A data-driven power system simulation optimization method

Abstract

The present invention belongs to the field of power system simulation application technology, and in particular relates to a data-driven power system simulation optimization method. Based on a data-driven dynamic adaptive time-space segmentation and collaborative model, the fast-changing electromagnetic transient process is mapped to a quantum processor, the slow-changing electromechanical transient process is retained in the classical computer, the grid node voltage is encoded as a quantum bit amplitude, the power network dynamics is equivalent to a quantum Hamiltonian, the quantum domain of the fast-changing electromagnetic transient and the classical domain of the slow-changing electromechanical dynamics are decoupled, and adaptive balance simulation accuracy and efficiency are achieved. Quantum computing sparse codes and dynamic error correction are constructed, non-uniform code distance distribution is used to reduce cross-scale data transmission errors, avoid high-frequency harmonic distortion, and achieve high-fault-tolerant simulation under limited quantum resources. The code distance is dynamically adjusted through the reinforcement learning DDPG strategy to respond to noise fluctuations and grid operating conditions in real time, optimizing resources and fault-tolerant balance.

Description

Technical Field

[0001] The present invention belongs to the technical field of power system simulation applications, and in particular relates to a data-driven power system simulation optimization method. Background Art

[0002] With the large-scale integration of renewable energy into power systems in recent years, the increasing number of power electronic devices has posed new challenges to the stability and security of power systems. Power electronic devices contain a large number of high-frequency switching devices. The interaction between high-frequency power electronic circuits and control systems has made the dynamic characteristics of power systems increasingly complex, placing higher demands on power system simulation. Traditional simulation tools face problems such as reduced accuracy and limited modeling fidelity when simulating power systems with power converters. Traditional algorithms lack good numerical accuracy and stability, suffer from high computational complexity, and have difficulty capturing high-frequency harmonics. Summary of the Invention

[0003] Aiming at the technical problems existing in power system simulation, the present invention proposes a data-driven power system simulation optimization method which has reasonable design, simple method, strong theoretical basis and can achieve accuracy and consistency in power system simulation.

[0004] In order to achieve the above object, the technical solution adopted by the present invention is: a data-driven power system simulation optimization method, characterized by comprising the following steps:

[0005] S1. Build a data-driven, dynamic, adaptive space-time segmentation and coordination model to map fast-changing electromagnetic transients to quantum processors, retain slow-changing electromechanical transients on classical computers, encode grid node voltages as quantum bit amplitudes, and equate the power network dynamics to a quantum Hamiltonian. This decouples the quantum domain of fast-changing electromagnetic transients from the classical domain of slow-changing electromechanical dynamics, satisfying the following requirements:

[0006]

[0007] Where Δt q (t) is the quantum domain step size at time t, Δt c (t) is the classical domain step size at time t, N is the proportional coefficient, ε is the error tolerance limit, e(t) is the local truncation error, p is the implicit trapezoidal method order, the quantum domain boundary conditions are used to predict the switching transients through LSTM, and the final interface voltage is predicted by compensating the interpolation error through CNN;

[0008] S2. Construct quantum computing sparse code and dynamic error correction, non-uniform code distance distribution, divide the quantum processor into the critical zone CZ and the non-critical zone NCZ, implement non-uniform code distance distribution, and calculate the sparse logic error rate as follows:

[0009]

[0010] Among them, p phys is the physical fault tolerance, d CZ is the key area code distance, d NCZ is the code distance in the non-critical area, p th =1%, C CZ is the error correction capability coefficient of the critical area, C NCZ is the error correction capability coefficient of the non-critical area, C CZ <C NCZ , and then dynamically adjust the code distance through reinforcement learning DDPG strategy;

[0011] S3. Model implementation: In the power grid topology, the converter nodes are first marked as critical areas, and the others are non-critical areas. The initial code distance of the critical area is set to 5, and the initial code distance of the non-critical area is set to 3. LSTM, CNN, and DDPG strategies are trained offline. The computing tasks are allocated through dynamic adaptive spatiotemporal segmentation and collaborative models to realize a data-driven power system simulation optimization method.

