Platform for testing dynamic characteristics and adjusting capability of network-forming type energy storage system

Abstract

The invention discloses a dynamic characteristic and regulation capability test platform for a network-building type energy storage system, and relates to the field of power system control, and the system comprises a hardware test platform building module, a software test platform building module, a multi-disturbance test scene building module, an acquisition and analysis module and a regulation capability evaluation module. By comprehensively evaluating the dynamic response, the steady-state error, the recovery capability and the anti-disturbance capability of the energy storage system and combining the particle swarm optimization and the deep learning optimization control strategy, the adjustment capability and the stability of the energy storage system in a complex power grid environment can be improved, the control strategy can be adjusted in real time to adapt to different load disturbances, and the stability of the energy storage system is improved. The overall performance and reliability of the energy storage system are improved, and stable operation of a power grid and efficient utilization of renewable energy sources are effectively supported.

Description

Technical Field

[0001] The present invention relates to the technical field of power system control, and particularly to a test platform for the dynamic characteristics and regulation capabilities of a grid-forming energy storage system. Background Art

[0002] The grid-forming energy storage system is an important part of modern power systems. Especially in the integrated application of renewable energy sources such as wind energy and solar energy, the energy storage system can smooth the load fluctuations of the power grid, regulate the grid frequency, and enhance the stability of the power grid. With the increasing requirements for energy consumption and environmental protection, the proportion of renewable energy in the power system is continuously increasing. The performance of the energy storage system is directly related to its stable role in the power system. The dynamic response ability and regulation ability of the energy storage system are key indicators to measure its performance. Most of the existing testing methods for energy storage systems focus on steady-state characteristics, while the research on the response and recovery capabilities of the system under dynamic loads, impact loads, and grid disturbances is relatively weak.

[0003] Currently, traditional power system control methods rely on simple mathematical models and control algorithms based on linear theory. However, with the increasing complexity and disturbances of the power grid, traditional control methods face great challenges. Therefore, new control strategies based on deep learning, machine learning, and intelligent optimization algorithms (such as particle swarm optimization) have become an important direction for improving the regulation ability and stability of power systems. These advanced control strategies can adaptively cope with complex load fluctuations, disturbances, and uncertainties, and improve the operation efficiency and stability of power systems. Summary of the Invention

[0004] In view of the problems such as insufficient dynamic performance testing of existing energy storage systems, insufficient control strategies and regulation capabilities, and insufficient adaptability to load disturbances, the present invention is proposed.

[0005] Therefore, the problems to be solved by the present invention are to comprehensively evaluate the dynamic response, steady-state error, recovery ability, and anti-disturbance ability of the system, and optimize its control strategy to enhance the stability of the power grid.

[0006] To solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, an embodiment of the present invention provides a test platform for the dynamic characteristics and regulation capabilities of a grid-forming energy storage system, which includes a hardware test platform construction module for simulating photovoltaic or wind energy input using a programmable DC power supply to provide a controllable power source, and simulating the main body of the grid-forming energy storage system based on a grid-forming energy storage converter and a battery pack; a software test platform construction module for establishing a virtual test environment according to a dynamic simulation platform and performing self-learning analysis on test data in combination with a machine learning framework; a multi-disturbance test scenario construction module for setting different load mutations, recording the power, frequency, voltage response time, and overshoot of the energy storage system, simulating random load fluctuations through a noise function, and using a high-power instantaneous impact load to observe the voltage / frequency recovery ability of the system; a collection and analysis module for collecting parameters such as voltage, frequency, power, phase, and total harmonic distortion, analyzing the system response characteristics using wavelet transform and Fourier transform, training a deep neural network model to predict the dynamic response ability of the system, and improving the system control strategy through a particle swarm algorithm; and a regulation ability evaluation module for calculating the steady-state error, then evaluating the rise time, overshoot, and recovery time of the system, and then testing the stability of the system under different load disturbances.

[0008] As a preferred solution of the test platform for the dynamic characteristics and regulation capabilities of the grid-forming energy storage system described in the present invention, wherein: simulating photovoltaic or wind energy input using a programmable DC power supply to provide a controllable power source includes:

[0009] Setting the I-V curve of the photovoltaic array to test the maximum power point tracking ability of the energy storage system, and the calculation formula of the I-V curve is as follows: ,

[0010] wherein, I sc is the short-circuit current, V oc is the open-circuit voltage, V t is the temperature-related parameter;

[0011] Setting the wind speed change curve and adopting high-frequency dynamic power regulation to simulate the wind speed-power curve, and the calculation formula of the wind speed-power curve is as follows: ,

[0012] wherein, C p is the power coefficient, ρ is the air density, A is the swept area of the wind turbine, and v is the wind speed.

