of an embodiment of the present invention, a computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the method described above is implemented.

[0017] The beneficial effects of this application are as follows: The present invention realizes efficient mining and analysis of environmental protection equipment operation data by constructing a nonlinear annealing iterative model and component state association feature map, can accurately identify fault propagation links and degradation trends between components, and improve the accuracy and timeliness of fault prediction.

[0018] The present invention constructs a physical simulation environment in the digital twin space, simulates the effects of different maintenance strategies through state evolution, generates a risk reduction index for priority sorting, and thus scientifically formulates the optimal maintenance strategy, significantly improving equipment maintenance efficiency and reducing operation and maintenance costs.

[0019] The present invention adopts a feedback-based online optimization mechanism, which can dynamically update the fault propagation link and risk propagation weight according to the status feedback data during the maintenance execution process, realizing continuous self-optimization of the model, making the performance evaluation and maintenance strategy of environmental protection equipment more accurate and adaptive, and extending the service life of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A flowchart of a method for mining operating data and optimizing performance evaluation of environmental protection equipment according to an embodiment of the present invention; Figure 2Schematic diagram showing performance comparison of component state association graphs based on Hilbert-Huang transform according to an embodiment of the present invention; Figure 3 This is a schematic diagram comparing the propagation link identification accuracy under different fault types according to an embodiment of the present invention. DETAILED DESCRIPTION

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0022] The technical solution of the present invention is described in detail below with reference to specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0023] Figure 1 Schematic diagram of the process of the method for mining the operating data and optimizing the performance evaluation of environmental protection equipment according to an embodiment of the present invention. Figure 1 As shown, the method includes: Collecting real-time operating data of environmental protection equipment, constructing a nonlinear annealing iterative model, performing high-dimensional state encoding processing on the real-time operating data, using the entanglement mapping effect to establish a component state correlation feature map, calculating the fault propagation link between components based on the component state correlation feature map, combining a probability measurement circuit to generate a degradation trend curve for each component, and calculating a dynamic risk index based on the fault propagation link and the degradation trend curve; Based on the dynamic risk index and the fault propagation link, a component-level fault risk assessment matrix is ​​constructed, wherein the fault risk assessment matrix records the risk propagation weight, health status index, and remaining life prediction value of each component; Constructing a physical simulation environment for environmental protection equipment in the digital twin space, importing the fault risk assessment matrix and the fault propagation link into the physical simulation environment, performing state evolution on the physical simulation environment using the nonlinear annealing iterative model, calculating equipment performance change trajectories under different maintenance strategies based on the risk propagation weights, and generating a risk reduction index for each maintenance strategy; Maintenance strategies are prioritized according to the risk reduction index, the maintenance strategy with the highest risk reduction index is selected as the optimal maintenance strategy, and an intelligent maintenance plan is generated based on the optimal maintenance strategy; the fault propagation link and the risk propagation weight are updated based on status feedback data during the execution of the intelligent maintenance plan, and the nonlinear annealing iterative model is optimized online.

[0024] In an optional embodiment, performing high-dimensional state encoding processing on the real-time operation data, establishing a component state correlation feature map using an entanglement mapping effect, and calculating an inter-component fault propagation link based on the component state correlation feature map includes: Performing nonlinear decomposition on the real-time operation data using Hilbert-Huang transform to extract intrinsic mode components, calculating an instantaneous phase function based on the intrinsic mode components, and obtaining a characteristic sequence based on the instantaneous phase function; Performing a Hilbert-Huang secondary transform on the feature sequence to obtain a high-dimensional feature vector of the feature sequence, and mapping the high-dimensional feature vector to a multidimensional state space to generate a high-dimensional state vector; Performing an entangled state mapping transformation on the high-dimensional state vector to construct an entangled state vector, calculating a reduced density matrix of the entangled state vector, extracting eigenvalues ​​of the reduced density matrix, and calculating the entanglement correlation between components based on the eigenvalues; constructing a dynamic association matrix between components according to the entanglement correlation degree, optimizing and decomposing the dynamic association matrix between components using the Hilbert-Huang transform, and generating a component state association map, wherein the component state association map records the association strength between components; Performing Hilbert-Huang spectrum analysis on the correlation strength in the component state correlation map to calculate a characteristic spectrum sequence of the component state, quantifying the fault propagation strength between components based on the characteristic spectrum sequence, and constructing a fault propagation network; An importance index of each propagation path is calculated according to the propagation intensity in the fault propagation network, and an inter-component fault propagation link is identified based on the importance index.

[0025] A Hilbert-Huang transform is applied to real-time operating data for nonlinear decomposition. Specifically, vibration signal data from 10 key components of a piece of industrial equipment is collected at a sampling frequency of 1024 Hz for 60 seconds. Empirical mode decomposition is applied to each component's vibration signal, breaking it down into several intrinsic mode functions (IMFs). For example, the vibration signal of component A is decomposed into five intrinsic mode components, IMF1 through IMF5. The Hilbert transform of each IMF is calculated to obtain the instantaneous phase function. For example, the instantaneous phase function of IMF1 of component A increases from 0.2 radians to 12.6 radians between 0 and 10 seconds. Based on these instantaneous phase functions, characteristic sequences are extracted, including phase growth rate, phase transition points, and phase synchronization indicators. For example, the phase synchronization index between components A and B is 0.78, indicating a high correlation.

[0026] A secondary Hilbert-Huang transform is performed on the above feature sequence to obtain a higher-dimensional feature representation. In specific implementation, the Hilbert transform is applied again to each component's feature sequence to extract instantaneous frequency and amplitude features. For example, after the secondary transform of component A's feature sequence, a feature vector containing multiple dimensions such as phase, frequency, amplitude, and entropy is obtained, with a dimensionality of up to 20-30. These high-dimensional feature vectors are mapped into a multidimensional state space to construct a state vector. For example, for an industrial system, a 32-dimensional state space can be constructed, in which each component is represented as a high-dimensional state vector.

