Distributed electric energy storage system combined power supply control method based on user cluster load
By acquiring grid connection point parameters to construct a dynamic virtual impedance mapping model and phase compensation offset, the response lag and oscillation problems of distributed energy storage systems under nonlinear loads are solved, asynchronous power compensation is realized, and the system's adaptive adjustment capability and hardware lifespan are improved.
Patent Information
- Application Number
- CN202610336078.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-03-19
- Publication Date
- 2026-04-21
AI Technical Summary
Existing distributed energy storage systems cannot accurately and adaptively adjust the output power of storage units when facing nonlinear loads and impulsive loads, resulting in power overcompensation, response lag and voltage oscillation in the electrical topology area. Hardware devices are subjected to high-frequency pulsating current impacts, shortening their service life.
By acquiring the local voltage and current parameters of the storage unit's grid connection point, calculating the voltage support strength characterization quantity and sensitivity factor, constructing a dynamic virtual impedance mapping model, and combining the phase compensation offset, adjusting the output power slope and inverter pulse width modulation wave, asynchronous power compensation response is achieved.
It enables storage units to autonomously perceive their topological location without external communication support, eliminates cluster oscillations, enhances the system's adaptive adjustment capability under complex time-varying topologies, and extends the service life of energy storage media and power electronic devices.
Smart Images

Figure CN121906624A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to a joint power supply control method for a distributed energy storage system based on user cluster load, belonging to the technical field of energy storage systems. Background Technology
[0002] In the current distributed energy storage system's joint power supply control, droop control is used to adjust the output power of the storage unit. This method is based on the physical assumption that the impedance characteristics of the distribution network are inductive and constant, and power distribution is achieved by adjusting the voltage and frequency offset.
[0003] The increase of nonlinear and impulsive loads at the end of the distribution network causes the distribution lines to exhibit resistive characteristics and high time-varying behavior. The start-up and shutdown of high-power loads cause instantaneous fluctuations in the electrical stiffness of the local bus. Existing strategies cannot capture the evolution characteristics of bus stiffness and rely on fixed adjustment slopes, causing power overcompensation in the core area of the electrical topology storage units. The superposition of response lag and electrical deviation triggers cluster voltage oscillations, subjecting the converter power devices and storage media to high-frequency pulsating current impacts and shortening the hardware service life. Hardware layout limitations have extended to the level of refined coordination of control strategies. For example, the utility model patent with authorization announcement number CN220022338U discloses a method to improve the lifespan of energy storage batteries. Distributed energy storage systems use current sampling to drive relays. When lithium battery packs are in operation, they switch to nickel-metal hydride battery packs to participate in grid dispatch. By utilizing the physical characteristics of different battery media, the impact of charging and discharging switching on lifespan is avoided. This type of hardware-based hard switching mechanism is an islanded static protection system. It cannot perceive the dynamic stiffness characteristics of the local bus and is difficult to solve the physical resonance caused by the homogeneity of control logic when deploying large-scale clusters. When faced with transient impacts from nonlinear loads, the lack of adaptive mapping of electrical topology positions and phase decoupling capabilities leads to power response sequence misalignment between nodes, making it impossible to support the load. Instead, the relay operation and current surges accelerate the fatigue wear of power electronic devices.
[0004] Therefore, how to enable storage units to achieve accurate adaptive mapping of dynamic bus stiffness while maintaining local awareness and without external communication support, and to construct a physically differentiated asynchronous response mechanism to suppress cluster oscillations, has become the technical problem to be solved by this invention. Summary of the Invention
[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: A joint power supply control method for a distributed energy storage system based on user cluster load, comprising the following steps:
[0006] Step 101: Obtain the local voltage and current parameters of the storage unit grid connection point, and calculate the voltage support strength characterization of the grid connection point based on the local voltage and current parameters.
[0007] Step 102: Construct the available energy adjustment coefficient based on the real-time state of charge of the storage cell;
[0008] Step 103: Extract the instruction tracking deviation between the instruction setpoint and the actual output value in the power control loop of the storage unit, and generate a sensitivity factor that characterizes the weight of the topological position of the storage unit in the distributed network based on the energy spectrum distribution of the instruction tracking deviation within a preset time window.
[0009] Step 104: Construct a dynamic virtual impedance mapping model by integrating the voltage support strength characterization quantity and the sensitivity factor. When the voltage support strength characterization quantity decreases and the sensitivity factor exceeds the preset sensitivity threshold, reduce the virtual impedance rise slope of the memory cell.
[0010] Step 105: Calculate the cross-correlation coefficient between instruction tracking deviation and phase fluctuation in local voltage parameters, and generate a phase compensation offset related to the hardware execution characteristics of the memory cell based on the cross-correlation coefficient.
[0011] Step 106: Combining the available energy adjustment coefficient, the dynamic virtual impedance mapping model, and the phase compensation offset, adjust the output power slope of the storage cell and the inverter pulse width modulation wave so that multiple storage cells generate asynchronous power compensation responses according to their respective electrical positions in the topology.
