Electrical operation and maintenance monitoring platform management system based on digital twinning
By building an electrical operation and maintenance monitoring platform using digital twin technology, the problems of low efficiency and fragmented integration in traditional electrical operation and maintenance monitoring systems are solved. This enables accurate fault prediction and optimized control of electrical equipment, improving equipment operating efficiency and stability.
Patent Information
- Application Number
- CN202511182016.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-22
- Publication Date
- 2025-11-21
AI Technical Summary
Traditional electrical operation and maintenance monitoring systems rely on offline data analysis and regular manual inspections, which cannot effectively cope with the complexity of operation and maintenance management and the real-time requirements brought about by the surge in the types and numbers of electrical equipment. The system integration is relatively fragmented, lacking unified standards and processes, resulting in poor coordination between various modules and affecting production control efficiency.
An electrical operation and maintenance monitoring platform management system based on digital twins is adopted. The system acquires equipment status data through the signal generation module, constructs state equations and generates control signals; the data simulation module uploads data to update the digital twin model and generates simulation results; the function optimization module constructs a loss function for optimization; and the equipment control module performs adaptive compensation to ensure that the equipment operates in the optimal state.
It enables accurate fault prediction and performance optimization of electrical equipment, improves equipment efficiency and stability, ensures equipment operates in optimal condition, and reduces operation and maintenance costs and risks.
Smart Images
Figure CN120999901A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of electrical monitoring technology, and in particular to an electrical operation and maintenance monitoring platform management system based on digital twins. Background Technology
[0002] With the continued growth of global energy demand and the accelerated transformation of the energy structure, the power system is facing an increasingly complex operating environment. The grid connection of large-scale renewable energy, the popularization of distributed energy, and the rapid development of smart grids have led to a surge in the types and quantities of electrical equipment, while also significantly increasing the complexity and real-time requirements of operation and maintenance management.
[0003] However, traditional electrical operation and maintenance monitoring systems rely heavily on offline data analysis and regular manual inspections, which cannot effectively address these challenges. In addition, the system integration of many electrical operation and maintenance monitoring platforms is relatively fragmented, lacking unified standards and processes, resulting in poor coordination between various modules and unsmooth data flow, thereby affecting the efficiency of production control. Summary of the Invention
[0004] This invention provides an electrical operation and maintenance monitoring platform management system based on digital twins, the main purpose of which is to solve the problem of low operation and maintenance efficiency of existing electrical operation and maintenance monitoring platform management systems based on digital twins.
[0005] To achieve the above objectives, this invention provides an electrical operation and maintenance monitoring platform management system based on digital twins. The system includes: a signal generation module, a data simulation module, a function optimization module, and an equipment control module. Specifically: The signal generation module is used to acquire electrical operating status data in the target device, construct the state equation of the target device based on the operating status data, calculate the target state information of the target device based on the state equation, and generate the control signal of the target device based on the target state information. The data simulation module is used to upload the running status data to the cloud to obtain the uploaded data, update the preset digital twin model according to the uploaded data to obtain the updated digital twin model, and generate the digital twin simulation result corresponding to the running status according to the updated digital twin model. The function optimization module is used to construct a loss function based on the digital twin simulation results, optimize and solve the loss function based on the target state information and the control signal, and obtain preliminary optimization results. The equipment control module verifies the preliminary optimization results, uses the verified preliminary optimization results as optimization data, calculates the difference between the target state information and the electrical sensor data in the operating state data, and performs adaptive compensation on the target equipment based on the difference, the control signal, and the optimization data to obtain the compensation result.
[0006] Optionally, when the signal generation module performs the function of constructing the state equation of the target device based on the operating state data, it is specifically used for: The operating status data is denoised to obtain denoised data; The denoised data is then subjected to outlier removal to obtain the data to be processed; Extract state variables from the data to be processed; Data features are extracted from the data to be processed to obtain data features; The state features and the state variables are concatenated to obtain the state equation.
[0007] Optionally, when the signal generation module performs the function of calculating the target state information of the target device based on the state equation, it is specifically used for: Calculate the current slope of the state equation; Substituting the current slope into the state equation yields the updated state equation; Based on the updated state equation and the current slope, the next state is predicted. The target state information is obtained by iterating the equations based on the next state and the state equation.
[0008] Optionally, when the data simulation module performs the function of uploading the running status data to the cloud to obtain the uploaded data, it is specifically used for: Principal component analysis was performed on the operational status data to obtain principal component data; The principal component data is encrypted to obtain encrypted data. A data transmission packet is constructed based on the encrypted data; Obtain the cloud upload endpoint, and upload the data transmission packet to the cloud upload endpoint using a time-sensitive network to obtain the uploaded data.
[0009] Optionally, when the data simulation module performs the function of updating the preset digital twin model based on the uploaded data to obtain the updated digital twin model, it is specifically used for: The uploaded data is parsed to obtain parsed data; The parsed data is filled with missing values to obtain the filled data; The filled data is timestamped to obtain aligned data; The alignment data and the preset digital twin model are used to perform dynamic parameter inversion to obtain inversion data; The digital twin model is calibrated based on the inversion data to obtain an updated digital twin model.