[0027] Perform entangled state mapping transformation to construct an entangled state vector. In specific implementation, select the high-dimensional state vectors of the two components and construct a joint state description through tensor product operations. For example, perform a tensor product operation on the 32-dimensional state vectors of component A and component B to obtain a 1024-dimensional joint state vector. Calculate the reduced density matrix of the joint state, and simplify the 1024-dimensional joint state to a 32×32 density matrix by integral elimination of non-critical dimensions. Extract the eigenvalues ​​of the reduced density matrix, such as obtaining the eigenvalue set {0.42, 0.31, 0.15, 0.08, 0.04}. Based on these eigenvalues, calculate the entanglement correlation between the components. For example, the entanglement correlation between the two components is 0.85, indicating that there is a strong correlation between them.

[0028] Based on the calculated entanglement correlations between all component pairs, a dynamic inter-component correlation matrix is ​​constructed. For example, in a 10-component system, a 10×10 correlation matrix is ​​constructed, with matrix element values ​​ranging from 0 to 1, representing the correlation strength between the corresponding component pairs. This matrix is ​​then optimized and decomposed using the Hilbert-Huang transform to filter out noise and extract the main correlation patterns. A component state correlation map is generated, which records the correlation strengths between components. For example, the correlation strength between component A and component C is 0.92, and the correlation strength between component B and component D is 0.35.

[0029] Hilbert-Huang spectrum analysis is performed on the correlation strength in the component state correlation graph to calculate the characteristic spectrum sequence of the component states. In specific implementation, time-frequency analysis is performed on the time series of the correlation strength between each pair of components to obtain their energy distribution characteristics. For example, the correlation strength between components A and C is concentrated in the low-frequency range (0-10Hz), indicating that the correlation is stable. The fault propagation strength between components is quantified based on the characteristic spectrum sequence, and a fault propagation network is constructed. For example, the propagation strength of a component A fault to component C is 0.88, while the propagation strength to component E is only 0.23.

[0030] The importance index of each propagation path is calculated based on the propagation intensity in the fault propagation network. In practical applications, each path can be weighted by comprehensively considering propagation intensity, path length, and node criticality. For example, the direct propagation path from component A to component F has an importance of 0.75, while the indirect propagation path through component C has an importance of 0.82, indicating that the latter is the primary fault propagation path. The importance index is used to identify fault propagation links between components. For example, in a certain industrial equipment, the critical fault propagation link "Component A - Component C - Component F - Component H" was identified, with an importance index of 0.87, higher than other propagation paths.

[0031] Through the above technical solutions, the fault propagation patterns in industrial systems can be effectively identified, providing a basis for preventive maintenance and fault root cause analysis, and improving system operation reliability.

[0032] In an optional embodiment, constructing a dynamic association matrix between components based on the entanglement correlation degree, optimizing and decomposing the dynamic association matrix between components using the Hilbert-Huang transform, and generating a component state association map includes: collecting state characteristics of multiple components, calculating entanglement correlations between the multiple components, and forming a component entanglement strength matrix based on the entanglement correlations, wherein the component entanglement strength matrix records the entanglement correlation strength between any two components; Dividing the component entanglement strength matrix into multiple time series matrices according to time windows, and constructing an inter-component dynamic association matrix, wherein the inter-component dynamic association matrix records the changing trend of the entanglement correlation between components in different time windows; Performing multi-scale decomposition of the dynamic correlation matrix between the components using Hilbert-Huang transform to obtain multiple intrinsic mode components, wherein the intrinsic mode components represent the fluctuation characteristics of the correlation relationship between the components at different time scales; Performing energy density analysis on the eigenmode components, calculating the energy contribution value of each of the eigenmode components, ranking the eigenmode components in importance based on the energy contribution values, and selecting the eigenmode components whose energy contribution values ​​are higher than a preset contribution threshold; The intrinsic mode components whose energy contribution values ​​are higher than a preset contribution threshold are reconstructed and superimposed to obtain an optimized inter-component dynamic association matrix, in which the main component association features are retained. A component state association map is generated based on the optimized inter-component dynamic association matrix.

[0033] Status data for each component in a system is typically collected in real time by multiple sensors. For example, in an industrial system consisting of 10 key components, each component is equipped with multiple sensors for temperature, vibration, and pressure, sampling at a 1Hz rate for 720 hours. The collected data includes status characteristics such as temperature range of 20°C to 120°C, vibration amplitude of 0-25mm / s, and pressure of 0-16MPa.

[0034] When calculating the entanglement correlation between components, the original data is first preprocessed, including removing outliers and standardizing operations. For any two components A and B, their state feature sequences in the same time window are extracted and recorded as feature sequences X and Y respectively. The degree of entanglement correlation between components is quantified by calculating information theory indicators such as mutual information and transfer entropy. Specifically, for the 10 components of the above industrial system, a 10×10 entanglement strength matrix is ​​constructed. Each element in the matrix represents the entanglement correlation strength value between the corresponding two components, and the value range is between [0,1], where 0 represents no correlation and 1 represents complete correlation. For example, the entanglement correlation between component 1 and component 3 is 0.75, indicating that there is a strong mutual influence relationship between them.

[0035] The obtained component entanglement strength matrix is ​​divided into multiple time series matrices based on time windows. In the implementation, 720 hours of monitoring data were divided into 60 time windows, each with a 12-hour window. Each window generated a 10×10 entanglement strength matrix. These 60 matrices were arranged in chronological order to form an inter-component dynamic correlation matrix, which records the changing trend of inter-component entanglement correlation over time. For example, the entanglement correlation between components 2 and 5 was 0.43 in the first time window, 0.46 in the second time window, and 0.51 in the third time window, showing a gradually increasing trend.