[0012] Preferably, the step of generating the sensitivity factor in step 103 includes: step 201, acquiring the power setpoint and actual output power of the inner power loop of the storage unit in real time, and calculating the difference between the two to obtain the command tracking deviation; step 202, performing second-order time derivative on the command tracking deviation, and extracting the convergence characteristics of the command tracking deviation changing with time; step 203, determining the real-time impedance magnitude of the grid connection point based on the local voltage parameters and current parameters, and mapping the convergence characteristics with the real-time impedance magnitude to determine the sensitivity factor.
[0013] Preferably, the step of constructing the dynamic virtual impedance mapping model in step 104 includes: step 301, setting a positive correlation rule between the sensitivity factor and the power output share of the storage unit; step 302, when the voltage support strength characterization quantity decreases and the sensitivity factor is in a preset high value range, reducing the virtual impedance of the storage unit so that the storage unit closer to the load center can preferentially compensate for the power deficit.
[0014] Preferably, the step of generating the phase compensation offset in step 105 includes: step 401, extracting the high-frequency components higher than the base frequency in the command tracking deviation, and capturing the instantaneous phase angle deviation value of the grid connection point voltage parameter; step 402, performing cross-correlation operation on the high-frequency components and the instantaneous phase angle deviation value to determine the disturbance quantity characterizing the hardware execution specificity of the storage unit; step 403, converting the disturbance quantity into the corresponding phase gain value and superimposing it on the inverter pulse width modulation wave, and realizing phase decoupling of the inter-node response through the hardware execution residual of the storage unit.
[0015] Preferably, the method further includes a hardware protection step based on instruction tracking deviation: Step 501, performing an integral operation on the time axis on the instruction tracking deviation to obtain the deviation integral characteristic value; Step 502, using the internal resistance thermal effect law of the internal power medium of the storage cell, mapping the deviation integral characteristic value to a thermal accumulation stress factor; Step 503, when the thermal accumulation stress factor exceeds the preset safety limit, increasing the impedance base in the dynamic virtual impedance mapping model to forcibly reduce the output ratio of the storage cell.
[0016] Preferably, the method further includes a step of compensating for response hysteresis: Step 601, calculate the rate of change of the grid connection point voltage and extract the second-order time derivative of the local voltage parameter to determine the load acceleration operator; Step 602, multiply the load acceleration operator with the available energy adjustment coefficient to determine the transient inertia compensation amount. It satisfies: ,in, This is the transient inertia compensation amount. For load acceleration operator, The available energy adjustment coefficient. The preset gain coefficient is used; in step 603, a transient inertial compensation amount is superimposed on the power compensation response to compensate for the response lag of the storage unit in the initial stage of the load step.
[0017] Preferably, the method further includes the step of performing asymmetric damping regulation: step 701, determining the voltage recovery polarity of the grid connection point and identifying whether the distributed network is in the voltage recovery period; step 702, when the distributed network is in the voltage recovery period and the storage unit is in the output power limit state, using the asymmetric correction coefficient to reduce the impedance drop slope in the dynamic virtual impedance mapping model; step 703, controlling the regulation speed of the storage unit returning to steady state, so that the regulation speed is greater than its response speed when entering the load compensation stage.
[0018] Preferably, the method further includes a step of resonant avoidance during the grid connection phase: step 801, injecting a frequency of [frequency value missing] into the power drive pulse of the storage cell. The characteristic excitation signal; Step 802, collect the feedback current of the grid connection point, and use the bandpass filtering algorithm to extract the response characteristic current corresponding to the characteristic excitation signal; Step 803, compare the response characteristic current with the preset impedance safety envelope, and according to the comparison result, translate the frequency domain center of the dynamic virtual impedance mapping model to avoid the inherent resonance point of the distributed network.
[0019] Preferably, the method further includes a phase adaptive calibration step: step 901, monitoring the phase relationship between the voltage and current jump vectors at the grid connection point and obtaining the phase deviation factor of the line impedance angle; step 902, identifying the impedance control attribute of the current grid connection point based on the phase deviation factor, the impedance control attribute including inductive attribute, resistive attribute and resistive-inductive hybrid attribute; step 903, adjusting the reactive power damping weight in the dynamic virtual impedance mapping model so that the phase of the current vector output by the storage unit is aligned with the polar axis of the local voltage parameter.
[0020] Preferably, the method further includes the step of performing collaborative arbitration: step 1001, setting a trigger window for energy fluctuation attenuation, and activating transient inertial compensation only when the absolute value of the rate of change of the grid connection point voltage exceeds a preset sudden change threshold; step 1002, dynamically sorting the access priority of each storage unit using a sensitivity factor; step 1003, within the trigger window, controlling the storage units sorted in the preset high priority range to perform power compensation response.
[0021] Compared with the prior art, the beneficial effects of the present invention are:
[0022] 1. In an energy storage system, the instantaneous bus stiffness characteristic and the physical identifier of the storage unit form a nonlinear decoupling factor that interacts with each other, breaking the synchronization of the response of large-scale storage clusters to user cluster load fluctuations. The bus stiffness reflects the physical sensitivity of the voltage at the access point to the rate of change of current. Combined with the hardware identifier, a slight offset is generated, which causes storage units distributed in different topological locations to generate staggered power compensation response sequences, eliminating feedback resonance caused by the homogenization of node control logic.