[0012] Preferably, the input for predicting the switching transient by LSTM is the historical voltage and the switching state to obtain the switching transient waveform. The input for predicting the final interface voltage by compensating the interpolation error by CNN is the interpolation voltage, the quantum domain current and the voltage gradient to obtain the voltage correction. The final interface voltage is:

[0013] V final =V interp +ΔV

[0014] Among them, V final is the final interface voltage, V interp is the interpolation voltage, and ΔV is the voltage correction amount.

[0015] Preferably, the code distance is dynamically adjusted by reinforcement learning DDPG strategy, the state space is quantum fidelity and voltage change rate, the action space is a combination of critical area code distance and non-critical area code distance, and the reward function formula is:

[0016] R(t)=w1Accuracy-w2Latency-w3Qubitcost

[0017] Among them, w1, w2 and w3 are weight coefficients, Accuracy represents the inverse of the error between the simulation result and the voltage amplitude of the real physical system, Latency represents the time delay from the start of the simulation task to the output of the result, and Qubitcost represents the total number of physical quantum bits consumed during the simulation process.

[0018] Compared with the prior art, the advantages and positive effects of the present invention are:

[0019] The present invention designs a data-driven power system simulation optimization method, constructs a data-driven dynamic adaptive space-time segmentation and coordination model, maps fast-changing electromagnetic transient processes to a quantum processor, retains slow-changing electromechanical transient processes on a classical computer, encodes grid node voltages as quantum bit amplitudes, and dynamically equates the power network to a quantum Hamiltonian. This decouples the quantum domain of fast-changing electromagnetic transients from the classical domain of slow-changing electromechanical dynamics, achieving adaptive balanced simulation accuracy and efficiency.

[0020] Construct quantum computing sparse codes and dynamic error correction, non-uniform code distance distribution, reduce cross-scale data transmission errors, avoid high-frequency harmonic distortion, achieve high fault tolerance simulation under limited quantum resources, and dynamically adjust the code distance through reinforcement learning DDPG strategy to respond to noise fluctuations and changes in grid conditions in real time, optimizing resources and fault tolerance balance. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following is a brief introduction to the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0022] Figure 1 Schematic diagram of the structure of a data-driven power system simulation optimization method. DETAILED DESCRIPTION

[0023] In order to more clearly understand the above-mentioned objects, features and advantages of the present invention, the present invention is further described below in conjunction with the accompanying drawings and embodiments. It should be noted that, in the absence of conflict, the embodiments of the present application and the features therein can be combined with each other.

[0024] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways than those described herein. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0025] Examples, such as Figure 1 As shown, considering that power electronic equipment contains a large number of high-frequency switching devices, the interaction between high-frequency power electronic circuits and control systems makes the dynamic characteristics of the power system increasingly complex, and also puts higher requirements on power system simulation. Traditional simulation tools face problems such as reduced accuracy and limited modeling fidelity in the simulation scenario of power systems with power converters connected. Traditional algorithms do not have good numerical accuracy and stability, and face problems such as high computational complexity and difficulty in capturing high-frequency harmonics. The present invention proposes a data-driven power system simulation optimization method.

[0026] As power systems continue to evolve, their complexity and data volumes continue to increase. Traditional computing methods and technologies are no longer able to meet the power system's demands for efficient and accurate calculations and decision-making. Quantum computing, as a new computing model, has enormous potential for application in power systems. In power grid planning and optimization, quantum computing can efficiently solve complex optimization problems.

[0027] LSTM networks are a special type of recurrent neural network with the advantages of high prediction accuracy and strong recursive capabilities. They can solve the long-term dependency problem of recurrent neural networks. By adding a "gate" structure to add or discard information, they can provide long-term memory.

[0028] Convolutional Neural Networks (CNNs) have been widely used in fields such as image classification and object detection due to their excellent feature extraction capabilities and good parallelism. CNNs contain numerous repetitive multiplication-accumulation operations. Reconfigurable processors, with their regular parallel structure and flexible configuration capabilities, have demonstrated significant advantages in CNN computational research.