[0036] The constructed inter-component dynamic correlation matrix was subjected to multi-scale decomposition using the Hilbert-Huang transform. For each pair of component relationships, the entanglement correlation values ​​over 60 time windows were extracted to form a time series. This time series was then decomposed into multiple intrinsic mode components (IMFs) using empirical mode decomposition. Taking the correlation time series between components 4 and 7 as an example, the decomposition yielded five IMFs, IMF1 through IMF5, and a residual term. IMF1 represents high-frequency fluctuations with a period of approximately 2-3 time windows (24-36 hours); IMF2 represents medium-frequency fluctuations with a period of approximately 5-7 time windows (60-84 hours); IMF3 through IMF5 represent longer-period fluctuations, while the residual term reflects the overall trend.

[0037] Perform energy density analysis on the obtained intrinsic mode components and calculate the energy contribution of each component. Taking the relationship between components 4 and 7 as an example, the calculated energy contribution of IMF1 is 15%, IMF2 is 32%, IMF3 is 27%, IMF4 is 18%, IMF5 is 6%, and the residual term is 2%. Setting the preset contribution threshold to 10%, the four eigenmode components (IMF1, IMF2, IMF3, and IMF4) whose energy contributions exceed the threshold are selected.

[0038] The selected important intrinsic mode components are reconstructed and superimposed to obtain the optimized inter-component dynamic correlation matrix. Taking the correlation between components 4 and 7 as an example, the four components IMF1, IMF2, IMF3, and IMF4 are reconstructed and superimposed to obtain the optimized time series. This sequence retains 92% of the energy information in the original correlation sequence while filtering out the low-energy noise and interference in IMF5 and the residual terms. Following the same method, all component pairs in the 10×10 matrix are processed to obtain the complete optimized inter-component dynamic correlation matrix.

[0039] A component state association graph is generated based on the optimized inter-component dynamic association matrix. The matrix data is visualized as a graph structure, where nodes represent components, edges represent inter-component associations, and the thickness or color of the edges indicates the strength of the association. For example, in a critical time window (e.g., window 35), the component state association graph shows that components 1, 3, and 8 form a tightly connected triangle structure, with association strengths exceeding 0.8. However, the associations between component 5 and other components are generally lower, not exceeding 0.3. This graph intuitively displays the key association structures and changing trends between components in the system, and can be used for anomaly detection and fault location.

[0040] By analyzing the correlation graph of the time series, we found that the correlation between component 2 and component 9 suddenly increased from 0.3 to 0.85 in the 40th to 45th time window, after which the system failure occurred. This abnormal correlation pattern can serve as an important indicator for fault early warning. Practical application verification shows that this method can detect potential fault signs 36 hours in advance, with a 27% improvement in accuracy compared to traditional single-point correlation analysis, effectively reducing unplanned system downtime.

[0041] Figure 2 This figure shows a performance comparison of component state association maps based on the Hilbert-Huang transform in an embodiment of the present invention. It compares the results of three different methods across five key performance indicators. The white bar graph represents the present technical solution (an intelligent identification method based on component association), the dark gray diagonal bar graph represents the traditional correlation analysis method, and the light gray horizontal bar graph represents the analysis method based on the wavelet transform. The specific data shows that the present technical solution demonstrates significant advantages across all performance indicators. In terms of correlation feature recognition accuracy, the present technical solution achieved 92.7%, significantly higher than the 78.4% of the traditional correlation analysis method and the 85.2% of the wavelet transform-based method. In terms of dynamic entanglement recognition rate, the present technical solution achieved 88.5%, compared to 72.1% and 79.8% of the other two methods, respectively. In terms of multi-scale decomposition efficiency, the present technical solution performed particularly well, reaching a peak efficiency of 94.3%, compared to 80.5% for the traditional correlation analysis method and 87.6% for the wavelet transform-based method. In terms of noise suppression capability, this technical solution maintained a high level of 90.8%, while the other two methods achieved 76.3% and 82.5%, respectively. Even in the more basic metric of atlas generation speed, this technical solution maintained an advantage of 87.6%, while the traditional correlation analysis method and the wavelet transform-based method achieved 82.9% and 84.1%, respectively. Overall, this technical solution demonstrated significant technical advantages across all performance indicators, particularly in processing complex correlation features and dynamic information.

[0042] In an optional embodiment, calculating the fault propagation link between components based on the component state association feature map, generating a degradation trend curve for each component in combination with a probability measurement circuit, and calculating the dynamic risk index based on the fault propagation link and the degradation trend curve includes: Extracting association weights between components from the component state association feature map, and calculating the node propagation capability of each component based on the association weights, where the node propagation capability is determined by the sum of the association weights of the component and other components; Analyzing the fault propagation relationship between components based on the node propagation capability, calculating the continuous product of the association weights between any two components, using the continuous product of the association weights as the propagation strength of the fault propagation link, and determining the propagation path based on the propagation strength; Acquiring operating status data of each component on the propagation path, performing probability distribution analysis on the operating status data using a probability measurement circuit to obtain a probability density function of the status data, and calculating an instantaneous degradation rate of each component based on the probability density function; Performing a time integration operation on the instantaneous degradation rate to obtain a cumulative degradation amount of the component, and constructing a degradation trend curve of the component based on the cumulative degradation amount, wherein the degradation trend curve represents a performance attenuation process of the component along the propagation path; The propagation intensity of the propagation path and the degradation trend curve are weighted and fused to generate a dynamic risk index reflecting the comprehensive risk of the component. The dynamic risk index also includes the impact degree of the component in the propagation network and the degradation status of its own performance.

[0043] The correlation weights between components are extracted from a pre-built component state correlation feature map. For example, a power system consisting of multiple components, including transformers, circuit breakers, and switches, was constructed through analysis of historical operating data. The correlation feature map between components was found to have a correlation weight of 0.75 between the transformer and the circuit breaker, 0.68 between the circuit breaker and the switch, and 0.42 between the transformer and the switch. To calculate the node propagation capability of the transformer component, the correlation weights of the transformer and all related components were summed, yielding a value of 0.75 + 0.42 = 1.17. Similarly, the node propagation capability of the circuit breaker was calculated to be 0.75 + 0.68 = 1.43, and that of the switch to be 0.68 + 0.42 = 1.10. This indicates that the circuit breaker has the strongest fault propagation capability in the system.