[0010] Optionally, when the function optimization module performs the function of constructing the loss function based on the digital twin simulation results, it is specifically used for: Key data indicators are extracted from the digital twin simulation results; The key data indicators are weighted and fused to obtain a single loss; The single loss is constrained to obtain the loss function.
[0011] Optionally, when the function optimization module performs the function of optimizing the loss function based on the target state information and the control signal to obtain preliminary optimization results, it is specifically used for: The target state information is discretized to obtain a discretized parameter space; The loss function is transformed to obtain a quadratic unconstrained binary optimization matrix. Constraints are constructed based on the control signal and the discretized parameter space; A preliminary optimization result is obtained by performing a hybrid solution based on the quadratic unconstrained binary optimization matrix, the constraints, and the discretized parameter space.
[0012] Optionally, when the function optimization module performs the function of performing a hybrid solution based on the quadratic unconstrained binary optimization matrix and the discretized parameter space to obtain preliminary optimization results, it is specifically used for: Using the discretized parameter space, a low-energy solution search is performed on the quadratic unconstrained binary optimization matrix to obtain candidate parameters; The gradient of the quadratic unconstrained binary optimization matrix is calculated using the candidate parameters; The quadratic unconstrained binary optimization matrix is locally optimized based on the gradient to obtain preliminary optimization results.
[0013] Optionally, when the device control module performs the function of verifying the preliminary optimization results and using the verified preliminary optimization results as optimization data, it is specifically used for: Calculate the Lyapunov index based on the preliminary optimization results; The stability of the preliminary optimization results was verified using the Lyapunov index, and the stability verification results were obtained. The preliminary optimization results are then subjected to boundary condition verification to obtain the boundary verification results. Based on the boundary verification results and the stability verification results, intermediate optimization results are selected from the preliminary optimization results; The intermediate optimization results are subjected to edge device shadow testing to obtain edge verification results; Optimized data is selected from the intermediate optimization results based on the edge verification results.
[0014] Optionally, when the device control module performs the function of adaptively compensating the target device based on the difference, the control signal, and the optimization data to obtain a compensation result, it is specifically used for: A gain matrix is constructed based on the optimized data; The signal is reconstructed based on the difference, the gain matrix, and the control signal to obtain the compensation signal; Collect real-time operating status data of the target device, and calculate the compensation increment of the target device based on the compensation signal and the real-time operating status data; The target device is adjusted using the compensation increment to obtain the compensation result.
[0015] This invention utilizes digital twin technology to model equipment and update the model in real time, enabling more accurate prediction of potential future equipment failures or performance degradation, thereby improving the operating efficiency of electrical equipment. By adaptively compensating for discrepancies between the target equipment's state information and actual sensor data, it ensures the equipment operates in its optimal state. Therefore, the electrical operation and maintenance monitoring platform management system proposed in this invention can solve the problem of low operational efficiency in existing digital twin-based electrical operation and maintenance monitoring platform management systems. Attached Figure Description
[0016] Figure 1 A functional module diagram of an electrical operation and maintenance monitoring platform management system based on digital twins, provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the process for constructing a state equation according to an embodiment of the present invention; Figure 3 A schematic diagram of a cloud upload process provided in an embodiment of the present invention; Figure 4 This is a flowchart illustrating a management method for an electrical operation and maintenance monitoring platform based on digital twins, as provided in an embodiment of the present invention.
[0017] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation
[0018] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
[0019] Reference Figure 1 The diagram shown illustrates the functional modules of an electrical operation and maintenance monitoring platform management system based on digital twins, according to an embodiment of the present invention. In this embodiment, the electrical operation and maintenance monitoring platform management system 100 based on digital twins can be installed in an electronic device. Depending on the functions implemented, the electrical operation and maintenance monitoring platform management system 100 based on digital twins may include a signal generation module 101, a data simulation module 102, a function optimization module 103, and a device control module 104. The modules described in this invention can also be referred to as units, which are a series of computer program segments that can be executed by the processor of an electronic device and can perform fixed functions, stored in the memory of the electronic device.
[0020] In this embodiment of the invention, the model building module 101 includes acquiring electrical operating status data of the target device, constructing a state equation of the target device based on the operating status data, calculating target state information of the target device based on the state equation, and generating control signals for the target device based on the target state information. In this embodiment of the invention, the signal generation module 102 includes uploading the running status data to the cloud to obtain uploaded data, updating the preset digital twin model according to the uploaded data to obtain an updated digital twin model, and generating a digital twin simulation result corresponding to the running status according to the updated digital twin model. In this embodiment of the invention, the data simulation module 103 includes constructing a loss function based on the digital twin simulation results, constructing a loss function based on the digital twin simulation results, optimizing the loss function based on the target state information and the control signal, and obtaining preliminary optimization results; In this embodiment of the invention, the device control module 104 includes verifying the preliminary optimization results, using the verified preliminary optimization results as optimization data, calculating the difference between the target state information and the electrical sensor data in the operating state data, and performing adaptive compensation on the target device based on the difference, the control signal, and the optimization data to obtain a compensation result.