[0023] 2. By utilizing the dynamic residual characteristics between the command setpoint and the actual execution result of the power control loop, the storage unit can autonomously identify its electrical topology location. The power execution residual reflects the tightness of the electrical coupling between the storage unit and the local bus. The residual signal feature extraction combined with the bus stiffness calculation of the sensitivity factor enables nodes closer to the load center to spontaneously obtain higher power regulation weights. This uncovers the accompanying information within the control system, avoids dependence on external communication links, eliminates the constraints of communication delays or interruptions on the reliability of power storage scheduling, and improves the system's adaptive adjustment capability under complex time-varying topologies.
[0024] 3. The correlation mapping between the power execution residual evolution trend and the thermal accumulation stress of the storage hardware enables cross-domain collaborative protection of power distribution targets and equipment health status. The nonlinear shift of the residual signal with the change of bus stiffness is monitored, and the physical fatigue and thermal stress state generated by the high-frequency adjustment of power devices are inverted. Based on this, the energy adequacy operator is dynamically corrected. When the node thermal stress is detected to be close to the safety threshold, the dynamic virtual impedance is increased to force the node output share to transfer to neighboring nodes, balance the service intensity within the cluster, and extend the overall service life of the power storage medium and power electronic devices. Attached Figure Description
[0025] Figure 1 This is a schematic diagram of the distributed joint power supply control process for user cluster loads according to the present invention;
[0026] Figure 2 This is a comparison chart of the bus voltage fluctuation suppression effect of the present invention under different load steps;
[0027] Figure 3 This is a three-tier architecture diagram of the intelligent control system integrating perception modeling and execution according to the present invention. Detailed Implementation
[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. The following embodiments are intended to explain the present invention, and not to limit the scope of protection of the present invention.
[0029] This invention provides a joint power supply control method for a distributed energy storage system based on user cluster loads, comprising five stages: local electrical feature extraction, energy state characterization, topology sensitivity analysis, virtual impedance model construction, and phase decoupling control. It collects local voltage and current parameters at the grid connection point through a converter (power conversion system), and performs deep feature mining based on the command execution deviation of the power control loop. This enables the storage unit to autonomously perceive its topological location, thereby suppressing voltage oscillations caused by homogenized cluster responses and ensuring network stability even without external communication support by dynamically adjusting virtual impedance and phase compensation. In distribution networks with high user cluster load penetration, the resistive characteristics of the end lines fluctuate drastically with load characteristics, causing the inductive impedance model based on the traditional droop control assumption to fail. To address this issue, this invention acquires the local voltage parameters at the grid connection point of the storage unit. With current parameters And based on local voltage parameters With current parameters Calculate the voltage support strength characterization of the grid connection point. The steps are as follows: The controller uses its built-in sensors to collect the voltage and current vectors at the grid connection point, and obtains the rate of change of voltage through single-cycle differential processing. With the rate of change of current Using the ratio formula Obtain voltage support strength characterization ; based on the grid connection point voltage at From the inside Fall to And the current from Increase to For example, the calculation yields and Then the instantaneous voltage support strength characterization quantity at that moment This characterization quantity The system reflects the physical sensitivity of the access point voltage to load fluctuations in real time; to ensure that the distributed storage units have sufficient energy during adjustment, the system executes adjustments based on the real-time state of charge of the storage units. Constructing the available energy regulation coefficient The steps; this coefficient The calculation uses a method based on The piecewise correlation function, when In to When the interval is, and Maintain a positive correlation; when Exceed or below At the same time, by adjusting the slope of the coefficient, the depth of the node's participation in the adjustment is reduced, thereby protecting the storage medium from physical damage.
[0030] In complex distributed networks, nodes often cannot distinguish between global disturbances on the power grid side and control interventions from neighboring nodes. This perceptual ambiguity is a contributing factor to system control instability. The system extracts the command setpoint from the power control loop of the storage unit. Compared with the actual output value Instruction tracking deviation And generate a sensitivity factor based on its energy spectrum distribution; the specific procedure is as follows: calculate the instantaneous difference of the inner power loop. Using a sliding time window Perform second-order time differentiation to extract its convergence characteristic over time; then compare this convergence characteristic with... and The determined real-time impedance magnitude is correlated and mapped; if the node is located in a high-load area at the end of the line, its command tracking deviation is... The convergence trajectory has higher fluctuation energy, and the corresponding sensitivity factor value increases; based on this, the system identifies the importance of the node in the topology, completes the data to decision support conversion process, and the controller acquires the discrete sequence of command tracking deviation at a sampling rate of 20kHz. The Hanning window function is applied to perform windowing processing, and the energy integral value of the low-frequency oscillation range from 2Hz to 10Hz is extracted by Fast Fourier Transform. Combined with the real-time impedance modulus of the grid connection point Construct a nonlinear mapping operator k and calculate the sensitivity factor. Comparison With the mean energy distribution within the quasi-steady-state period and standard deviation If satisfied This triggers the high-sensitivity topology output enhancement logic, operator Based on the experimental determination of the impedance gradient of the distribution line length-to-diameter ratio; and addressing the problem of power overshoot caused by the inability of existing technologies to dynamically adjust the output elasticity according to the bus stiffness, this invention integrates voltage support strength characterization. A dynamic virtual impedance mapping model is constructed using the sensitivity factor; a logical closed loop is established between voltage, current, and virtual impedance, in the voltage support strength characterization quantity. When the sensitivity factor decreases and exceeds the preset sensitivity threshold, the virtual impedance rise slope of the storage cell is reduced. When a node near the load center senses that the bus support capability is weakening and its sensitivity is high, it reduces the virtual impedance to make it respond to the power deficit first, preventing voltage drops caused by adjustment lag. Meanwhile, nodes at the topology edge maintain a higher virtual impedance. This differentiated configuration enables tens of thousands of nodes to generate asynchronous power compensation responses, avoiding logic resonance.