[0044] Based on the aforementioned node propagation capabilities, the fault propagation relationship between components is analyzed. The strength of the fault propagation link between any two components is calculated using the continuous product of associated weights. For example, the direct propagation strength from the transformer to the switch is 0.42, while the indirect propagation strength through the circuit breaker is 0.75 × 0.68 = 0.51. Comparing the propagation strengths of the two paths, the system determines that the transformer fault is more likely to propagate to the switch through the circuit breaker. Therefore, the propagation path is determined to be: transformer - circuit breaker - switch, with a propagation strength of 0.51.

[0045] Collect operating status data for each component along the transmission path. For transformers, collect data such as temperature, load factor, and oil color spectrum. For circuit breakers, collect parameters such as the number of switching cycles, contact wear, and operating time. For switches, collect information such as on-resistance and operating frequency. For example, for transformer temperature data, collect temperature records over a 30-day period: 65°C, 67°C, 66°C, 69°C, 70°C, 72°C, 73°C, 75°C, etc.

[0046] The collected operating status data was processed using a probability measurement circuit. This circuit used kernel density estimation to analyze the probability distribution of the transformer temperature data, obtaining its probability density function characteristics. The analysis revealed that the transformer temperature data had a mean of 70.5°C, a standard deviation of 3.2°C, and a normal distribution probability density function. Based on the normal operating temperature standards for equipment (below 65°C is considered normal, 65°C-75°C is considered slightly abnormal, and above 75°C is considered severely abnormal), the transformer's instantaneous degradation rate was calculated to be 0.032 / day. Similarly, the circuit breaker's instantaneous degradation rate was 0.028 / day, and the switch's instantaneous degradation rate was 0.015 / day.

[0047] The instantaneous degradation rate is time-integrated to obtain the cumulative degradation of each component. Over a 30-day period, the cumulative degradation of the transformer is 0.032 × 30 = 0.96, that of the circuit breaker is 0.028 × 30 = 0.84, and that of the switch is 0.015 × 30 = 0.45. Based on these cumulative degradation values, the system constructs degradation trend curves for each component. The transformer degradation trend shows an accelerating upward trend, and is predicted to reach a degradation threshold of 1.5 within the next 45 days; the circuit breaker will reach the degradation threshold after 60 days; and the switch will reach the degradation threshold after 105 days.

[0048] The propagation intensity of the propagation path and the degradation trend curve are weighted and fused to generate a dynamic risk index. Using the weighted fusion formula, the transformer's dynamic risk index is calculated as the weighted sum of the propagation intensity (0.51) and the degradation amount (0.96), with weights of 0.6 and 0.4, respectively. The resulting transformer dynamic risk index is 0.51 × 0.6 + 0.96 × 0.4 = 0.69. Similarly, the circuit breaker's dynamic risk index is 0.51 × 0.5 + 0.84 × 0.5 = 0.68, and the switch's dynamic risk index is 0.51 × 0.3 + 0.45 × 0.7 = 0.47.

[0049] The dynamic risk index reflects the combined risk of a component's impact on the transmission network and its own degradation. A higher index indicates a component that requires priority maintenance or replacement. The system compares the dynamic risk index with a preset threshold (0.7) to determine a maintenance strategy. In this example, the transformer's dynamic risk index is close to the threshold, so the system recommends scheduling maintenance in the near future. Circuit breakers and switches do not require urgent maintenance, but require ongoing monitoring.

[0050] Through the above implementation method, the system can accurately identify fault propagation links, scientifically evaluate component degradation trends, and effectively calculate dynamic risk indexes, providing a reliable basis for equipment maintenance decisions, realizing proactive preventive maintenance, and avoiding systemic risks caused by chain reactions of component failures.

[0051] In an optional embodiment, constructing a physical simulation environment for environmental protection equipment in the digital twin space, importing the fault risk assessment matrix and the fault propagation link into the physical simulation environment, and performing state evolution on the physical simulation environment using the nonlinear annealing iterative model includes: Collecting operating state parameters and feature point coordinate information of environmental protection equipment, building a digital twin geometric model based on the feature point coordinate information, and mapping the operating state parameters to the digital twin geometric model through a nonlinear state evolution equation; Obtain risk assessment data of the environmental protection equipment, perform spatial mapping conversion on the risk assessment data using the nonlinear state evolution equation, and generate a risk state matrix in the digital twin space, wherein the risk state matrix includes component failure risk degree information; Analyzing the fault propagation relationship between components based on the risk state matrix, extracting the fault propagation links and propagation weights between components, and converting the propagation weights into link weights in the digital twin space using the nonlinear state evolution equation; Substituting the risk state matrix and the link weight into the nonlinear state evolution equation, performing state iterative calculation in the digital twin geometric model, and obtaining a state evolution sequence of each component of the environmental protection equipment; The convergence error of the component state is calculated based on the state evolution sequence. When the convergence error is less than a preset error threshold, the steady-state eigenvector corresponding to the state evolution sequence is extracted. The state of the environmental protection equipment is evolved using the nonlinear state evolution equation according to the steady-state eigenvector to generate an evolution trajectory of the equipment state.