[0021] In detail, the modules described in the electrical operation and maintenance monitoring platform management system 100 based on digital twins in this embodiment of the invention use the same technical means as those in the electrical operation and maintenance monitoring platform management system based on digital twins shown in the accompanying drawings, and can produce the same technical effects, which will not be repeated here.
[0022] The following describes, with reference to specific embodiments, each component and its specific workflow of the electrical operation and maintenance monitoring platform management system based on digital twins: The signal generation module 101 is used to acquire electrical operating status data in the target device, construct the state equation of the target device based on the operating status data, calculate the target state information of the target device based on the state equation, and generate the control signal of the target device based on the target state information.
[0023] In this embodiment of the invention, by acquiring operational status data in real time and constructing state equations, the current operational status of the target device can be accurately understood, and potential problems can be detected in a timely manner.
[0024] In this embodiment of the invention, the operating status data refers to various key parameters and status information collected in real time or periodically from electrical equipment and its operating environment, including electrical sensor data (such as voltage, current, power and energy data), equipment status data (such as temperature of key components, mechanical status data, etc.), and control signal and actuator status data (such as controller output signal, actual actuator response, etc.).
[0025] See Figure 2 As shown in this embodiment of the invention, when the signal generation module performs the function of constructing the state equation of the target device based on the operating state data, it is specifically used for: S21. Denoise the operating status data to obtain denoised data; S22. Remove outliers from the denoised data to obtain the data to be processed; S23. Extract state variables from the data to be processed; S24. Extract data features from the data to be processed to obtain data features; S25. Perform state concatenation on the data features and the state variables to obtain the state equation.
[0026] In this embodiment of the invention, the data denoising is achieved by performing multi-scale decomposition of the signal on the running status data using a preset wavelet basis, then performing threshold filtering based on the decomposition results to obtain filtering results, and finally performing inverse wavelet transform based on the filtering results to obtain denoised data.
[0027] In this embodiment of the invention, outlier removal is achieved by calculating the mean and standard deviation, removing data and subtracting data whose absolute value of the mean is greater than three times the standard deviation, to obtain the data to be processed.
[0028] In this embodiment of the invention, the extraction of state variables refers to performing data clustering on the data to be processed, and extracting state variables with strong separability by maximizing the inter-class variance.
[0029] In this embodiment of the invention, the data feature extraction is obtained by using a residual neural network with a preset number of layers to extract features from the data to be processed. By concatenating the data to be processed, the state dimension and control dimension corresponding to the full-dimensional data are obtained. Then, the full-dimensional data is extracted using several residual blocks to obtain several layers of dimensions, which are the data features.
[0030] In this embodiment of the invention, features and state variables are fused into a state vector to construct a differential equation. The differential equation is of the form: dx / dt=f(x,u) Where x is the state variable, u is the control input in the data feature, and f(x,u) is the state equation when the state variable is x and the control input is u.
[0031] In this embodiment of the invention, when the signal generation module performs the function of calculating the target state information of the target device based on the state equation, it is specifically used for: Calculate the current slope of the state equation; Substituting the current slope into the state equation yields the updated state equation; Based on the updated state equation and the current slope, the next state is predicted. The target state information is obtained by iterating the equations based on the next state and the state equation.
[0032] In this embodiment of the invention, the rate of change of the state at the current moment is calculated using a state equation. For example, in electrical equipment, the state may be parameters such as current and voltage, while the control signal may be an external adjustment signal. The current slope refers to the rate of change of the system at a certain moment, obtained through the state equation based on the current state and the control signal.
[0033] In this embodiment of the invention, by substituting the current state and control signal into the state equation, the rate of change of the state at that moment (i.e., the slope) can be obtained, thereby providing a basis for the next prediction. The purpose is to deduce the state change trend of the system at the next moment based on the current state and control signal.
[0034] In this embodiment of the invention, the next state is predicted based on the updated state equation and the current slope. This can be achieved by the Euler method or other numerical integration methods. The prediction formula of the Euler method is: x(t+Δt)=x(t)+x(-)(t)Δt, where Δt is the time step, x(-)(t) is the current slope, and x(t+Δt) is the next state.
[0035] In this embodiment of the invention, the equation iteration refers to using the state equation to perform multiple iterations until the system reaches the target state. This means that by continuously updating the state and control signals, the state of the system converges to the expected target state.
[0036] In this embodiment of the invention, by continuously adjusting the system state to approach the target state, the optimal operation of the target device can ultimately be achieved.
[0037] In this embodiment of the invention, the control signal is generated to enable the operating state of the target device to approach or maintain the value of the target state information, and a control strategy or control algorithm is implemented based on the target state information. The difference between the target state information and the operating state data is obtained by comparing the target state information with the operating state data. The control algorithm (such as a PID control algorithm or an adaptive control algorithm) is then used to calculate the adjustable controller parameters based on the error, thus obtaining the control signal.
[0038] In this embodiment of the invention, by acquiring real-time operating status data and constructing state equations, accurate modeling of the target device's operating status is achieved, thereby dynamically adjusting the device's operating parameters and improving the system's response speed and operating efficiency.