[0031] To eliminate response synchronization issues when large-scale storage clusters handle load fluctuations, the system calculates instruction tracking deviations. The cross-correlation coefficient between the phase fluctuation and the local voltage parameters is used to generate the phase compensation offset; the controller extracts... The high-frequency components above the fundamental frequency are captured, and the instantaneous phase angle deviation of the grid-connected point voltage parameters is captured. Cross-correlation is performed on these two components to determine the hardware execution specificity of the node. This offset is converted into a phase gain value and superimposed on the inverter's pulse width modulation wave. Due to slight differences in the aging degree and internal characteristics of the hardware at each node, the resulting phase compensation offset has decoupling properties. Without the need for physical identifiers, staggered power compensation sequences are ensured between nodes. The high-frequency component Ph above 100Hz in the command tracking deviation is extracted, and the instantaneous phase angle deviation of the grid-connected point voltage is captured. Perform cross-correlation operation ,in For discrete sampling sequence numbers, This is a time delay quantity, used to obtain the disturbance quantity that reflects the delay characteristics of the drive circuit. Substitute into the phase gain conversion function The phase compensation offset is calculated and superimposed onto the reference angular frequency of the inverter's space vector pulse width modulation algorithm. Phase decoupling of the output current vector is achieved through the microsecond-level hardware execution residual generated by the drive pulse delay of the switching devices at each node. The conversion coefficient α is calibrated based on the experimental curve of the heterogeneous converter's drive pulse response. When dealing with the step impact caused by the start-up of large inductive load groups, the system performs a compensation response hysteresis step. The load acceleration operator is determined by calculating the rate of change of the grid-connected voltage and extracting the second-order time derivative of the voltage parameters. ; this operator With available energy adjustment coefficient Perform product operations to determine the transient inertia compensation amount. , This is the transient inertia compensation amount. For load acceleration operator, The available energy adjustment coefficient. The preset gain coefficient, and this transient inertial compensation amount satisfy: This compensation amount is superimposed on the power compensation response.
[0032] When load unloading causes the system to enter the recovery period, if the return speeds of each node are different, cross-node reactive power circulation will occur. The system executes an asymmetric damping adjustment step; it determines the voltage recovery polarity of the grid connection point and identifies whether the system is in the voltage recovery period; when the voltage recovers and the storage unit is in a power-limited state, it uses an asymmetric correction coefficient to reduce the impedance drop slope in the dynamic virtual impedance mapping model; this strategy controls the adjustment speed of the storage unit to steady-state regression to be greater than its response speed when entering the load compensation phase, thereby achieving high synchronization of cluster nodes when exiting the saturation state and eliminating the risk of circulation current caused by physical differences between nodes; to ensure the safety of the system during dynamic access, this invention executes a resonance avoidance procedure during the grid connection phase; a frequency of [missing information] is injected into the power drive pulse of the storage unit. The characteristic excitation signal has an amplitude lower than the power frequency voltage. The system collects feedback current from the grid connection point and extracts the corresponding response characteristic current using a bandpass filter algorithm; it compares this current with the preset impedance safety envelope; if an abnormal increase in amplitude is detected, indicating the presence of a potential resonant point, the system shifts the frequency domain center of the dynamic virtual impedance mapping model to avoid the inherent resonant point; to address the hardware thermal aging risk caused by long-term operation, the system executes hardware protection steps; and it tracks instruction deviations. Perform integration along the time axis to obtain the deviation integral characteristic value; utilize the internal resistance thermal effect of the storage medium to map this characteristic value into a thermal accumulation stress factor. ;exist When the preset safety limit is exceeded, the impedance base in the dynamic virtual impedance mapping model is increased, and the output ratio of the storage unit is forcibly reduced, so that the adjustment load is transferred to the neighboring nodes with better health status, thereby extending the overall service life of the system.
[0033] Example 1: In an urban power distribution terminal area containing two hundred distributed storage units, when a large number of electric vehicle fast charging piles and a central air conditioning system cluster are started simultaneously in the evening, the voltage support strength characterization of the grid connection point is as follows: From the normal state Down to The system obtains the local voltage parameters of each storage unit's grid connection point. With current parameters By calculating the voltage change rate With the rate of change of current The ratio determines the voltage support strength characterization quantity. Synchronous with the state of charge Constructing the available energy regulation coefficient And extract the command setpoint from the power control loop. Compared with the actual output value Instruction tracking deviation ,in, It is a characterization of voltage support strength. The rate of change of voltage. The rate of change of current, In a charged state, The available energy adjustment coefficient. For instruction tracking deviation, The instruction is given a value. This is the actual output value.