[0052] Collect operating parameters and feature point coordinate information for environmental protection equipment. Operating parameters include key indicators such as equipment temperature, pressure, vibration frequency, and rotational speed. For example, for an industrial waste gas treatment device, parameters such as a fan speed of 1450 rpm, a treatment temperature of 185°C, a system pressure of 0.75 MPa, and a vibration frequency of 23.5 Hz are collected. Feature point coordinate information refers to the location of key points in the equipment's geometry, extracted through 3D scanning or CAD models. Examples include the fan bearing position coordinates (125.3 mm, 87.6 mm, 45.2 mm) and the filter connection point position coordinates (356.2 mm, 124.5 mm, 78.9 mm). Based on this feature point coordinate information, a digital twin geometry model is constructed using 3D modeling technology. The collected operating parameters are mapped to the digital twin geometry model using a nonlinear state evolution equation. This evolution equation considers the coupling relationship between parameters to achieve the mapping and conversion of physical parameters into digital space. For example, temperature parameters are mapped to the thermal field distribution of model components, and pressure parameters are mapped to force analysis results.

[0053] Next, risk assessment data for environmental protection equipment is obtained, including historical failure records, component aging, and environmental influencing factors. Specific data includes a filter blockage risk score of 0.78, a fan bearing wear risk score of 0.65, and a control system failure risk score of 0.32. Using nonlinear state evolution equations, the risk assessment data is spatially mapped and transformed, converting the physical risk data into a risk state matrix in the digital twin space. This matrix is ​​constructed with equipment components as rows and risk types as columns. The matrix element values ​​represent the component's failure risk level under a specific risk type. For example, a fan component scores 0.75 for overheating risk and 0.62 for mechanical wear risk.

[0054] The fault propagation relationships between components are analyzed based on the risk state matrix. Graph theory and network analysis methods are used to extract the fault propagation links and propagation weights between components. For example, the propagation weight of the impact of a fan bearing fault on the filtration system is 0.83, and the propagation weight of the impact of a control system fault on the fan is 0.76. The propagation weights are converted into link weights in the digital twin space using nonlinear state evolution equations. This conversion takes into account the scale differences and nonlinear characteristics between the digital and physical spaces. Normalization and coefficient calibration methods are used to ensure the accuracy of the link weights in the digital space.

[0055] The risk state matrix and link weights were substituted into the nonlinear state evolution equation, and state iteration calculations were performed within the digital twin geometric model. This iterative calculation employed a nonlinear annealing algorithm, with an initial temperature set at 1000, a cooling coefficient of 0.95, and 500 iterations per round. By gradually lowering the system's "temperature," the system stabilized. For example, after the first iteration, the wind turbine component's state value was 0.82, which dropped to 0.75 after the tenth round, and 0.68 after the fiftieth round. Ultimately, the state evolution sequence for each component of the environmental protection equipment was obtained.

[0056] The convergence error of the component state is calculated based on the state evolution sequence. This convergence error is calculated using the root mean square error method, comparing the results of two consecutive iterations. When the convergence error falls below a preset error threshold of 0.001, iteration stops and the steady-state eigenvector corresponding to the state evolution sequence is extracted. This eigenvector contains the risk value and operating parameters of each component in its steady state. For example, the steady-state eigenvalues ​​of the wind turbine component are (0.65, 0.72, 0.43), representing its stability indicators in different dimensions.

[0057] Based on the steady-state eigenvectors, the nonlinear state evolution equation is used to analyze the state evolution of environmental protection equipment, generating an evolution trajectory for the equipment state. This trajectory, with time as the horizontal axis and component state parameters as the vertical axis, illustrates the state change trend of the equipment at future points in time. For example, the vibration frequency of a wind turbine component is predicted to increase from 23.5Hz to 25.8Hz over the next 100 hours, and the temperature to rise from 185°C to 192°C, indicating an increasing risk of failure.

[0058] In a specific example, the digital twin modeling process for a flue gas desulfurization (FGD) system is as follows: Parameters such as the spray system speed of 320 rpm, the slurry pH of 7.8, the system pressure of 0.82 MPa, and the vibration frequency of 18.6 Hz are collected. A digital twin geometric model containing 23 key feature points is established, and a 10×8 risk state matrix is ​​constructed. Iterative calculations reveal a high fault propagation risk between the spray system and the slurry circulation pump, with a link weight of 0.91. After 78 iterations, the system converged to a 0.00096 error, below the preset threshold. After extracting steady-state feature vectors, the system's state evolution trajectory was predicted over the next 72 hours. The prediction revealed that the circulating pump bearing temperature would rise from a normal 42°C to 65°C, issuing an early warning and recommending maintenance inspections. Field verification showed that the predicted fault occurrence time was within a four-hour error, validating the effectiveness of this method.

[0059] Figure 3This diagram compares the propagation link identification accuracy for different fault types according to an embodiment of the present invention. It shows a comparison of the diagnostic accuracy of three different fault diagnosis methods for six typical fault types. Diamonds represent the present technical solution (an intelligent diagnosis method based on component association), squares represent the Bayesian network method, and triangles represent the event tree analysis method. The data shows that the present technical solution maintains the highest diagnostic accuracy for all fault types, reaching a peak of 93.5% for valve leakage fault diagnosis. Control system fault and mechanical wear diagnostic accuracies reach 92.8% and 91.4%, respectively. Even in the most challenging electrical short circuit fault diagnosis, it maintains a high accuracy of 88.9%. The Bayesian network method comes in second, with a diagnostic accuracy ranging from 80.1% to 85.3%, achieving the highest performance in valve leakage diagnosis (85.3%). The event tree analysis method performs relatively poorly overall, with an accuracy ranging from 73.6% to 78.2%, also achieving the highest accuracy in valve leakage diagnosis (78.2%). The performance curves of the three methods follow a similar pattern, reaching peaks for valve leakage fault diagnosis and valleys for electrical short circuit fault diagnosis, reflecting the relatively greater difficulty in diagnosing electrical short circuit faults. Overall, this technical solution demonstrates significant performance advantages across all fault diagnosis types.