[0013] As a preferred solution of the test platform for the dynamic characteristics and regulation capabilities of the grid-forming energy storage system described in the present invention, wherein: simulating the main body of the grid-forming energy storage system based on a grid-forming energy storage converter and a battery pack includes:

[0014] Disconnecting the external power grid, starting the grid-forming PCS, setting the target voltage and frequency, gradually increasing the load, observing whether the system can maintain a stable voltage and frequency, and recording the steady-state error;

[0015] Set the initial load to 50% of the rated power, instantaneously increase the load to 100%, record the rise time, instantaneously reduce the load to 10%, and observe whether overshoot or undervoltage occurs;

[0016] Simulate a grid voltage dip, observe whether the PCS can operate stably, introduce a sudden load, observe whether the system frequency overshoot is ≤2 Hz, inject a 2% frequency perturbation, and test the frequency recovery time.

[0017] As a preferred solution of the test platform for the dynamic characteristics and regulation ability of the grid-forming energy storage system described in the present invention, wherein: establish a virtual test environment according to the dynamic simulation platform, and perform self-learning analysis on the test data in combination with the machine learning framework, including:

[0018] Use LSTM to predict the voltage and frequency fluctuations of the system under different load conditions: ,

[0019] where y t is the system state predicted at the current time t, y t-n is the system state at the past n time steps, and n is the number of past time steps;

[0020] Use CNN to identify abnormal grid states, monitor the response data of the PCS under different load conditions, optimize the PI controller parameters through a neural network, adaptively adjust the battery charge and discharge strategy, and optimize the battery discharge timing in combination with the grid load prediction.

[0021] As a preferred solution of the test platform for the dynamic characteristics and regulation ability of the grid-forming energy storage system described in the present invention, wherein: simulate the random fluctuation of the load through a noise function, use a high-power instantaneous impact load, and observe the voltage / frequency recovery ability of the system, including:

[0022] Introduce noise disturbance on the load in the simulation environment, observe the steady-state error and short-term oscillation of the voltage and frequency under fluctuations, and use the random pulse of Gaussian white noise for disturbance. The expression of the Gaussian white noise is as follows: ,

[0023] where P load (t) is the load power at time t, P base is the reference load power, σ is the disturbance amplitude, and N(0,1) is the random noise of the standard normal distribution;

[0024] Apply a high-power impact on the load end in a short time. For the step load, instantaneously increase it by 50% of the rated load. For the pulse load, apply a high-power load for 2 - 5 s and then remove it. For the periodic impact load, apply a 150% rated power load for 1 s every 10 s, calculate the recovery time and overshoot, and observe whether the responses are consistent under multiple impacts.

[0025] As a preferred solution of the test platform for the dynamic characteristics and regulation ability of the network-forming energy storage system described in the present invention, it includes: collecting voltage, frequency, power, phase and total harmonic distortion parameters, analyzing the system response characteristics by using wavelet transform and Fourier transform, training a deep neural network model, and predicting the dynamic response ability of the system, including:

[0026] Analyze the harmonic characteristics of voltage and frequency by using fast Fourier transform, analyze the harmonic content, calculate the total harmonic distortion, and extract transient oscillation, low-frequency fluctuation and mutation characteristics through continuous wavelet transform;

[0027] Adopt deep feedforward neural network training. The input layer includes the original signal, FFT and wavelet features. The output layer regresses and outputs the predicted recovery time, and classifies and outputs stable / oscillatory.

[0028] As a preferred solution of the test platform for the dynamic characteristics and regulation ability of the network-forming energy storage system described in the present invention, it includes: improving the system control strategy by using particle swarm optimization algorithm, including:

[0029] Construct a suitable objective function J, and the objective function is set as follows: ,

[0030] where V ref , f ref are the target voltage and frequency, P loss is the power loss of the system, THD is the total harmonic distortion, and α1, α2, α3, α4 are weight coefficients;

[0031] Set the number of particles N, randomly initialize the control parameters of each particle and give the velocity, set the maximum number of iterations T of the particle, the inertia weight w and the learning factors c1, c2. For each particle X i , calculate its fitness. If J i is less than the current optimal value, then update the individual optimal solution P best,i and the global optimal solution G best .

[0032] As a preferred solution of the test platform for the dynamic characteristics and regulation ability of the network-forming energy storage system described in the present invention, it includes: calculating the steady-state error, and then testing the stability of the system under different load disturbances, including:

[0033] After the system response is stable, record the difference between the target output y ref and the actual output y act . Calculate the steady-state error e ss to evaluate the accuracy of the system, and adjust the gain and control strategy parameters;

[0034] Apply a load disturbance, which causes fluctuations in voltage, current, and frequency parameters by changing the load power ΔP(t); observe the system response, record the voltage and frequency parameters of the system after the disturbance, and calculate the response time, rise time, recovery time, and steady-state error; analyze the system's recovery ability. For a large-power instantaneous impact load, evaluate whether the system can restore the voltage and frequency to normal levels within a short time.