[0034] Instruction execution tracking deviation for each memory unit The second-order time derivative is used to extract the convergence feature that varies with time, and the energy spectrum distribution of this convergence feature within the sliding time window is calculated. This generates a sensitivity factor characterizing the weight of each storage cell's topological location in the distributed network. For storage cells located at the load center, the instruction tracking deviation... The convergence trajectory exhibits specific energy distribution characteristics. Based on this, the system adjusts the corresponding sensitivity factor to above the preset sensitivity threshold and utilizes a dynamic virtual impedance mapping model to reduce the virtual impedance rise slope of the memory cell. Simultaneously, it calculates the instruction tracking deviation. The cross-correlation coefficient with the phase fluctuations in the local voltage parameters generates a phase compensation offset related to the hardware execution characteristics of the storage cells, which is then superimposed on the inverter pulse width modulation wave. Due to the differences in the physical characteristics of each storage cell, this phase compensation offset causes each storage cell to generate an asynchronous power compensation response sequence. The system monitors the rate of change of the grid-connected voltage and extracts the second-order time derivative of the voltage parameters to determine the load acceleration operator. and the load acceleration operator With available energy adjustment coefficient Perform product operations to determine the transient inertia compensation amount. , This is the transient inertia compensation amount. For load acceleration operator, The available energy adjustment coefficient. The preset gain coefficient, and this transient inertial compensation amount satisfy: The transient inertial compensation is superimposed in the compensation response. The system adjusts the output current feedback component. When the load is unloaded and the voltage rises, the system determines the voltage recovery polarity of the grid connection point. When the storage unit is in the output power limit state, the system uses the asymmetric correction coefficient to reduce the impedance drop slope in the dynamic virtual impedance mapping model, so that the adjustment rate of the storage unit returning to steady state is faster than its response rate when entering the load compensation stage, and finally limits the voltage fluctuation of the grid connection point to within the preset safety threshold.
[0035] Example 2: This experiment verifies the dynamic response accuracy and network operation stability of a distributed energy storage system when handling nonlinear user cluster loads. The experimental platform simulates random power fluctuations in an urban transformer area during the evening peak load period using a hardware-in-the-loop simulation environment. The experimental data comes from a high-sampling-rate data acquisition link, which has... Time resolution and voltage signal Current signal Sensing accuracy; core sampling period The settings involve balancing the real-time nature of data updates with the load calculated by the controller; due to the monitored bus voltage vector The transient bandwidth is in to Within the interval, to satisfy the Nyquist sampling criterion and provide phase margin, the sampling frequency needs to be set at [value missing]. The above; applied in the experiment The sampling interval is used to determine the voltage support strength characterization quantity. The extraction does not have calculation errors due to spectral aliasing.
[0036] To simulate the electromagnetic environment at the end of the power distribution line, a signal-to-noise ratio of [value missing] was injected into the acquisition channel during the experiment. Gaussian white noise, and introduce Gaussian white noise containing The low-frequency background disturbance of the total harmonic distortion rate is used to examine the instruction tracking deviation of each memory cell. The convergence characteristics adjust the output performance; the test group at the load occurrence to When the gradient changes abruptly, synchronous monitoring commands track the deviation. The correlation with grid connection point voltage fluctuations; among which, command tracking deviation Give a value to the instruction Compared with the actual output value The difference; the sensitivity factor is determined by the instruction tracking deviation. The second time derivative in The energy integral within the sliding window is determined.
[0037] Table 1: Comparison of experimental data of the method of the present invention under different load intensity gradients
[0038]
[0039] See Table 1, when the load step is in to When the voltage fluctuation is within the specified range, the peak value of the voltage fluctuation after adjustment by this method stabilizes at... The noise component is suppressed to the voltage fundamental frequency amplitude below. Within; this is due to the nonlinear reconstruction of node output achieved by the dynamic virtual impedance mapping model based on the sensitivity factor; when the total load increases to At that time, the peak voltage fluctuation rose to And the response latency increased to Experimental results confirm that, even under environmental interference, the method of this invention, utilizing the topology sensitivity factor derived from the instruction execution residual, can drive each distributed node to generate asynchronous power compensation responses, thereby effectively suppressing cluster oscillations caused by load jumps; transient inertial compensation amount Satisfy the formula This compensation amount is in the initial stage of voltage drop. Internal damping support is provided.