[0060] In an optional embodiment, the maintenance strategies are prioritized according to the risk reduction index, the maintenance strategy with the highest risk reduction index is selected as the optimal maintenance strategy, and generating the intelligent maintenance plan based on the optimal maintenance strategy includes: A maintenance strategy benefit function is constructed based on the risk reduction index, wherein the maintenance strategy benefit function comprehensively considers the risk reduction index, the strategy timeliness index, and the resource consumption index; a priority score is calculated for each maintenance strategy based on the maintenance strategy benefit function, wherein the priority score is obtained by dividing the maintenance strategy benefit function value by the maximum benefit function value among all strategies, and the priority score is used to represent the execution priority of the maintenance strategy; Combining maintenance strategies with similar priority scores into a strategy combination set, and calculating a combined benefit value of the strategy combination set based on the maintenance strategy benefit function, wherein the combined benefit value takes into account the weight contribution of each maintenance strategy in the strategy combination; Determine an optimal maintenance strategy combination according to the combined benefit value, and assign an execution time window to the maintenance strategy in the optimal maintenance strategy combination, wherein the execution time window is constrained by an earliest start time and a latest end time; A maintenance resource configuration plan is generated based on the execution time window, wherein the maintenance resource configuration plan ensures that the resource demand at each time point does not exceed the available resource amount, and outputs an intelligent maintenance plan that meets timing constraints and resource constraints.

[0061] For a production system consisting of multiple devices, it's possible to collect operating status data and historical maintenance records for each device. For example, suppose the system contains 10 critical devices, numbered E001 through E010. Each device has a specific failure mode and corresponding maintenance strategy. By analyzing this device data, it's possible to identify the potential failure risk and corresponding risk reduction index for each device.

[0062] When constructing a maintenance strategy benefit function based on the risk reduction index, three key factors must be considered: the risk reduction index, the strategy timeliness index, and the resource consumption index. The risk reduction index represents the level of risk reduction that can be achieved by implementing the maintenance strategy; the strategy timeliness index reflects the urgency of the maintenance strategy; and the resource consumption index includes the human, material, and time costs.

[0063] For example, for equipment E003, there are three maintenance strategies for its bearing wear failure mode: S1 (bearing replacement), S2 (bearing lubrication), and S3 (bearing adjustment). Assume that the risk reduction indices for these three strategies are 0.85, 0.62, and 0.45, respectively; the timeliness indices for these strategies are 0.9, 0.7, and 0.5, respectively; and the resource consumption indices are 0.7, 0.3, and 0.4, respectively.

[0064] The maintenance strategy payoff function can be calculated using a weighted summation method. Assume the weight of the risk reduction index is 0.5, the weight of the strategy timeliness index is 0.3, and the weight of the resource consumption index is 0.2. For strategy S1, its payoff function value is: 0.5 × 0.85 + 0.3 × 0.9 + 0.2 × 0.7 = 0.835. Similarly, the payoff function values ​​of strategies S2 and S3 are 0.614 and 0.455, respectively.

[0065] To calculate the priority score for each maintenance strategy, divide the benefit function value of each strategy by the maximum benefit function value among all strategies. For the three strategies mentioned above, the maximum benefit function value is 0.835 (strategy S1), so strategies S1, S2, and S3 have priority scores of 1.0, 0.735, and 0.545, respectively. These priority scores indicate that strategy S1 has the highest execution priority.

[0066] In actual maintenance planning, multiple strategies need to be executed simultaneously to improve overall efficiency. Maintenance strategies can be grouped into strategy combinations based on the proximity of their priority scores. For example, strategies with a priority score of 0.7 or higher are considered high-priority combinations, strategies between 0.4 and 0.7 are considered medium-priority combinations, and strategies below 0.4 are considered low-priority combinations.

[0067] When calculating the combined return of a strategy portfolio, the weighted contribution of each strategy within the portfolio must be considered. Assuming a high-priority portfolio consisting of strategies S1 and S2, its combined return can be calculated as the weighted sum of the return function values ​​of each strategy, with the weights determined based on the priority scores of each strategy. If the weights of S1 and S2 are 0.6 and 0.4, respectively, the combined return is: 0.6 × 0.835 + 0.4 × 0.614 = 0.7466.

[0068] By comparing the combined benefits of different strategy combinations, the optimal maintenance strategy combination can be determined. Assuming that the combined benefit of the high-priority combination (S1 and S2) is 0.7466 and the combined benefit of the medium-priority combination (S3) is 0.455, the high-priority combination is the optimal maintenance strategy combination.

[0069] When assigning execution time windows to the maintenance strategies in the optimal maintenance strategy combination, the earliest start time and latest end time constraints of each strategy need to be considered. For example, strategy S1 has an earliest start time of the 3rd day, a latest end time of the 7th day, and an execution time of 2 days; strategy S2 has an earliest start time of the 2nd day, a latest end time of the 5th day, and an execution time of 1 day.

[0070] Under the condition that time constraints are met, the specific execution time of each strategy can be determined through a sorting algorithm. For example, strategy S1 can be scheduled for execution between the 3rd and 4th days, and strategy S2 can be scheduled for execution on the 2nd day. In this way, the execution time of S1 and S2 is within their allowed time windows.

[0071] When generating a maintenance resource allocation plan based on the execution time window, it is necessary to ensure that the resource demand at each time point does not exceed the available resources. Assume that the system has 5 maintenance personnel available, strategy S1 requires 3 personnel, and strategy S2 requires 2 personnel. According to the previously determined execution schedule, executing S2 on day 2 requires 2 personnel, and executing S1 on days 3 and 4 requires 3 personnel. At no point in time does the demand exceed the 5 personnel available in the system, thus satisfying the resource constraint.

[0072] The generated intelligent maintenance plan includes the execution time and required resources for each maintenance strategy, ensuring the feasibility and efficiency of the plan. The plan fully considers factors such as risk reduction, timeliness, and resource constraints, providing a scientific basis for equipment maintenance decision-making.