[0035] In a second aspect, an embodiment of the present invention provides a computer device, including a memory and a processor, where the memory stores a computer program, and: when the computer program instructions are executed by the processor, the steps of the test platform for the dynamic characteristics and regulation ability of the network-forming energy storage system as described in the first aspect of the present invention are implemented.

[0036] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, and: when the computer program instructions are executed by the processor, the steps of the test platform for the dynamic characteristics and regulation ability of the network-forming energy storage system as described in the first aspect of the present invention are implemented.

[0037] The beneficial effects of the present invention are as follows: By comprehensively evaluating the dynamic response, steady-state error, recovery ability, and anti-disturbance ability of the energy storage system, and combining the particle swarm algorithm and deep learning optimization control strategy, the present invention can not only improve the regulation ability and stability of the energy storage system in a complex power grid environment, but also adjust the control strategy in real time to adapt to different load disturbances, enhance the overall performance and reliability of the energy storage system, and effectively support the stable operation of the power grid and the efficient utilization of renewable energy. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0039] Figure 1 It is a schematic structural diagram of the test platform for the dynamic characteristics and regulation ability of the network-forming energy storage system.

[0040] Figure 2 It is a diagram of the computer device of the test platform for the dynamic characteristics and regulation ability of the network-forming energy storage system. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0041] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following will provide a detailed description of the specific embodiments of the present invention with reference to the accompanying drawings of the specification.

[0042] In the following description, numerous specific details are set forth to provide a thorough understanding of the present invention. However, the present invention may be practiced in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the spirit of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0043] Secondly, the so-called "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation manner of the present invention. The appearances of "in one embodiment" in different places in this specification do not all refer to the same embodiment, nor are they separate or alternative embodiments that exclude each other with other embodiments.

[0044] Embodiment 1

[0045] Referring to Figures 1 - 2 , which is the first embodiment of the present invention. This embodiment provides a test platform for the dynamic characteristics and regulation capabilities of a network-forming energy storage system, including

[0046] S1: Use a programmable DC power supply to simulate photovoltaic or wind energy input, provide a controllable power source, and simulate the main body of the network-forming energy storage system based on the network-forming energy storage converter and the battery pack.

[0047] Preferably, set the I-V curve of the photovoltaic array to test the maximum power point tracking ability of the energy storage system. The calculation formula of the I-V curve is as follows: ,

[0048] where I sc is the short-circuit current, V oc is the open-circuit voltage, and V t is the temperature-related parameter;

[0049] Set the wind speed change curve, adopt high-frequency dynamic power regulation to simulate the wind speed-power curve. The calculation formula of the wind speed-power curve is as follows: ,

[0050] where C p is the power coefficient, ρ is the air density, A is the swept area of the wind turbine, and v is the wind speed.

[0051] Preferably, disconnect the external power grid, start the PCS, set the target voltage and frequency, gradually increase the load, observe whether the system can maintain stable voltage and frequency, and record the steady-state error;

[0052] Set the initial load to 50% of the rated power, instantaneously increase the load to 100%, record the rise time, instantaneously reduce the load to 10%, and observe whether there is overshoot or undervoltage;

[0053] Simulate the grid voltage dip, observe whether the PCS can operate stably, introduce a sudden load, observe whether the system frequency overshoots ≤ 2 Hz, inject a 2% frequency perturbation, and test the frequency recovery time.

[0054] Further, during the test, first set the light intensity and temperature to obtain the I-V curve of the photovoltaic array. According to the I-V curve of the photovoltaic array, calculate the output power P = V × IP at different voltages and determine the corresponding maximum power point. Use the MPPT algorithm or incremental conductance method to adjust the working voltage of the photovoltaic array in real time so that the output power is as close as possible to the maximum power point. Under different environmental conditions (such as changes in light intensity and temperature), test whether the energy storage system can accurately track the maximum power point of the photovoltaic array and operate stably.

[0055] According to the changes in different wind speeds, set a series of wind speed values and calculate the corresponding power output.

[0056] For low wind speeds (below 3 m / s), the power output is very low. In the medium wind speed range (about 5 - 12 m / s), the power output reaches a relatively high value. At high wind speeds (above 15 m / s), the fan output reaches the maximum power and may be limited by the designed maximum power output of the fan. Through the set dynamic wind speed change curve, simulate the rapid change of wind speed and calculate the power output at each moment. The energy storage system performs high-frequency dynamic regulation according to the relationship between the real-time wind speed and power to ensure that the power output is close to the maximum power point and provides real-time regulation feedback. Dynamically adjust the load demand of the energy storage system according to the wind speed change to ensure matching with the output power of the wind turbine and achieve maximum power. When using photovoltaic and wind energy simultaneously, the energy storage system needs to reasonably allocate the load according to the different power curves of the two to ensure the stable power output of the overall system.

