Phase prediction-based energy storage system black start load recovery control method and system

Abstract

This invention belongs to the field of new energy and energy-saving technology, specifically relating to a black-start load recovery control method and system for an energy storage system based on phase prediction. First, a real-time dynamic safety window and real-time prediction confidence level are calculated for the predicted phase difference value. Corresponding thresholds are preset and compared. Only when the predicted phase difference value is within the dynamic safety window and the prediction confidence level is not lower than the threshold, in the predicted future... The system issues closing commands at all times, achieving precise contact closure through mechanical delay. The confidence model is constructed by combining predicted stability terms and system drastic change penalty terms. At the same time, for cold-start loads with no residual charge, it is optimized to close the circuit at the zero-crossing point of the grid voltage, effectively suppressing inrush current and realizing the transformation of closing from blind control to precise control, thereby improving the safety, reliability and control accuracy of closing.

Description

A Black-Start Load Recovery Control Method and System for Energy Storage Systems Based on Phase Prediction Technical Field

[0001] This invention belongs to the field of new energy and energy-saving technology, specifically relating to a black-start load recovery control method and system for energy storage systems based on phase prediction. Background Technology

[0002] With the rapid development of electrochemical energy storage technology, energy storage systems have become an indispensable power source in grid black start schemes due to their excellent fast power response capability and flexible four-quadrant operation capability. In the energy storage-based black start process, the energy storage system first establishes an islanded microgrid with stable voltage and frequency, and then gradually restores important loads within the grid according to a predetermined sequence. The reliability of this load restoration phase directly determines the success or failure of the entire black start process.

[0003] Currently, when using energy storage systems for black start, load restoration mainly relies on two technologies: one is automated program control based on fixed timing, and the other is remote manual control by dispatchers. The core logic of these two existing technologies is to issue a circuit breaker closing command immediately or after a fixed delay after detecting that the target line is energized and the voltage is normal. This method treats the energy storage system as a simple power source and fails to fully utilize its advantages as a power electronic converter with rapid controllability.

[0004] Existing closing strategies completely ignore the dynamic relationship between the grid voltage phase and load characteristics at the actual closing moment of the circuit breaker. In the weak grid environment established by energy storage systems, this "random closing" method is prone to closing the circuit at unfavorable phase points (such as near voltage peaks), especially when restoring inductive loads such as transformers and motors, which can generate inrush currents as high as 6-8 times the rated current. This may not only cause the overcurrent protection of the energy storage converter to trip, interrupting the restoration process, but also cause irreversible electromagnetic damage to the circuit breaker equipment and the load itself.

[0005] Existing methods simplify energy storage systems to the use of traditional rotating generators, failing to leverage their core advantages as fully controllable power electronic devices. Energy storage converters inherently possess millisecond-level precise control capabilities, adjusting output characteristics within each switching cycle; however, existing coarse-grained control strategies completely waste this potential, failing to achieve refined control of the closing process. During the unique process of black start, the grid structure is fragile and has poor damping characteristics, yet existing methods still employ the same closing logic as normal grids. They neither consider the differentiated impact characteristics of different load types (such as motors, transformers, and nonlinear loads) nor dynamically adjust the closing strategy based on the system's real-time stability, lacking necessary risk warning and avoidance mechanisms. If a recovery path fails due to inrush current, existing technology lacks automatic reconfiguration capabilities, requiring manual intervention to redo the recovery plan. This significantly delays valuable recovery time and may even lead to the failure of the entire black start process. These technical problems severely restrict the full effectiveness of energy storage systems in black start applications. Summary of the Invention

[0006] This invention provides a black-start load restoration control method and system for energy storage systems based on phase prediction. It addresses the existing load restoration closing strategies for black-start energy storage systems, which fail to utilize the advantages of fast and controllable power electronic equipment in energy storage converters. They employ the same coarse fixed / manual closing logic as the normal power grid, neglecting the dynamic relationship between grid voltage phase and load characteristics, as well as the differentiated characteristics of weak grids and different loads. This leads to a series of faults caused by large inrush currents, and lacks dynamic adjustment, risk warning and avoidance, and automatic fault reconfiguration capabilities, severely restricting the application effect and restoration reliability of energy storage black-start systems.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] A black-start load recovery control method for an energy storage system based on phase prediction includes the following steps:

[0009] The grid-side voltage phase and the residual voltage phase of the load-side line of the circuit breaker are monitored and acquired. Based on the grid-side voltage phase and the residual voltage phase of the load-side line, the real-time phase difference is calculated.