[0039] The data simulation module 102 is used to upload the operating status data to the cloud to obtain uploaded data, update the preset digital twin model according to the uploaded data to obtain an updated digital twin model, and generate the digital twin simulation result corresponding to the operating status according to the updated digital twin model.
[0040] In this embodiment of the invention, actual monitoring data is uploaded to the cloud for model updates, ensuring that the digital twin model can reflect the latest equipment status and enhancing the model's realism and reliability.
[0041] See Figure 3 As shown, when the data simulation module performs the function of uploading the running status data to the cloud to obtain the uploaded data, it is specifically used for: S31. Perform principal component analysis on the operating status data to obtain principal component data; S32. Encrypt the principal component data to obtain encrypted data; S33. Construct a data transmission packet based on the encrypted data; S34. Obtain the cloud upload endpoint, and upload the data transmission packet to the cloud upload endpoint using a time-sensitive network to obtain the uploaded data.
[0042] In this embodiment of the invention, the principal component analysis is performed by standardizing or normalizing the running state data to ensure that different features have the same scale, calculating the covariance matrix of the standardized dataset, solving for the eigenvalues and eigenvectors of the covariance matrix, sorting the eigenvalues by size, selecting the eigenvectors corresponding to the n largest eigenvalues as principal components, and using the selected principal components to transform the original data to a new coordinate system to obtain the dimensionality-reduced principal component data.
[0043] In this embodiment of the invention, a public key and a private key pair are generated for the principal component data. The public key is used to encrypt the principal component data. The corresponding data items in the principal component data are multiplied by the public key to obtain encrypted data.
[0044] In this embodiment of the invention, the format of the data transmission packet is determined, including header information (such as source address and destination address), metadata (such as timestamp and data type) and body (encrypted data). According to the defined data structure, the encrypted data and other relevant information are combined into a complete data transmission packet.
[0045] In this embodiment of the invention, a network environment supporting the TSN protocol is set up to ensure that the network has low latency, high bandwidth and high reliability. The target address (IP address and port number, etc.) for data upload is obtained from the cloud service provider. The constructed data transmission packet is sent to the specified cloud address through the configured TSN network. The cloud server is checked to see if it has successfully received the data and to confirm the data integrity.
[0046] In this embodiment of the invention, when the data simulation module performs the function of updating the preset digital twin model based on the uploaded data to obtain the updated digital twin model, it is specifically used for: The uploaded data is parsed to obtain parsed data; The parsed data is filled with missing values to obtain the filled data; The filled data is timestamped to obtain aligned data; The alignment data and the preset digital twin model are used to perform dynamic parameter inversion to obtain inversion data; The digital twin model is calibrated based on the inversion data to obtain an updated digital twin model.
[0047] In this embodiment of the invention, the encoding format of the uploaded data (such as JSON, XML, etc.) is determined, and the data transmission packet is decoded using a corresponding parsing library (such as the json module in Python). Useful information is extracted from the decoded data to form parsed data.
[0048] In this embodiment of the invention, data analysis tools (such as Pandas) are used to check for missing values in the parsed data, and an appropriate imputation method is selected based on the data characteristics and requirements. For example, for time series data, linear interpolation can be used, and the selected imputation method can be applied to fill in the missing values to generate an imputed dataset.
[0049] In this embodiment of the invention, the timestamp format of all data is ensured to be consistent (such as the ISO 8601 standard), and time resampling technology (such as the resample function of Pandas) is used to align all data on the same time interval. Multi-source data is merged through inner join or other methods to share the same time axis and obtain aligned data.
[0050] In this embodiment of the invention, the dynamic parameter inversion is achieved by providing the aligned data as input to the digital twin model, and the model calculates the state or behavior of the system (such as temperature, current, etc.) based on the aligned data to obtain the inversion data.
[0051] In this embodiment of the invention, the optimal parameter values obtained during the inversion process are used to replace the corresponding parameters in the original model, and the performance of the updated model is tested using a portion of data that did not participate in the inversion, thereby obtaining an updated digital twin model.
[0052] In this embodiment of the invention, data calibration involves adjusting the digital twin model based on the inverted data to make the model's parameters and behavior more consistent with the actual state of the real system. First, it is necessary to verify whether the existing digital twin model is consistent with the actual system behavior, analyzing the differences between the inverted data and the actual data. These differences are typically represented in an error function. Based on the results of the error analysis, the parameters in the digital twin model are adjusted. For example, if the temperature predicted by the model differs significantly from the actual measured temperature, it may be necessary to adjust parameters such as the thermal conductivity coefficient.
[0053] In this embodiment of the invention, the timeliness and accuracy of the digital twin model are ensured by uploading the latest operational status data to the cloud and updating the model. Furthermore, by generating simulation results using the updated digital twin model, the operating status of the equipment can be simulated in a virtual environment, potential problems can be identified in advance, and operational procedures can be optimized, reducing risks in actual operation.
[0054] Furthermore, the generation of simulation results using the updated digital twin model involves simulating the operational state based on the updated digital twin model to obtain simulation results. The simulation will simulate the behavior of the equipment under different operating conditions, such as the response under different loads, ambient temperatures, and operating parameters.