[0057] Through an intelligent scheduling strategy, maximize the power output of photovoltaic and wind energy to meet the grid demand. When the power output of photovoltaic or wind energy generation exceeds the demand, the energy storage system stores the excess electrical energy. When the power demand is greater than the real-time available photovoltaic and wind energy, the energy storage system releases electrical energy to ensure stable power supply. Test the power output of the system under changes in light and wind speed, and evaluate the maximum power tracking ability of the photovoltaic array and the fan.

[0058] Test the response speed of the energy storage system under dynamic loads, especially the power regulation ability of the system when the wind speed and light intensity change rapidly. Conduct long-term light and wind speed change tests to evaluate the operating stability of the energy storage system under different seasons and weather conditions.

[0059] S2: Establish a virtual test environment based on the dynamic simulation platform and perform self-learning analysis on the test data in combination with the machine learning framework.

[0060] Preferably, use the LSTM prediction system to predict voltage and frequency fluctuations under different loads: ,

[0061] where y t is the system state predicted at the current time t, and y t-n is the system state at the past n time steps, and n is the number of past time steps;

[0062] Adopt CNN to identify abnormal states of the power grid, monitor the response data of PCS under different load conditions, optimize the parameters of the PI controller through a neural network, adaptively adjust the battery charge and discharge strategy, and optimize the battery discharge timing in combination with the power grid load prediction.

[0063] Furthermore, collect the real-time response data of PCS, including various electrical parameters such as voltage, current, and power. Perform time-domain and frequency-domain analysis on the data, extract feature maps, and form the input suitable for CNN. The input data is the time-series data of electrical signals such as voltage and current. Construct a CNN network including convolutional layers, pooling layers, and fully connected layers to automatically extract spatial and temporal features in the data. Set appropriate hyperparameters such as the size of the convolutional kernel, pooling layer parameters, and learning rate. Use the labeled dataset to train the CNN model and optimize it with the cross-entropy loss function. Verify the abnormal recognition ability of the CNN model on the test set, and evaluate the accuracy, recall rate, and F1 score of the model. Deploy the trained CNN model to monitor the output response data of PCS in real time and identify abnormal states in the power grid. According to the abnormal recognition results, trigger an alarm or adjust system parameters for adaptive adjustment.

[0064] Use a neural network to optimize the proportional and integral parameters of the PI controller based on the power grid load prediction data and the charge and discharge state of the battery. The inputs of the neural network are information such as the current state of the battery, load demand, and power grid state, and the output is the optimal PI parameters. Use reinforcement learning or supervised learning to train the neural network, and adjust the parameters of the PI controller through experimental data so that the battery charge and discharge strategy can cope with load fluctuations and power grid abnormalities. During operation, the neural network adjusts the parameters of the PI controller through real-time feedback to ensure a stable and efficient battery charge and discharge process and cope with load mutations or power grid abnormalities.

[0065] Use the LSTM model to predict the changing trend of the power grid load and identify the peak and trough periods of the power grid load in advance. Identify the abnormal state of the power grid through the CNN model and adjust the discharge timing in combination with real-time feedback. When the load demand is high, the battery discharges in advance to meet the power grid demand; when the power grid is abnormal, the battery adjusts the discharge mode through rapid response to avoid imposing an additional burden on the power grid. Adopt a dynamic adjustment strategy, combine a neural network model to optimize the discharge timing, so as to reduce power grid fluctuations and excessive charge and discharge of the battery. Feed the optimized discharge timing back to the energy storage control system to adjust the discharge process of the battery in real time.

[0066] S3: Set different load mutations, record the power, frequency, voltage response time and overshoot of the energy storage system, simulate the random fluctuation of the load through the noise function, and use a high-power instantaneous impact load to observe the voltage / frequency recovery ability of the system.

[0067] Preferably, on the simulation environment load, introduce noise disturbance, observe the steady-state error and short-term oscillation of voltage and frequency under fluctuations, and use the random pulse of Gaussian white noise for disturbance. The expression of the Gaussian white noise is as follows: ,

[0068] where P load (t) is the load power at time t, P base is the reference load power, σ is the disturbance amplitude, and N(0,1) is the random noise of the standard normal distribution;

[0069] Apply a high-power impact at the load end for a short time. The step load instantaneously increases by 50% of the rated load, the pulse load applies a high-power load for 2 - 5 s and then is removed, and the periodic impact load applies a 150% rated power load for 1 s every 10 s. Calculate the recovery time and overshoot, and observe whether the responses are consistent under multiple impacts.

[0070] Furthermore, the power system used in this embodiment is a microgrid composed of photovoltaic power generation, wind power generation and an energy storage system. The system parameters are as follows: the maximum power of the photovoltaic array is 10 kW, the rated power of the fan is 5 kW,

[0071] the capacity of the energy storage system is 20 kWh, the reference load power is 8 kW, the power grid frequency is 50 Hz, the power grid voltage is 230 V, the load disturbance amplitude is 0.1. After each disturbance, the time required for the system to stabilize to the steady state is 5 s.