[0029] Considering the superposition and entanglement characteristics of quantum states, large-scale linear equations can be solved in parallel, reducing complexity. Quantum evolution follows unitary transformations, theoretically avoiding the energy consumption thermodynamic limit of classical computers. Electromagnetic transient (fast-changing) processes are mapped to quantum processors, while electromechanical transient (slow-changing) processes are retained on classical computers to achieve time-space separation, calculate boundary conditions, and achieve decoupling of time and space dimensions. Therefore, we first construct a data-driven dynamic adaptive time-space segmentation and coordination model, mapping fast-changing electromagnetic transient processes to quantum processors, while retaining slow-changing electromechanical transient processes on classical computers. Grid node voltages are encoded as quantum bit amplitudes, and the power network dynamics are equivalent to the quantum Hamiltonian. This decouples the quantum domain of fast-changing electromagnetic transients from the classical domain of slow-changing electromechanical dynamics, satisfying the following requirements:

[0030]

[0031] Where Δt q (t) is the quantum domain step size at time t, Δt c(t) is the classical domain step size at time t, N is the proportional coefficient, ε is the error tolerance limit, e(t) is the local truncation error, and p is the order of the implicit trapezoidal method, which realizes the relationship between the progress and efficiency of the adaptive balance simulation. The step size is dynamically adjusted according to the local error. The step size is automatically reduced (μs level) in the high-frequency switching stage of the converter (large error), and the step size is enlarged (ms level) in the low-frequency dynamic stage of the power grid (small error) to achieve balance accuracy and efficiency. Compared with the fixed step size method, the overall computational complexity is reduced, which is suitable for long-term simulation scenarios. The quantum domain boundary conditions predict the switching transients through LSTM, and the final interface voltage is predicted by compensating the interpolation error through CNN. Specifically, the input of the switch transient prediction through LSTM is the historical voltage and switch state to obtain the switch transient waveform. The input of the final interface voltage prediction through CNN compensating the interpolation error is the interpolation voltage, quantum domain current and voltage gradient to obtain the voltage correction. The final interface voltage is:

[0032] V final =V interp +ΔV

[0033] Among them, V final is the final interface voltage, V interp is the interpolated voltage, and αV is the voltage correction. The LSTM captures the timing characteristics of switching events, and the CNN corrects local nonlinear errors, reducing the relative error of the interface voltage at μs-level mutation points. The fusion of quantum domain current and classical domain gradient information eliminates the phase lag problem of traditional interpolation methods (such as cubic splines), improves the accuracy of harmonic analysis, and achieves cross-scale consistency.

[0034] Deep reinforcement learning (DRL) is a core AI technology. Through interactive learning between an intelligent agent and its environment, it provides an efficient approach for decision-making in various fields. To address the problem of continuous action spaces, the Deterministic Policy Gradient (DPG) algorithm was proposed. Combining the Deep Q-Network (DQN) and DPG concepts, the Deep Deterministic Policy Gradient (DDPG) algorithm was created.

[0035] Then, considering that traditional surface codes require quantum bits to be arranged into a dense two-dimensional grid, the resource overhead is huge. In view of the characteristics of power system simulation, the converter nodes (high-frequency switching) require high fidelity, while the transmission lines (low-frequency dynamics) can tolerate higher noise. The instantaneous error rate of the converter switch increases sharply, but is lower in steady state. Construct quantum computing sparse codes and dynamic error correction, non-uniform code distance distribution, divide the quantum processor into the critical zone CZ and the non-critical zone NCZ, implement non-uniform code distance distribution, and calculate the sparse logic error rate as follows:

[0036]

[0037] Among them, p phys is the physical fault tolerance, d CZ is the key area code distance, d NCZ is the code distance in the non-critical area, p th =1%, C CZ is the error correction capability coefficient of the critical area, C NCZ is the error correction capability coefficient of the non-critical area, C CZ <C NCA , and then dynamically adjust the code distance through reinforcement learning DDPG strategy. The non-uniform code distance design reduces the consumption of physical quantum bits and allows dynamic adjustment of the CZ / NCZ division to respond to changes in grid topology (such as the access of new energy sources) without the need to reconstruct the entire error correction code. Specifically, the code distance is dynamically adjusted through reinforcement learning DDPG strategy. The state space is quantum fidelity and voltage change rate, and the action space is a combination of critical area code distance and non-critical area code distance. The reward function formula is:

[0038] R(t)=w1Accuracy-w2Latency-w3Qubitcost

[0039] Where w1, w2, and w3 are weight coefficients. Accuracy represents the inverse of the voltage amplitude error between the simulation result and the actual physical system. Latency represents the time delay from simulation task initiation to result output. Qubitcost represents the total number of physical qubits consumed during the simulation. The weight coefficients seek the optimal balance between accuracy, speed, and resources to improve overall performance. Dynamic weight adjustment is supported, such as during fault conditions, and to adapt to different operating modes during steady state. By penalizing qubit cost, excessive resource usage is avoided on noisy, medium-sized quantum devices, improving task parallelization capabilities.