[0040] Example 3: This example combines Figures 1 to 3 This document describes a joint power supply control method for a distributed energy storage system based on user cluster load, as follows: Figure 1As shown, the system performs steps to collect grid connection point parameters to obtain local voltage U and current I. Based on the rate of change of voltage and current, it calculates the voltage support strength characterization quantity, extracts the command tracking deviation (the difference between the given value and the output value), and generates a sensitivity factor by analyzing the deviation energy spectrum distribution. The system integrates the voltage support strength characterization quantity and the sensitivity factor to construct a dynamic virtual impedance mapping model. Simultaneously, the system acquires the real-time state of charge (SoC) of the storage unit and constructs the available energy adjustment coefficient based on the SoC state segmentation mapping. Furthermore, it generates a phase compensation offset by calculating the cross-correlation coefficient between the deviation and the voltage phase. Finally, it combines the available energy adjustment coefficient, the dynamic virtual impedance mapping model, and the phase compensation offset to adjust the power output slope and the inverter pulse width modulation wave, thereby generating an asynchronous power compensation response that eliminates feedback resonance and extends lifespan.
[0041] like Figure 2 As shown in the chart, the horizontal axis represents the total load step, and the vertical axis represents the peak voltage fluctuation V. Four specific load step test points of 100kW, 200kW, 350kW, and 450kW are set in the chart. By comparing the original peak voltage fluctuation of the bus voltage with the peak voltage fluctuation after adjustment using a bar chart, the data shows that under all test conditions, the peak voltage fluctuation after adjustment using this method is lower than the original state. Under the high-power step condition of 450kW, the adjusted voltage fluctuation still maintains a low amplitude, verifying the system's voltage smoothing capability over a wide load range. Figure 3 As shown, the overall architecture is centered on the intelligent controller of the storage unit and is vertically divided into three capability levels. Capability level one is the multi-dimensional feature perception layer, which includes three functional modules: calculating voltage support strength characterization, extracting command tracking deviation convergence features, and constructing available energy adjustment coefficients. Its data interaction logic is as follows: the voltage support strength and deviation convergence features flow to capability level two, i.e., the dynamic model construction layer, to generate the topology position sensitivity factor, hardware characteristic phase compensation offset, and construct a dynamic virtual impedance mapping model. The available energy adjustment coefficients generated by capability level one are directly passed to the subsequent links and converge with the model parameters output by capability level two to capability level three, i.e., the collaborative response execution layer. This layer integrates the above parameters to adjust the inverter pulse width modulation wave and execute asynchronous power compensation response, thereby forming a complete closed loop from feature perception to execution response.
[0042] Example 4: In a microgrid test platform containing ten distributed storage units, each storage unit employs a controller with floating-point arithmetic capabilities, and the sampling rate of its voltage sensor is set to... And the quantization accuracy is better than The sensing accuracy of the current sensor is no less than During the startup phase, each node executes the initial state definition procedure, which involves monitoring the voltage amplitude of the local bus under no-load conditions and setting it as... The per-unit value benchmark is used, and the initial impedance base in the dynamic virtual impedance mapping model is set to the impedance modulus per unit length of the microgrid line. This ensures the system has sufficient damping margin at the moment of grid connection; the system then enters the parameter calibration process to address command tracking deviations. The controller processes the data. Step size extraction A series of continuous power deviation sampling points are used to preprocess the time-domain signal using the Hanning window function to suppress spectral leakage, followed by the execution of a Fast Fourier Transform. The energy spectral density distribution of the power deviation is obtained, and the following is extracted. to The energy integral value within the frequency band is used to characterize the coupling strength between the node and the oscillation component generated by load changes.
[0043] To determine the preset sensitivity threshold, this embodiment uses the statistical distribution method for parameter calibration, while the system is operating in a quasi-steady state. Within each sampling period, the controller records the fluctuation sequence of the energy integral value in real time and calculates its mean. with standard deviation And set the preset sensitivity threshold to This enables the system to incorporate sensitivity compensation logic when the coupled energy deviates from the statistical normal due to load fluctuations, while also addressing the load acceleration operator. The preset mutation threshold is determined by applying the second-order time derivative of the grid connection point voltage to... The peak values within the observation window are statistically sampled, and a threshold value is set as the maximum acceleration of voltage fluctuation within that observation window. This establishes an adaptive judgment criterion for specific topological environments; among which, For instruction tracking deviation, For Fast Fourier Transform, The mean, Standard deviation, For load acceleration operators; tracking deviations during command execution. With thermal cumulative stress factor When mapping, the system multiplies the eigenvalue of the deviation integral by a dimensionless number. The internal resistance equivalent temperature rise coefficient is determined by the measured internal resistance and specific heat capacity parameters of the storage medium, thereby determining the thermal accumulation stress factor. It can quantitatively reflect the temperature rise trend of power devices, when the thermal accumulation stress factor Exceeding the safe temperature rise value of the corresponding device When the forced adjustment procedure of the trigger impedance base is applied; Following a step load impact, each node, based on the sensitivity factor determined according to the aforementioned calibration procedure, [is subject to certain conditions]. Gradient reconstruction of the output force is completed internally, and the convergence trajectory of the grid connection point voltage is compared with the voltage support strength characterization quantity. The predicted physical model matches, and no current exceeding the rated current was generated throughout the process. The transient impact of multiples ultimately confirmed the logical consistency and operational stability of the control loop under engineering constraints; among which, This is the thermal cumulative stress factor. It is a characterization of voltage support strength.