[0072] At the load end, instantaneously apply a sudden impact of 50% of the rated load for 2 - 5 s, and then return to the reference load. Assume P base= 8 kW, the load is 150% of the rated load within 0 - 5 seconds and then returns to the reference value. A high-power load impact is applied every 10 seconds for 1 second, with the load increasing by 50% of the rated power. This means that every 10 seconds, the load increases by 50% of the rated power and lasts for 1 second, after which it returns to the reference load. The initial load power is 8 kW. A 1-second load impact is applied every 10 seconds, with the load increasing to 12 kW and then returning to the reference load of 8 kW. After each pulse load is applied, the changes in system voltage, current, and frequency are recorded. Within 10 seconds after the load disturbance occurs, observe the changes in system voltage and frequency, paying particular attention to the system's recovery time and steady-state error.

[0073] Through simulation observations, the following results are obtained: When the load disturbance occurs, there are certain fluctuations in the system voltage and frequency. The disturbance duration is 1 second, the maximum fluctuation of the system voltage is about 5%, and the frequency fluctuation range is ±1 Hz. Under periodic load impacts, the fluctuations of voltage and frequency are more obvious, but the voltage and frequency can return to the steady state within 2 seconds. After the disturbance, the system voltage finally drops to 228 V, with an error of 0.87% compared to the reference voltage of 230 V. The system frequency finally stabilizes at 50.3 Hz, with an error of 0.6% compared to the reference frequency of 50 Hz.

[0074] After each load impact is applied, the average recovery time of the system voltage and frequency is 3 seconds. During the recovery process, the voltage fluctuation gradually decreases, and the frequency tends to the reference value. After each load impact is applied, the overshoot of the system voltage is about 6%, and the overshoot of the frequency is 1.5 Hz.

[0075] S4: Acquisition and analysis module, used to acquire voltage, frequency, power, phase, and total harmonic distortion parameters,

[0076] Analyze the system response characteristics using wavelet transform and Fourier transform, train a deep neural network model, predict the dynamic response ability of the system, and improve the system control strategy through the particle swarm algorithm.

[0077] Preferably, construct a suitable objective function J, and the objective function is set as follows: ,

[0078] where V ref , f ref are the target voltage and frequency, P loss is the power loss of the system, THD is the total harmonic distortion, and α1, α2, α3, α4 are weight coefficients;

[0079] Set the number of particles N, randomly initialize the control parameters of each particle and assign a velocity, set the maximum number of iterations T of the particle, the inertia weight w, and the learning factors c1, c2. For each particle X i , calculate its fitness. If Ji If it is less than the current optimal value, update the individual optimal solution P best,i and the global optimal solution G best .

[0080] Furthermore, in the particle swarm optimization algorithm, each particle represents a possible solution of the PI controller. The position and velocity of the particle represent the values of the gains Kp and Ki of the PI controller respectively. The specific algorithm parameter settings are as follows: The number of particles N is set to 50, that is, there are 50 particles in the particle swarm. The maximum number of iterations is set to 100 times. The inertia weight w is set to 0.7, indicating the forward inertia of the particle. The learning factors c1 = c2 = 1.5 represent the learning ability of the particle to its individual optimal solution and the global optimal solution during the search process. The position of the particle, that is, the controller gains Kp and Ki, represents the objective to be optimized. The update of the particle velocity represents the adjustment speed of the controller gain. The initial position of the particle (i.e., the initial gain of the PI controller) and velocity are randomly initialized. For each particle, the initial values of Kp and Ki are randomly selected, and the initial velocity is set. The selection range of the initial particle position and velocity can be set according to the requirements of the actual power system. For example, assume that the initial ranges of the gains Kp and Ki of the PI controller are [0, 10].

[0081] After calculating the fitness of each particle, if the fitness Ji of the current particle is less than the fitness of its individual optimal solution, update the individual optimal solution. If the fitness Ji of the current particle is less than the fitness of the global optimal solution, update the global optimal solution. Randomly initialize the position and velocity of the particle. For each particle, calculate the fitness Ji. Adjust the state of the particle according to the velocity update formula and the position update formula. Update the individual optimal solution and the global optimal solution according to the fitness. If the convergence condition is met, stop the iteration. After 100 iterations, the particle swarm optimization algorithm will converge to the optimal PI controller gain values of Kp and Ki. The following optimization results can be obtained: The controller gains Kp and Ki optimized by the particle swarm optimization algorithm will significantly reduce the fluctuations of the system voltage and frequency, making the system more stable. Through optimization, the power loss and total harmonic distortion are significantly reduced, thereby improving the operating efficiency of the power system.