[0055] In this embodiment of the invention, by performing simulations using a continuously updated model, the behavior of the equipment can be predicted, the operation process optimized, and potential problems identified in advance without affecting actual production.
[0056] The function optimization module 103 is used to construct a loss function based on the digital twin simulation results, optimize the loss function based on the target state information and the control signal, and obtain preliminary optimization results.
[0057] In this embodiment of the invention, a loss function is constructed using digital twin simulation results, and then optimized and solved to find the optimal or near-optimal operating parameter settings. This allows for minimizing energy consumption or other key performance indicators while ensuring equipment performance.
[0058] In this embodiment of the invention, when the function optimization module performs the function of constructing a loss function based on the digital twin simulation results, it is specifically used for: Key data indicators are extracted from the digital twin simulation results; The key data indicators are weighted and fused to obtain a single loss; The single loss is constrained to obtain the loss function.
[0059] In this embodiment of the invention, the equipment operating efficiency, energy consumption level, and temperature fluctuation range are calculated based on the digital twin simulation results to obtain key indicators. These key indicators are then normalized to obtain data key indicators. The equipment operating efficiency is typically calculated as the ratio between the equipment's output power and its input power. The energy consumption level typically refers to the energy consumed by the equipment per unit time, obtained by calculating the total energy consumed by the equipment in the simulation results and the total operating time of the equipment. The temperature fluctuation range refers to the range of temperature changes during equipment operation, especially under load fluctuations and changes in ambient temperature. The temperature fluctuation range can be obtained by simulating the temperature changes of the equipment under different operating conditions using a digital twin simulation model.
[0060] In this embodiment of the invention, a comprehensive single loss is generated by weighted fusion of multiple key indicators.
[0061] In this embodiment of the invention, constraints are constructed based on the single loss, the constraints are converted into penalty terms, the single loss value is combined with the constraints to form a complete loss function, and the penalty terms are added to the single loss. The constraints include the maximum load of the equipment, temperature range, maximum power, the control signal (voltage or current) of the motor having an effective operating range, and the response time of the equipment should complete the operation within a certain time frame, etc.
[0062] In this embodiment of the invention, when the function optimization module performs the function of optimizing the loss function based on the target state information and the control signal to obtain preliminary optimization results, it is specifically used for: The target state information is discretized to obtain a discretized parameter space; The loss function is transformed to obtain a quadratic unconstrained binary optimization matrix. Constraints are constructed based on the control signal and the discretized parameter space; A preliminary optimization result is obtained by performing a hybrid solution based on the quadratic unconstrained binary optimization matrix, the constraints, and the discretized parameter space.
[0063] In this embodiment of the invention, the data discretization refers to defining a discretization interval based on the actual range of the target state information (such as minimum value, maximum value), and obtaining a discretization parameter space by dividing the data range into several equally spaced intervals.
[0064] In this embodiment of the invention, the data transformation is achieved by mathematical modeling, expressing the loss function as a quadratic polynomial, and filling its coefficients into a quadratic unconstrained binary matrix to obtain a quadratic unconstrained binary optimization matrix.
[0065] In this embodiment of the invention, the control signal is calculated and generated based on target state information and is used to regulate the operation of the equipment. The constraint conditions constructed based on the control signal and the discretized parameter space are achieved by setting the maximum and minimum values of the control signal according to the nature of the control signal and the actual operational limitations of the equipment. The discretized parameter space reflects multiple discrete intervals of the target state, each interval corresponding to a different equipment operating state. Corresponding constraint conditions are set based on the discretized target state intervals. After constraining the control signal and the discretized parameter space, composite constraint conditions need to be generated. By superimposing their respective constraint conditions, the constraint conditions can collectively ensure that changes in the control signal and the target state meet the requirements of actual operation during the optimization process. For example, if the control signal is voltage and the target state is temperature, then the voltage adjustment must not cause the temperature to exceed a predetermined safe range.
[0066] In this embodiment of the invention, when the function optimization module performs the function of performing a hybrid solution based on the quadratic unconstrained binary optimization matrix and the discretized parameter space to obtain preliminary optimization results, it is specifically used for: Using the discretized parameter space, a low-energy solution search is performed on the quadratic unconstrained binary optimization matrix to obtain candidate parameters; The gradient of the quadratic unconstrained binary optimization matrix is calculated using the candidate parameters; The quadratic unconstrained binary optimization matrix is locally optimized based on the gradient to obtain preliminary optimization results.
[0067] In this embodiment of the invention, the parameter values are iteratively adjusted in the discretized parameter space to find the solution that minimizes the quadratic unconstrained binary optimization matrix. During the search process, the energy value of each iteration is recorded, the current optimal solution is retained, and several solutions with the lowest energy are selected as candidate parameters.
[0068] In this embodiment of the invention, an objective function is constructed based on the quadratic unconstrained binary optimization matrix, the gradient of the objective function is calculated, candidate parameters are substituted into the gradient formula, and the gradient value of the objective function at that point is calculated. The gradient is stored in vector form, representing the optimization direction and magnitude of each parameter.