[0073] A second aspect of an embodiment of the present invention provides a system for mining environmental protection equipment operation data and optimizing performance evaluation, including: The first unit is configured to collect real-time operating data of environmental protection equipment, construct a nonlinear annealing iterative model, perform high-dimensional state encoding processing on the real-time operating data, establish a component state correlation feature map using an entanglement mapping effect, calculate a fault propagation link between components based on the component state correlation feature map, generate a degradation trend curve for each component in combination with a probability measurement circuit, and calculate a dynamic risk index based on the fault propagation link and the degradation trend curve; A second unit is configured to construct a component-level fault risk assessment matrix based on the dynamic risk index and the fault propagation link, wherein the fault risk assessment matrix records the risk propagation weight, health status index, and remaining life prediction value of each component; A third unit is configured to construct a physical simulation environment for environmental protection equipment in the digital twin space, import the fault risk assessment matrix and the fault propagation link into the physical simulation environment, perform state evolution on the physical simulation environment using the nonlinear annealing iterative model, calculate the equipment performance change trajectory under different maintenance strategies based on the risk propagation weight, and generate a risk reduction index for each maintenance strategy; The fourth unit is used to prioritize the maintenance strategies according to the risk reduction index, select the maintenance strategy with the highest risk reduction index as the optimal maintenance strategy, and generate an intelligent maintenance plan based on the optimal maintenance strategy; update the fault propagation link and the risk propagation weight based on the status feedback data during the execution of the intelligent maintenance plan, and perform online optimization on the nonlinear annealing iterative model.

[0074] According to a third aspect of an embodiment of the present invention, an electronic device is provided, including: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the aforementioned method.

[0075] According to a fourth aspect of an embodiment of the present invention, a computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the method described above is implemented.

[0076] The present invention may be a method, an apparatus, a system and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for executing various aspects of the present invention.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for environmental protection equipment operation data mining and performance evaluation optimization, characterized in that: include: Collecting real-time operating data of environmental protection equipment, constructing a nonlinear annealing iterative model, performing high-dimensional state encoding processing on the real-time operating data, using the entanglement mapping effect to establish a component state correlation feature map, calculating the fault propagation link between components based on the component state correlation feature map, combining a probability measurement circuit to generate a degradation trend curve for each component, and calculating a dynamic risk index based on the fault propagation link and the degradation trend curve; Based on the dynamic risk index and the fault propagation link, a component-level fault risk assessment matrix is ​​constructed, wherein the fault risk assessment matrix records the risk propagation weight, health status index, and remaining life prediction value of each component; Constructing a physical simulation environment for environmental protection equipment in the digital twin space, importing the fault risk assessment matrix and the fault propagation link into the physical simulation environment, performing state evolution on the physical simulation environment using the nonlinear annealing iterative model, calculating equipment performance change trajectories under different maintenance strategies based on the risk propagation weights, and generating a risk reduction index for each maintenance strategy; Prioritizing the maintenance strategies according to the risk reduction index, selecting the maintenance strategy with the highest risk reduction index as the optimal maintenance strategy, and generating an intelligent maintenance plan based on the optimal maintenance strategy; The fault propagation link and the risk propagation weight are updated based on the status feedback data during the execution of the intelligent maintenance plan, and the nonlinear annealing iterative model is optimized online.

2. The method according to claim 1, characterized in that Performing high-dimensional state encoding processing on the real-time operation data, establishing a component state correlation feature map using an entanglement mapping effect, and calculating an inter-component fault propagation link based on the component state correlation feature map includes: Performing nonlinear decomposition on the real-time operation data using Hilbert-Huang transform to extract intrinsic mode components, calculating an instantaneous phase function based on the intrinsic mode components, and obtaining a characteristic sequence based on the instantaneous phase function; Performing a Hilbert-Huang secondary transform on the feature sequence to obtain a high-dimensional feature vector of the feature sequence, and mapping the high-dimensional feature vector to a multidimensional state space to generate a high-dimensional state vector; Performing an entangled state mapping transformation on the high-dimensional state vector to construct an entangled state vector, calculating a reduced density matrix of the entangled state vector, extracting eigenvalues ​​of the reduced density matrix, and calculating the entanglement correlation between components based on the eigenvalues; constructing a dynamic association matrix between components according to the entanglement correlation degree, optimizing and decomposing the dynamic association matrix between components using the Hilbert-Huang transform, and generating a component state association map, wherein the component state association map records the association strength between components; Performing Hilbert-Huang spectrum analysis on the correlation strength in the component state correlation map to calculate a characteristic spectrum sequence of the component state, quantifying the fault propagation strength between components based on the characteristic spectrum sequence, and constructing a fault propagation network; An importance index of each propagation path is calculated according to the propagation intensity in the fault propagation network, and an inter-component fault propagation link is identified based on the importance index.

3. The method according to claim 2, characterized in that Constructing a dynamic association matrix between components according to the entanglement correlation degree, optimizing and decomposing the dynamic association matrix between components using the Hilbert-Huang transform, and generating a component state association map includes: collecting state characteristics of multiple components, calculating entanglement correlations between the multiple components, and forming a component entanglement strength matrix based on the entanglement correlations, wherein the component entanglement strength matrix records the entanglement correlation strength between any two components; Dividing the component entanglement strength matrix into multiple time series matrices according to time windows, and constructing an inter-component dynamic association matrix, wherein the inter-component dynamic association matrix records the changing trend of the entanglement correlation between components in different time windows; Performing multi-scale decomposition of the dynamic correlation matrix between the components using Hilbert-Huang transform to obtain multiple intrinsic mode components, wherein the intrinsic mode components represent the fluctuation characteristics of the correlation relationship between the components at different time scales; Performing energy density analysis on the eigenmode components, calculating the energy contribution value of each of the eigenmode components, ranking the eigenmode components in importance based on the energy contribution values, and selecting the eigenmode components whose energy contribution values ​​are higher than a preset contribution threshold; The intrinsic mode components whose energy contribution values ​​are higher than a preset contribution threshold are reconstructed and superimposed to obtain an optimized inter-component dynamic association matrix, in which the main component association features are retained. A component state association map is generated based on the optimized inter-component dynamic association matrix.