[0082] S5: Calculate the steady-state error, then evaluate the rise time, overshoot, and recovery time of the system, and then test the stability of the system under different load disturbances.

[0083] Preferably, after the system response is stable, record the difference between the target output y ref and the actual output y act , calculate the steady-state error e ss to evaluate the accuracy of the system, and adjust the gain and control strategy parameters;

[0084] Apply a load disturbance, by changing the load power ΔP(t), to cause fluctuations in voltage, current, and frequency parameters; observe the system response, record the voltage and frequency parameters of the system after the disturbance, and calculate the response time, rise time, recovery time, and steady-state error; analyze the system's recovery ability. For a large-power instantaneous impact load, evaluate whether the system can restore the voltage and frequency to the normal level within a short time.

[0085] Further, assume that after the system stabilizes, the target voltage is V ref = 220V, and the target frequency is f ref = 50Hz. The actual voltage and frequency are V act = 221V and f act = 49.8Hz; for voltage and frequency, calculate respectively: e V = ∣V ref - V act ∣ = ∣220V - 221V∣ = 1V, e f = ∣f ref - f act ∣ = ∣50Hz - 49.8Hz∣ = 0.2Hz; the steady-state error at this time is e V = 1V and e f = 0.2Hz, and evaluate the system accuracy.

[0086] Assume that the load power increases by 50% of the rated power at time t = 5s: P old = 1000W,, P new = 1.5 × P old = 1500W

[0087] This causes fluctuations in voltage and frequency, and the amplitude of the load disturbance is ΔP(t) = 500W;

[0088] After applying the load disturbance, the system's response may exceed the target voltage or frequency, resulting in an overshoot. The overshoot Mp is defined as the ratio of the deviation when the system reaches the maximum value to the set value. Assume that the overshoot of the system voltage is Mp = 1.36%. Similarly, the overshoot for frequency is calculated as Mp = 2%.

[0089] For a large-power instantaneous impact load (e.g., the load power increases instantaneously by 50%), record the system's response and analyze whether the voltage and frequency can be restored to the normal level within a short time. Assume that the load mutation is applied at time t = 10s, the voltage and frequency fluctuate briefly, and the recovery time of the system is t recover = 0.8s.

[0090] To verify the stability of the system, multiple load impact tests are conducted. For example, a 150% rated power load is applied for 1 second every 10 seconds. After each load application, the changes in voltage and frequency are recorded, and the recovery time, overshoot, and steady-state error after each perturbation are calculated. For example, after 3 load impacts, the recovery times of the system are t recover1 = 0.9, t recover2 = 1.0 s, and t recover3 = 0.95 s, and the overshoots are Mp1 = 1.5%, Mp2 = 1.3%, and Mp3 = 1.4%. If the system recovery time is too long or the overshoot is too large, the system performance is optimized by adjusting the PI controller gain. Assuming the current gains are Kp = 1.5 and Ki = 0.5, the system response speed is increased by increasing the proportional gain Kp, or the steady-state error is reduced by increasing the integral gain Ki. For example, assuming the adjusted gains are Kp = 2.0 and Ki = 0.6, after a new load impact test, the system recovery time is shortened to t recover = 0.7 s, and the overshoot is reduced to Mp = 1.2%.

[0091] This embodiment also provides a computer device, which is applicable to the case of the dynamic characteristic and regulation ability test platform of the network-forming energy storage system, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the dynamic characteristic and regulation ability test platform of the network-forming energy storage system as proposed in the above embodiment.

[0092] This computer device can be a terminal. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected through a system bus. Among them, the processor of this computer device is used to provide computing and control capabilities. The memory of this computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of this computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a carrier network, NFC (Near Field Communication), or other technologies. The display screen of this computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of this computer device can be a touch layer covered on the display screen, or a button, a trackball, or a touchpad set on the shell of the computer device, or an external keyboard, touchpad, or mouse, etc.

[0093] This embodiment also provides a storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the dynamic characteristic and regulation ability test platform of the network-forming energy storage system as proposed in the above embodiment.

[0094] In summary, by comprehensively evaluating the dynamic response, steady-state error, recovery ability, and anti-disturbance ability of the energy storage system, and combining the particle swarm optimization algorithm and deep learning to optimize the control strategy, the present invention can not only improve the regulation ability and stability of the energy storage system in a complex power grid environment, but also adjust the control strategy in real time to adapt to different load disturbances, enhance the overall performance and reliability of the energy storage system, and effectively support the stable operation of the power grid and the efficient utilization of renewable energy.

[0095] Embodiment 2

[0096] This embodiment is the second embodiment of the present invention. This embodiment provides a test platform for the dynamic characteristics and regulation ability of a grid-forming energy storage system. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through simulation experiments.