[0069] Furthermore, by solving the quadratic unconstrained binary optimization matrix into univariate terms (linear terms) and multivariate interaction terms (quadratic terms), the diagonal elements of the quadratic unconstrained binary optimization matrix are used as coefficients of the univariate variables themselves, and the off-diagonal elements are used as coefficients of the interactions between variables, thus obtaining the objective function. The gradient value is obtained by calculating the derivative of the objective function with respect to each candidate parameter.
[0070] In this embodiment of the invention, the local optimization is achieved by inputting candidate parameters and gradient values, setting the learning rate (step size) and the maximum number of iterations, updating the parameters using the gradient descent formula, repeatedly calculating the gradient and updating the parameters, until the preset maximum number of iterations is reached or the gradient value is close to zero, indicating that a local optimum has been found or the energy value no longer decreases significantly, thus obtaining a preliminary optimization result.
[0071] In this embodiment of the invention, a set of high-quality candidate parameters can be effectively found by searching for low-energy solutions (i.e., the solutions with the lowest cost or the best performance) in the discretized parameter space.
[0072] In this embodiment of the invention, by constructing a loss function and optimizing it, the optimal solution can be found for key indicators such as energy consumption and efficiency. The preliminary results after optimization provide a more scientific and efficient basis for subsequent equipment control, further improving the overall performance of the system.
[0073] The device control module 104 verifies the preliminary optimization results, uses the verified preliminary optimization results as optimization data, calculates the difference between the target state information and the electrical sensor data in the operating state data, and performs adaptive compensation on the target device based on the difference, the control signal, and the optimization data to obtain the compensation result.
[0074] In this embodiment of the invention, by verifying the preliminary optimization results and adaptively compensating the target device based on the difference and optimization data, the working state of the device can be dynamically adjusted so that it is always near the optimal working point.
[0075] In this embodiment of the invention, when the device control module performs the function of verifying the preliminary optimization results and using the verified preliminary optimization results as optimization data, it is specifically used for: Calculate the Lyapunov index based on the preliminary optimization results; The stability of the preliminary optimization results was verified using the Lyapunov index, and the stability verification results were obtained. The preliminary optimization results are then subjected to boundary condition verification to obtain the boundary verification results. Based on the boundary verification results and the stability verification results, intermediate optimization results are selected from the preliminary optimization results; The intermediate optimization results are subjected to edge device shadow testing to obtain edge verification results; Optimized data is selected from the intermediate optimization results based on the edge verification results.
[0076] In this embodiment of the invention, the calculation of the Lyapunov exponent involves introducing a disturbance based on the initial state of the system corresponding to the preliminary optimization result. As the system evolves, this disturbance changes over time. If the system is stable, the disturbance gradually decreases over time; if the system is chaotic, the disturbance increases exponentially. The Lyapunov exponent is obtained by calculating the ratio of the disturbance at each moment to the disturbance at the initial state, and then performing a logarithmic calculation on this ratio.
[0077] In this embodiment of the invention, data where the Lyapunov index is less than 0 are selected as the stability results that pass the stability verification.
[0078] In this embodiment of the invention, the effective range (such as minimum value, maximum value), physical constraints, etc. of each parameter are determined, and each preliminary optimization result is traversed to check whether it exceeds the defined boundary conditions.
[0079] In this embodiment of the invention, the filtering refers to checking the output value of each optimization result (e.g., the temperature and current of the device) to verify whether it is within a predetermined legal range. If an optimization result exceeds the set boundary conditions (e.g., the device temperature exceeds the safe range), the result will be excluded. Only those optimization results that meet the boundary conditions will be retained as intermediate optimization results.
[0080] In this embodiment of the invention, the digital twin model of the edge device is updated using intermediate optimization results, simulation tests are run in a virtual environment, the system response and performance indicators are observed, and various performance indicators during the test are analyzed and recorded to evaluate the effect of the intermediate optimization results.
[0081] In this embodiment of the invention, the shadow test refers to simulating the optimization results in a virtual environment, simulating device operation, and comparing it with actual data. The shadow test can simulate the device's performance under different operating conditions to ensure the feasibility of the optimization results in real-world applications. Using the intermediate optimization results as input, the virtual device is run, and its operating status, output results, and device behavior are monitored. By comparing the simulated results with actual data, the performance of the optimization results in the real-world environment is evaluated. If the shadow test results show that the device can operate stably under the expected conditions and meet the expected performance goals, then the optimization result is considered valid.
[0082] In this embodiment of the invention, the performance of each intermediate optimization result is evaluated based on the results of the edge device shadow test. According to the set performance standards (such as lowest energy consumption and highest efficiency), the best-performing intermediate optimization result is selected as the final optimization data.
[0083] In this embodiment of the invention, by evaluating the impact of preliminary optimization results on the long-term behavior of the system, it is ensured that the proposed control strategy will not lead to unstable or chaotic behavior of the system.
[0084] In this embodiment of the invention, when the device control module performs the function of adaptively compensating the target device based on the difference, the control signal, and the optimization data to obtain a compensation result, it is specifically used for: A gain matrix is constructed based on the optimized data; The signal is reconstructed based on the difference, the gain matrix, and the control signal to obtain the compensation signal; Collect real-time operating status data of the target device, and calculate the compensation increment of the target device based on the compensation signal and the real-time operating status data; The target device is adjusted using the compensation increment to obtain the compensation result.