4. The method according to claim 1, wherein Calculating the fault propagation link between components based on the component state association feature map, generating a degradation trend curve for each component in combination with a probability measurement circuit, and calculating the dynamic risk index based on the fault propagation link and the degradation trend curve includes: Extracting association weights between components from the component state association feature map, and calculating the node propagation capability of each component based on the association weights, where the node propagation capability is determined by the sum of the association weights of the component and other components; Analyzing the fault propagation relationship between components based on the node propagation capability, calculating the continuous product of the association weights between any two components, using the continuous product of the association weights as the propagation strength of the fault propagation link, and determining the propagation path based on the propagation strength; Acquiring operating status data of each component on the propagation path, performing probability distribution analysis on the operating status data using a probability measurement circuit to obtain a probability density function of the status data, and calculating an instantaneous degradation rate of each component based on the probability density function; Performing a time integration operation on the instantaneous degradation rate to obtain a cumulative degradation amount of the component, and constructing a degradation trend curve of the component based on the cumulative degradation amount, wherein the degradation trend curve represents a performance attenuation process of the component along the propagation path; The propagation intensity of the propagation path and the degradation trend curve are weighted and fused to generate a dynamic risk index reflecting the comprehensive risk of the component. The dynamic risk index also includes the impact degree of the component in the propagation network and the degradation status of its own performance.

5. The method according to claim 1, wherein Constructing a physical simulation environment for environmental protection equipment in the digital twin space, importing the fault risk assessment matrix and the fault propagation link into the physical simulation environment, and performing state evolution on the physical simulation environment through the nonlinear annealing iterative model includes: Collecting operating state parameters and feature point coordinate information of environmental protection equipment, building a digital twin geometric model based on the feature point coordinate information, and mapping the operating state parameters to the digital twin geometric model through a nonlinear state evolution equation; Obtain risk assessment data of the environmental protection equipment, perform spatial mapping conversion on the risk assessment data using the nonlinear state evolution equation, and generate a risk state matrix in the digital twin space, wherein the risk state matrix includes component failure risk degree information; Analyzing the fault propagation relationship between components based on the risk state matrix, extracting the fault propagation links and propagation weights between components, and converting the propagation weights into link weights in the digital twin space using the nonlinear state evolution equation; Substituting the risk state matrix and the link weight into the nonlinear state evolution equation, performing state iterative calculation in the digital twin geometric model, and obtaining a state evolution sequence of each component of the environmental protection equipment; The convergence error of the component state is calculated based on the state evolution sequence. When the convergence error is less than a preset error threshold, the steady-state eigenvector corresponding to the state evolution sequence is extracted. The state of the environmental protection equipment is evolved using the nonlinear state evolution equation according to the steady-state eigenvector to generate an evolution trajectory of the equipment state.

6. The method according to claim 1, characterized in that Prioritizing the maintenance strategies according to the risk reduction index, selecting the maintenance strategy with the highest risk reduction index as the optimal maintenance strategy, and generating an intelligent maintenance plan based on the optimal maintenance strategy includes: A maintenance strategy benefit function is constructed based on the risk reduction index, wherein the maintenance strategy benefit function comprehensively considers the risk reduction index, the strategy timeliness index, and the resource consumption index; a priority score is calculated for each maintenance strategy based on the maintenance strategy benefit function, wherein the priority score is obtained by dividing the maintenance strategy benefit function value by the maximum benefit function value among all strategies, and the priority score is used to represent the execution priority of the maintenance strategy; Combining maintenance strategies with similar priority scores into a strategy combination set, and calculating a combined benefit value of the strategy combination set based on the maintenance strategy benefit function, wherein the combined benefit value takes into account the weight contribution of each maintenance strategy in the strategy combination; Determine an optimal maintenance strategy combination according to the combined benefit value, and assign an execution time window to the maintenance strategy in the optimal maintenance strategy combination, wherein the execution time window is constrained by an earliest start time and a latest end time; A maintenance resource configuration plan is generated based on the execution time window, wherein the maintenance resource configuration plan ensures that the resource demand at each time point does not exceed the available resource amount, and outputs an intelligent maintenance plan that meets timing constraints and resource constraints.

7. A system for environmental protection equipment operation data mining and performance evaluation optimization, used to implement the method according to any one of claims 1 to 6, characterized in that: include: The first unit is configured to collect real-time operating data of environmental protection equipment, construct a nonlinear annealing iterative model, perform high-dimensional state encoding processing on the real-time operating data, establish a component state correlation feature map using an entanglement mapping effect, calculate a fault propagation link between components based on the component state correlation feature map, generate a degradation trend curve for each component in combination with a probability measurement circuit, and calculate a dynamic risk index based on the fault propagation link and the degradation trend curve; A second unit is configured to construct a component-level fault risk assessment matrix based on the dynamic risk index and the fault propagation link, wherein the fault risk assessment matrix records the risk propagation weight, health status index, and remaining life prediction value of each component; A third unit is configured to construct a physical simulation environment for environmental protection equipment in the digital twin space, import the fault risk assessment matrix and the fault propagation link into the physical simulation environment, perform state evolution on the physical simulation environment using the nonlinear annealing iterative model, calculate the equipment performance change trajectory under different maintenance strategies based on the risk propagation weight, and generate a risk reduction index for each maintenance strategy; A fourth unit is configured to prioritize the maintenance strategies according to the risk reduction index, select the maintenance strategy with the highest risk reduction index as the optimal maintenance strategy, and generate an intelligent maintenance plan based on the optimal maintenance strategy; The fault propagation link and the risk propagation weight are updated based on the status feedback data during the execution of the intelligent maintenance plan, and the nonlinear annealing iterative model is optimized online.

8. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images