[0097] Specifically, the target voltage is 220V, the target frequency is 50Hz, initially set as Kp = 1.5, Ki = 0.5, the initial load power is 1000W, and the load increases by 50% after perturbation, that is, the power change is 1500W. Random noise perturbation (Gaussian white noise) and high-power impact load are adopted, set for 20s, and load perturbation is carried out every 10s, with a duration of 1s.

[0098] Apply a perturbation of 50% of the rated power, and the load power increases from 1000W to 1500W. Among them, ΔP = 0.5 (that is, the load power increases by 50%), the simulation start time is t = 5s, and it lasts for 1s. The system responds to the load perturbation during the simulation process and reaches a new steady state. Record the differences between the target voltage and the actual voltage, and between the target frequency and the actual frequency. Evaluate the accuracy of the controller by calculating the steady-state error. Assume that the errors of voltage and frequency in the simulation are: voltage steady-state error e ss,V = 1V; frequency steady-state error e ss,f = 0.2Hz; the time from the application of the load perturbation to the start of the system's reaction. The simulation results show that the response time of the system is t r = 0.2s, the time required from the start of the system's fluctuation to reaching the steady-state voltage and frequency. The simulation results show that the rise time is t rise = 0.5s. The time required for the system to return to the set voltage and frequency after the load perturbation is removed. The simulation results show that the recovery time is t recover = 1s.

[0099] Furthermore, by observing the system response waveform, the overshoots of the system voltage and frequency are calculated. The overshoot Mp is the ratio between the maximum deviation and the target value. Assume that in the simulation, the voltage overshoot of the system is Mp,V = 1.36% and the frequency overshoot is Mp,f = 2%. To verify the stability and consistency of the system, multiple load disturbance tests are conducted. A 150% rated load is applied every 10 s for 1 s. After each load disturbance, the changes in voltage and frequency are recorded, and the recovery time and overshoot of the system are calculated.

[0100] The simulation results are as follows:

[0101] The recovery time t of the first load disturbance recover1 = 1.0 s, and the voltage overshoot Mp = 1.4%;

[0102] The recovery time t of the second load disturbance recover2 = 0.9, and the voltage overshoot Mp = 1.3%;

[0103] The recovery time t of the third load disturbance recover3 = 1.1 s, and the voltage overshoot Mp = 1.5%;

[0104] According to the simulation results, the gains of the PI controller are adjusted to optimize the system response. The proportional gain Kp = 2.0 and the integral gain Ki = 0.6 are increased, and another simulation is conducted to observe the improvement of the system response.

[0105] The simulation results after adjustment: The recovery time is reduced to t recover = 0.8 s, and the voltage overshoot is reduced to Mp = 1.2%.

[0106] Through the simulation experiment, the effect of optimizing the load disturbance response of the present invention through the PI controller is verified. After adjusting the parameters of the PI controller, the recovery time and overshoot of the system are significantly optimized, and in the case of multiple load disturbances, the system can recover stably and quickly.

[0107] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

Claims

1. A test platform for the dynamic characteristics and regulation capabilities of a network-forming energy storage system, characterized in that, Including: A hardware test platform building module, which is used to simulate photovoltaic or wind energy input by using a programmable DC power supply to provide a controllable power source, and simulate the main body of a grid-forming energy storage system based on a grid-forming energy storage converter and a battery pack; A software test platform building module, which is used to establish a virtual test environment according to a dynamic simulation platform and perform self-learning analysis on test data in combination with a machine learning framework; A multi-disturbance test scenario building module, which is used to set different load mutations and grid disturbances for simulation, record the power, frequency, voltage response time and overshoot of the energy storage system, simulate random load fluctuations through a noise function, and use a high-power instantaneous impact load to observe the voltage / frequency recovery ability of the system; A collection and analysis module, which is used to collect parameters such as voltage, frequency, power, phase and total harmonic distortion; Analyze the system response characteristics by using wavelet transform and Fourier transform, train a deep neural network model, predict the dynamic response ability of the system, and improve the system control strategy through a particle swarm algorithm; A regulation ability evaluation module, which is used to calculate the steady-state error, then evaluate the rise time, overshoot and recovery time of the system, and then test the stability of the system under different load disturbances.

2. The test platform for the dynamic characteristics and regulation capabilities of the network-forming energy storage system according to claim 1, characterized in that, The use of a programmable DC power supply to simulate photovoltaic or wind energy input and provide a controllable power source includes: Set the I-V curve of the photovoltaic array and test the maximum power point tracking ability of the energy storage system. The calculation formula of the I-V curve is as follows: , Among them, I sc is the short-circuit current, V oc is the open-circuit voltage, V t is the temperature-related parameter; Set the wind speed change curve, adopt high-frequency dynamic power regulation, and simulate the wind speed-power curve. The calculation formula of the wind speed-power curve is as follows: , Among them, C p is the power coefficient, ρ is the air density, A is the swept area of the wind turbine, and v is the wind speed.