[0085] In this embodiment of the invention, the gain matrix is constructed based on optimization data. A gain matrix is constructed by analyzing the target device's operating status data, control signals, and optimization data. The function of this matrix is to generate appropriate compensation signals by adjusting the target device's state and control signals, thereby adjusting the device's operating state within the automatic control system.
[0086] Furthermore, the gain matrix is constructed by modeling the relationship between the device operating state and the control signal using linear or nonlinear functions. The specific construction process may include: using the operating state data and a linear regression model to derive the relationship between the control signal and the device state, thus obtaining the gain matrix. Each element of the gain matrix corresponds to a regression coefficient between the input signal and the output state.
[0087] In this embodiment of the invention, the control signal is transformed by a gain matrix to generate a compensation signal. A gain matrix K is introduced into the control system to transform the original control signal u to obtain the compensation signal.
[0088] In this embodiment of the invention, the difference between real-time monitoring data and expected values is analyzed, the compensation signal is adjusted according to the difference, and the compensation increment is calculated.
[0089] Furthermore, signal reconstruction was performed using the gain matrix and control signal to obtain a compensation signal generated according to an optimization algorithm. This compensation signal is used to adjust the control input of the equipment to make its operating state as close as possible to the ideal state or to achieve the optimization target. Based on the compensation signal and real-time operating status data, a compensation increment is generated, which is the amount of adjustment the equipment needs to make. The goal of the compensation increment is to adjust the equipment's operation based on the deviation between the current state and the compensation signal. For example, assuming the target temperature of the equipment is 100°C, and the current temperature is 90°C, the compensation increment might be a positive value to increase the temperature. The compensation signal is the input signal controlling the equipment's heater. The gain matrix can be used to calculate the required adjustment amount, i.e., the compensation increment, based on the equipment's thermal characteristics.
[0090] In this embodiment of the invention, an automatic control system such as a PLC or DCS executes the adjustment operation corresponding to the compensation increment. The compensation increment is used as an adjustment signal input, and the control system performs necessary adjustments to the equipment based on the compensation increment. The adjustment operation typically includes adjusting the equipment's input parameters (such as voltage, current, temperature, etc.) or changing the equipment's operating mode, thereby keeping the equipment in its optimal operating state as much as possible.
[0091] In this embodiment of the invention, by constructing a gain matrix using optimized data, precise adjustments can be made to the specific operating state and historical data of the target device, enabling the control system to respond more quickly and accurately to changes and abnormal situations during device operation, thereby improving the response speed and accuracy of the control system.
[0092] In this embodiment of the invention, the feasibility and effectiveness of the preliminary optimization results are verified to ensure their practical application, thereby avoiding unreasonable control strategies. By calculating the difference and performing adaptive compensation, the operating parameters of the equipment can be dynamically adjusted to reduce errors, improve the stability and accuracy of equipment operation, and extend the service life of the equipment.
[0093] like Figure 4 The diagram shown is a flowchart illustrating a digital twin-based electrical operation and maintenance monitoring platform management method according to an embodiment of the present invention. In this embodiment, the digital twin-based electrical operation and maintenance monitoring platform management method includes: S401. Obtain the electrical operating status data of the target device, construct the state equation of the target device based on the operating status data, calculate the target state information of the target device based on the state equation, and generate the control signal of the target device based on the target state information; S402. Upload the running status data to the cloud to obtain uploaded data. Update the preset digital twin model according to the uploaded data to obtain an updated digital twin model. Generate the digital twin simulation result corresponding to the running status according to the updated digital twin model. S403. Construct a loss function based on the digital twin simulation results, optimize the loss function based on the target state information and the control signal, and obtain preliminary optimization results; S404. Verify the preliminary optimization results, use the verified preliminary optimization results as optimization data, calculate the difference between the target state information and the electrical sensor data in the operating state data, and perform adaptive compensation on the target device based on the difference, the control signal and the optimization data to obtain the compensation result.
[0094] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology.
[0095] Artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.
[0096] Furthermore, it is clear that the word "including" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or systems described in the system can also be implemented by a single unit or system through software or hardware. Terms such as "first," "second," etc., are used to indicate names and do not indicate any specific order.
[0097] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.
Claims
1. An electrical operation and maintenance monitoring platform management system based on digital twins, characterized in that, The system includes: a signal generation module, a data simulation module, a function optimization module, and a device control module, specifically: The signal generation module is used to acquire electrical operating status data in the target device, construct the state equation of the target device based on the operating status data, calculate the target state information of the target device based on the state equation, and generate the control signal of the target device based on the target state information. The data simulation module is used to upload the running status data to the cloud to obtain the uploaded data, update the preset digital twin model according to the uploaded data to obtain the updated digital twin model, and generate the digital twin simulation result corresponding to the running status according to the updated digital twin model. The function optimization module is used to construct a loss function based on the digital twin simulation results, optimize and solve the loss function based on the target state information and the control signal, and obtain preliminary optimization results. The equipment control module verifies the preliminary optimization results, uses the verified preliminary optimization results as optimization data, calculates the difference between the target state information and the electrical sensor data in the operating state data, and performs adaptive compensation on the target equipment based on the difference, the control signal, and the optimization data to obtain the compensation result.