[0040] Finally, the model was implemented and run. In the power grid topology, the converter nodes were first marked as critical areas, and the others were non-critical areas. The initial code distance in the critical area was set to 5, and the initial code distance in the non-critical area was set to 3. LSTM, CNN, and DDPG strategies were trained offline. The computing tasks were allocated through dynamic adaptive spatiotemporal segmentation and collaborative models to achieve a data-driven power system simulation optimization method.

[0041] The above description is merely a preferred embodiment of the present invention and does not constitute any other form of limitation to the present invention. Any person skilled in the art may utilize the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes for application in other fields. However, any simple modification, equivalent change, and modification of the above embodiments made in accordance with the technical essence of the present invention without departing from the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A data-driven power system simulation optimization method, characterized in that: The steps include: S1. Build a data-driven, dynamic, adaptive space-time segmentation and coordination model to map fast-changing electromagnetic transients to quantum processors, retain slow-changing electromechanical transients on classical computers, encode grid node voltages as quantum bit amplitudes, and equate the power network dynamics to a quantum Hamiltonian. This decouples the quantum domain of fast-changing electromagnetic transients from the classical domain of slow-changing electromechanical dynamics, satisfying the following requirements: Where Δt q (t) is the quantum domain step size at time t, Δt c (t) is the classical domain step size at time t, N is the proportional coefficient, ε is the error tolerance limit, e(t) is the local truncation error, p is the implicit trapezoidal method order, the quantum domain boundary conditions are used to predict the switching transients through LSTM, and the final interface voltage is predicted by compensating the interpolation error through CNN; S2. Construct quantum computing sparse code and dynamic error correction, non-uniform code distance distribution, divide the quantum processor into the critical zone CZ and the non-critical zone NCZ, implement non-uniform code distance distribution, and calculate the sparse logic error rate as follows: Among them, p phys is the physical fault tolerance, d CZ is the key area code distance, d NCZ is the code distance in the non-critical area, p th =1%, C CZ is the error correction capability coefficient of the critical area, C NCZ is the error correction capability coefficient of the non-critical area, C CZ <C NCZ , and then dynamically adjust the code distance through reinforcement learning DDPG strategy; S3. Implement data-driven power system simulation optimization. In the power grid topology, first mark the converter node as the critical area and the others as non-critical areas. Set the initial critical area code distance to 5 and the initial non-critical area code distance to 3. Offline training of LSTM, CNN and DDPG strategies is carried out. Through dynamic adaptive spatiotemporal segmentation and collaborative model allocation of computing tasks, a data-driven power system simulation optimization method is implemented.

2. The data-driven power system simulation optimization method according to claim 1, characterized in that: The inputs of the LSTM prediction of the switching transient are the historical voltage and the switching state, and the switching transient waveform is obtained. The inputs of the CNN compensation of the interpolation error prediction of the final interface voltage are the interpolation voltage, the quantum domain current, and the voltage gradient, and the voltage correction is obtained. The final interface voltage is: V final =V interp +ΔV Among them, V final is the final interface voltage, V interp is the interpolation voltage, and ΔV is the voltage correction amount.

3. The data-driven power system simulation optimization method according to claim 1, characterized in that: The code distance is dynamically adjusted through reinforcement learning DDPG strategy. The state space is quantum fidelity and voltage change rate. The action space is the combination of critical area code distance and non-critical area code distance. The formula of the reward function is: R(t)=w1Accuracy-w2Latency-w3Qubitcost Among them, w1, w2 and w3 are weight coefficients, Accuracy represents the inverse of the error between the simulation result and the voltage amplitude of the real physical system, Latency represents the time delay from the start of the simulation task to the output of the result, and Qubitcost represents the total number of physical quantum bits consumed during the simulation process.

Images