3. The test platform for the dynamic characteristics and regulation capabilities of the network-forming energy storage system according to claim 1, characterized in that: The simulation of the main body of a grid-forming energy storage system based on a grid-forming energy storage converter and a battery pack includes: Disconnect the external power grid, start the PCS, set the target voltage and frequency, gradually increase the load, observe whether the system can maintain stable voltage and frequency, and record the steady-state error; Set the initial load to 50% of the rated power, instantaneously increase the load to 100%, record the rise time, instantaneously reduce the load to 10%, and observe whether there is overshoot or undervoltage; Simulate a grid voltage dip or rise, observe whether the PCS can operate in parallel - participate in regulation - operate in a steady state; introduce a sudden load, observe whether the system frequency overshoot is ≤2Hz, inject a 2% frequency disturbance, and test the frequency recovery time.

4. The test platform for the dynamic characteristics and regulation capabilities of the network-forming energy storage system according to claim 1, wherein: The establishment of a virtual test environment according to a dynamic simulation platform and the self-learning analysis of test data in combination with a machine learning framework include: Use the LSTM prediction system to predict voltage and frequency fluctuations under different load conditions: , where y t is the predicted system state at the current time t, and y t-n is the system state at the past n time steps, where n is the number of past time steps; Use a CNN to identify abnormal grid states, monitor the response data of the PCS under different load conditions, optimize the parameters of the PI controller through a neural network, adaptively adjust the battery charge and discharge strategy, and optimize the battery discharge timing in combination with grid load forecasting.

5. The test platform for the dynamic characteristics and regulation capabilities of the network-forming energy storage system according to claim 1, characterized in that, The simulation of random load fluctuations through a noise function and the use of a high-power instantaneous impact load to observe the voltage / frequency recovery ability of the system include: On the simulation environment load, introduce noise disturbances and observe the steady-state error and short-term oscillations of voltage and frequency under fluctuations. Use the random pulses of Gaussian white noise for disturbance. The expression of the Gaussian white noise is as follows: , where, P load (t) is the load power at time t, P base is the reference load power, σ is the disturbance amplitude, and N(0,1) is the random noise of the standard normal distribution; Apply a high-power impact at the load end for a short time. The step load instantaneously increases by 50% of the rated load, the pulse load applies a high-power load for 2 - 5s and then is removed, and the periodic impact load applies a 150% rated power load for 1s every 10s. Calculate the recovery time and overshoot, and observe whether the responses are consistent under multiple impacts.

6. The test platform for the dynamic characteristics and regulation ability of the network-forming energy storage system according to claim 1, characterized in that: The collection of parameters such as voltage, frequency, power, phase and total harmonic distortion, the analysis of system response characteristics by using wavelet transform and Fourier transform, the training of a deep neural network model, and the prediction of the dynamic response ability of the system include: Analyze the harmonic characteristics of voltage and frequency using the fast Fourier transform, analyze the harmonic content, calculate the total harmonic distortion, and extract transient oscillation, low-frequency fluctuation, and mutation characteristics through continuous wavelet transform; Adopt deep feedforward neural network training. The input layer includes the original signal, FFT, and wavelet features. The output layer regresses to output the predicted recovery time and classifies to output stable / oscillating.

7. The test platform for the dynamic characteristics and regulation capabilities of the network-forming energy storage system according to claim 1, characterized in that: Improve the system control strategy through the particle swarm algorithm, including: Construct an appropriate objective function J, and the objective function is set as follows: , Among them, V ref , f ref are the target voltage and frequency, P loss is the power loss of the system, THD is the total harmonic distortion, and α1, α2, α3, α4 are the weight coefficients; Set the number of particles N, randomly initialize the control parameters of each particle and assign a velocity, set the maximum number of iterations T of the particle, the inertia weight w, and the learning factors c1 and c2. For each particle X i , calculate its fitness. If J i is less than the current optimal value, then update the individual optimal solution P best,i and the global optimal solution G best .

8. The test platform for the dynamic characteristics and regulation ability of the network-forming energy storage system according to claim 1, wherein, Calculate the steady-state error and then test the stability of the system under different load disturbances, including: After the system response stabilizes, record the target output y ref and the actual output y act The difference between them is used to calculate the steady-state error e ss , evaluate the accuracy of the system, and adjust the gain and control strategy parameters; Apply a load disturbance, cause fluctuations in voltage, current, and frequency parameters by changing the load power ΔP(t); observe the system response, record the voltage and frequency parameters of the system after the disturbance, calculate the response time, rise time, recovery time, and steady-state error; analyze the system recovery ability. For a large-power instantaneous impact load, evaluate whether the system can restore the voltage and frequency to the normal level within a short time.

9. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the test platform for the dynamic characteristics and regulation ability of the network-forming energy storage system according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the test platform for the dynamic characteristics and regulation ability of the network-forming energy storage system according to any one of claims 1 to 8.

Images