[0044] Example 5: In a distributed distribution substation deployment scenario involving the integration of multiple brand converters, each storage unit performs a baseline parameter calibration process before grid connection to establish a physical correspondence. The system extracts the rated capacity and real-time state of charge of the storage unit. The internal resistance temperature rise slope of the storage medium under different charge levels was determined through controlled charge-discharge experiments, and the measured slope values were then recorded in the available energy adjustment coefficient. In the corresponding association mapping table, through to Real-time state of charge Sampling points are evenly distributed within the range to determine the available energy adjustment coefficient. The evolution trajectory of the energy surplus changes, so that hardware units from different batches can express the force elasticity on the same physical scale when adjusting the power output slope.
[0045] When the system faces the risk of feedback resonance due to the homogeneity of response logic among heterogeneous hardware nodes, each memory cell executes an identifier resolution procedure to convert the physical identifier of the memory cell into a discrete offset increment in the control loop. This procedure maps the unique hardware identifier to a current gain deviation value within a set range, and then uses the voltage support strength characterization value of the access point. During the grid-connected commissioning phase, weighted scaling is implemented. Each node generates a response misalignment in the time series based on the calculated phase compensation offset, and the rate of change of the grid-connected point voltage is synchronously acquired. With the rate of change of current This is used to verify the degree of asynchronicity of response sequences between nodes, enabling the distributed cluster to maintain the spatial decoupling of current vectors in each branch during the process of handling instantaneous fluctuations in user cluster load.
[0046] Example 6: During the delivery and commissioning of a heterogeneous microgrid system, each storage unit determines the transient inertia compensation amount by executing the front-end loop gain calibration procedure. Preset gain coefficient The procedure applies step values. Using a resistive step load as the reference excitation source, the controller synchronously monitors the rate of change of the grid connection point voltage and extracts the second-order time derivative of the voltage parameters to determine the load acceleration operator. By adjusting the preset gain coefficient The value of the value is observed and the time constant of the grid connection point voltage recovering from the drop point to the steady state is measured. When this time constant is within the range of... to The range and voltage support strength characterization quantity The fluctuation amplitude is lower than At that time, determine the corresponding preset gain coefficient. ;in, This is the transient inertia compensation amount. The preset gain coefficient, For load acceleration operator, It is a characterization of voltage support strength.
[0047] During the grid-connected self-test process, the system utilizes the energy state distribution consistency algorithm to adjust the available energy adjustment coefficient. Implementing baseline alignment, the controller reads the real-time state of charge inside the storage cell. The available energy adjustment coefficient is determined by combining the measured internal resistance temperature rise slope. The initial zero point is used to suppress internal circulating currents between nodes caused by differences in control parameters. The system injects a frequency of [frequency value missing] into the inverter pulse width modulation wave. The characteristic excitation signal is collected and the response characteristic current is acquired. The amplitude response characteristics of the response characteristic current are used to correct the initial impedance base in the dynamic virtual impedance mapping model. This procedure ensures that the absolute value of the deviation of the output power slope of each node when dealing with the same user cluster load fluctuations is less than 100%. This allows for convergence of cluster energy allocation while maintaining differentiated responses in the physical identifiers of storage cells, thus avoiding voltage recovery instability caused by initial parameter discrepancies; whereby... The available energy adjustment coefficient. For real-time state of charge, For frequency.
[0048] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.
[0049] 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. A joint power supply control method for a distributed energy storage system based on user cluster load, characterized in that, Includes the following steps: Step 101: Obtain the local voltage and current parameters of the storage unit grid connection point, and calculate the voltage support strength characterization of the grid connection point based on the local voltage and current parameters. Step 102: Construct the available energy adjustment coefficient based on the real-time state of charge of the storage cell; Step 103: Extract the instruction tracking deviation between the instruction setpoint and the actual output value in the power control loop of the storage unit, and generate a sensitivity factor that characterizes the weight of the topological position of the storage unit in the distributed network based on the energy spectrum distribution of the instruction tracking deviation within a preset time window. Step 104: Construct a dynamic virtual impedance mapping model by integrating the voltage support strength characterization quantity and the sensitivity factor. When the voltage support strength characterization quantity decreases and the sensitivity factor exceeds the preset sensitivity threshold, reduce the virtual impedance rise slope of the memory cell. Step 105: Calculate the cross-correlation coefficient between instruction tracking deviation and phase fluctuation in local voltage parameters, and generate a phase compensation offset related to the hardware execution characteristics of the memory cell based on the cross-correlation coefficient. Step 106: Combining the available energy adjustment coefficient, the dynamic virtual impedance mapping model, and the phase compensation offset, adjust the output power slope of the storage cell and the inverter pulse width modulation wave so that multiple storage cells generate asynchronous power compensation responses according to their respective electrical positions in the topology.
2. The method for joint power supply control of a distributed energy storage system based on user cluster load according to claim 1, characterized in that, The steps in step 103 to generate the sensitivity factor include: Step 201, acquiring the power setpoint and actual output power of the inner power loop of the storage unit in real time, and calculating the difference between the two to obtain the command tracking deviation; Step 202, performing second-order time derivative on the command tracking deviation, and extracting the convergence characteristics of the command tracking deviation over time; Step 203, determining the real-time impedance magnitude of the grid connection point based on the local voltage and current parameters, and mapping the convergence characteristics to the real-time impedance magnitude to determine the sensitivity factor.