2. The electrical operation and maintenance monitoring platform management system based on digital twin as described in claim 1, characterized in that, When the signal generation module performs the function of constructing the state equation of the target device based on the operating state data, it is specifically used for: The operating status data is denoised to obtain denoised data; The denoised data is then subjected to outlier removal to obtain the data to be processed; Extract state variables from the data to be processed; Data features are extracted from the data to be processed to obtain data features; The state features and the state variables are concatenated to obtain the state equation.
3. The electrical operation and maintenance monitoring platform management system based on digital twin as described in claim 1, characterized in that, When the signal generation module performs the function of calculating the target state information of the target device based on the state equation, it is specifically used for: Calculate the current slope of the state equation; Substituting the current slope into the state equation yields the updated state equation; Based on the updated state equation and the current slope, the next state is predicted. The target state information is obtained by iterating the equations based on the next state and the state equation.
4. The electrical operation and maintenance monitoring platform management system based on digital twin as described in claim 1, characterized in that, When the data simulation module performs the function of uploading the running status data to the cloud and obtaining the uploaded data, it is specifically used for: Principal component analysis was performed on the operational status data to obtain principal component data; The principal component data is encrypted to obtain encrypted data. A data transmission packet is constructed based on the encrypted data; Obtain the cloud upload endpoint, and upload the data transmission packet to the cloud upload endpoint using a time-sensitive network to obtain the uploaded data.
5. The electrical operation and maintenance monitoring platform management system based on digital twin as described in claim 1, characterized in that, When the data simulation module performs the function of updating the preset digital twin model based on the uploaded data to obtain the updated digital twin model, it is specifically used for: The uploaded data is parsed to obtain parsed data; The parsed data is filled with missing values to obtain the filled data; The filled data is timestamped to obtain aligned data; The alignment data and the preset digital twin model are used to perform dynamic parameter inversion to obtain inversion data; The digital twin model is calibrated based on the inversion data to obtain an updated digital twin model.
6. The electrical operation and maintenance monitoring platform management system based on digital twin as described in claim 1, characterized in that, When the function optimization module executes the function of constructing the loss function based on the digital twin simulation results, it is specifically used for: Key data indicators are extracted from the digital twin simulation results; The key data indicators are weighted and fused to obtain a single loss; The single loss is constrained to obtain the loss function.
7. The electrical operation and maintenance monitoring platform management system based on digital twin as described in claim 1, characterized in that, When the function optimization module performs the function of optimizing the loss function based on the target state information and the control signal to obtain preliminary optimization results, it is specifically used for: The target state information is discretized to obtain a discretized parameter space; The loss function is transformed to obtain a quadratic unconstrained binary optimization matrix. Constraints are constructed based on the control signal and the discretized parameter space; A preliminary optimization result is obtained by performing a hybrid solution based on the quadratic unconstrained binary optimization matrix, the constraints, and the discretized parameter space.
8. The electrical operation and maintenance monitoring platform management system based on digital twin as described in claim 7, characterized in that, When the function optimization module performs the function of performing a hybrid solution based on the quadratic unconstrained binary optimization matrix and the discretized parameter space to obtain preliminary optimization results, it is specifically used for: Using the discretized parameter space, a low-energy solution search is performed on the quadratic unconstrained binary optimization matrix to obtain candidate parameters; The gradient of the quadratic unconstrained binary optimization matrix is calculated using the candidate parameters; The quadratic unconstrained binary optimization matrix is locally optimized based on the gradient to obtain preliminary optimization results.
9. The electrical operation and maintenance monitoring platform management system based on digital twin as described in claim 1, characterized in that, When the device control module performs the function of verifying the preliminary optimization results and using the verified preliminary optimization results as optimization data, it is specifically used for: Calculate the Lyapunov index based on the preliminary optimization results; The stability of the preliminary optimization results was verified using the Lyapunov index, and the stability verification results were obtained. The preliminary optimization results are then subjected to boundary condition verification to obtain the boundary verification results. Based on the boundary verification results and the stability verification results, intermediate optimization results are selected from the preliminary optimization results; The intermediate optimization results are subjected to edge device shadow testing to obtain edge verification results; Optimized data is selected from the intermediate optimization results based on the edge verification results.
10. The electrical operation and maintenance monitoring platform management system based on digital twin as described in claim 1, characterized in that, When the device control module performs the function of adaptively compensating the target device based on the difference, the control signal, and the optimization data to obtain a compensation result, it is specifically used for: A gain matrix is constructed based on the optimized data; The signal is reconstructed based on the difference, the gain matrix, and the control signal to obtain the compensation signal; Collect real-time operating status data of the target device, and calculate the compensation increment of the target device based on the compensation signal and the real-time operating status data; The target device is adjusted using the compensation increment to obtain the compensation result.