3. The method for joint power supply control of a distributed energy storage system based on user cluster load according to claim 1, characterized in that, The steps in step 104 for constructing the dynamic virtual impedance mapping model include: step 301, setting a positive correlation rule between the sensitivity factor and the power output share of the storage unit; step 302, when the voltage support strength characterization quantity decreases and the sensitivity factor is in a preset high value range, reducing the virtual impedance of the storage unit so that the storage unit closer to the load center can preferentially compensate for the power deficit.
4. The method for joint power supply control of a distributed energy storage system based on user cluster load according to claim 1, characterized in that, The step of generating the phase compensation offset in step 105 includes: step 401, extracting the high-frequency components higher than the base frequency in the command tracking deviation, and capturing the instantaneous phase angle deviation value of the grid connection point voltage parameter; step 402, performing cross-correlation operation on the high-frequency components and the instantaneous phase angle deviation value to determine the disturbance quantity characterizing the hardware execution specificity of the storage unit; step 403, converting the disturbance quantity into the corresponding phase gain value and superimposing it on the inverter pulse width modulation wave, and realizing phase decoupling of the inter-node response through the hardware execution residual of the storage unit.
5. The method for joint power supply control of a distributed energy storage system based on user cluster load according to claim 1, characterized in that, The method also includes a hardware protection step based on instruction tracking deviation: Step 501, perform an integral operation on the time axis on the instruction tracking deviation to obtain the deviation integral characteristic value; Step 502, use the internal resistance thermal effect law of the internal power medium of the storage cell to map the deviation integral characteristic value into a thermal accumulation stress factor; Step 503, when the thermal accumulation stress factor exceeds the preset safety limit, increase the impedance base in the dynamic virtual impedance mapping model to forcibly reduce the output ratio of the storage cell.
6. The method for joint power supply control of a distributed energy storage system based on user cluster load according to claim 1, characterized in that, The method also includes a step to compensate for response hysteresis: Step 601, calculate the rate of change of the grid connection point voltage and extract the second-order time derivative of the local voltage parameters to determine the load acceleration operator; Step 602, multiply the load acceleration operator with the available energy adjustment coefficient to determine the transient inertia compensation amount. It satisfies: ,in, This is the transient inertia compensation amount. For load acceleration operator, The available energy adjustment coefficient. The preset gain coefficient is used; in step 603, a transient inertial compensation amount is superimposed on the power compensation response to compensate for the response lag of the storage unit in the initial stage of the load step.
7. The method for joint power supply control of a distributed energy storage system based on user cluster load according to claim 1, characterized in that, The method also includes the step of performing asymmetric damping regulation: Step 701, determine the voltage recovery polarity of the grid connection point and identify whether the distributed network is in the voltage recovery period; Step 702, when the distributed network is in the voltage recovery period and the storage unit is in the output power limit state, use the asymmetric correction coefficient to reduce the impedance drop slope in the dynamic virtual impedance mapping model; Step 703, control the regulation speed of the storage unit to return to steady state, so that the regulation speed is greater than its response speed when entering the load compensation stage.
8. The method for joint power supply control of a distributed energy storage system based on user cluster load according to claim 1, characterized in that, The method also includes a step of resonant avoidance during the grid connection phase: step 801, injecting a frequency of [frequency missing] into the power drive pulse of the storage cell. The characteristic excitation signal; Step 802, collect the feedback current of the grid connection point, and use the bandpass filtering algorithm to extract the response characteristic current corresponding to the characteristic excitation signal; Step 803, compare the response characteristic current with the preset impedance safety envelope, and according to the comparison result, translate the frequency domain center of the dynamic virtual impedance mapping model to avoid the inherent resonance point of the distributed network.
9. The method for joint power supply control of a distributed energy storage system based on user cluster load according to claim 1, characterized in that, The method also includes a phase adaptive calibration step: Step 901, monitor the phase relationship between the voltage and current jump vectors at the grid connection point and obtain the phase deviation factor of the line impedance angle; Step 902, identify the impedance control attribute of the current grid connection point based on the phase deviation factor, the impedance control attribute includes inductive attribute, resistive attribute and resistive-inductive hybrid attribute; Step 903, adjust the reactive power damping weight in the dynamic virtual impedance mapping model so that the phase of the current vector output by the storage unit is aligned with the polar axis of the local voltage parameter.
10. The method for joint power supply control of a distributed energy storage system based on user cluster load according to claim 1, characterized in that, The method also includes the step of performing collaborative arbitration: Step 1001, setting a trigger window for energy fluctuation attenuation, and activating transient inertial compensation only when the absolute value of the rate of change of the grid connection point voltage exceeds a preset sudden change threshold; Step 1002, dynamically sorting the access priority of each storage unit using a sensitivity factor; Step 1003, within the trigger window, controlling the storage units sorted in the preset high priority range to perform power compensation response.
Citation Information
Patent Citations
Distributed energy storage system for prolonging service life of energy storage battery
CN220022338U