[0010] Based on the real-time phase difference, calculate the rate of change of the phase difference; based on the rate of change of the phase difference, calculate the predicted value of the phase difference using a kinematic displacement prediction model.

[0011] The real-time dynamic safety window and real-time prediction confidence level are calculated based on the phase difference prediction value. A confidence level threshold is preset, and it is determined whether the phase difference prediction value is within the real-time dynamic safety window. The real-time prediction confidence level is compared with the confidence level threshold, and the circuit breaker closing control is performed based on the judgment result and the comparison result.

[0012] The step of calculating the phase difference change rate based on the real-time phase difference specifically involves: calculating the phase difference change rate using a frequency-adaptive change rate correction term based on the real-time phase difference. The formula for calculating the rate of change of phase difference is as follows:

[0013]

[0014] in, Let be the real-time phase difference at time t. Let be the grid-side voltage frequency at time t. Let be the frequency of the residual voltage on the load side at time t. The frequency change rate weighting coefficient is 360, which is the core conversion coefficient. The phase difference change rate calculation formula is as follows: This is the rate of change correction term for frequency adaptation.

[0015] The step of calculating the predicted phase difference value based on the phase difference change rate and the kinematic displacement prediction model is as follows:

[0016] Based on the obtained phase difference change rate and real-time phase difference Based on the kinematic displacement prediction model, the phase difference prediction model is derived, and the future time is calculated. The predicted phase difference value, and the kinematic displacement prediction model are as follows:

[0017]

[0018] in, For the predicted displacement at future time T, The instantaneous displacement at the initial moment. Let be the instantaneous velocity at the initial moment, and T be the prediction time interval.

[0019] The future moment The formula for calculating the phase difference prediction value is as follows:

[0020]

[0021] In the formula, For the future moment The predicted phase difference value.

[0022] In future moments Phase difference prediction value Using the system rated voltage, circuit breaker rated current, load equivalent impedance, and empirical coefficients for inrush current as inputs, the real-time dynamic safety window allowed for the current phase difference prediction value is calculated. Simultaneously, by recording the predicted and actual phase difference values ​​over the past N periods, the standard deviation and average value of the recent phase difference prediction error are calculated. The acceleration of the phase difference change is obtained by differential calculation of the phase difference change rate. Combining the standard deviation, average value, and acceleration of the phase difference change of the recent phase difference prediction error, an acceleration penalty coefficient and a smoothing constant are introduced to calculate the real-time prediction confidence level. According to the real-time dynamic security window and real-time prediction confidence For future moments A comprehensive evaluation is performed on the predicted phase difference values.

[0023] The real-time dynamic security window The calculation formula is as follows:

[0024]

[0025] In the formula, Basic security window, The surge current sensitivity coefficient, This is an estimated value of the impact current based on the predicted phase difference. ,in, This is an empirical coefficient for the impact current. Converting from degrees to radians This is the rated current of the circuit breaker. The system's rated voltage. This is the equivalent impedance of the load.

[0026] The real-time prediction confidence The calculation formula is as follows:

[0027]

[0028] In the formula, This represents the standard deviation of the recent phase difference prediction error. This represents the average value of recent phase difference prediction errors. This is the smoothing constant.

[0029] Preset reliability threshold Predicting future moments Phase difference prediction value Is it in a real-time dynamic security window? Within this context, the real-time prediction confidence level is compared with a confidence threshold; if and only if the real-time predicted future time... Phase difference prediction value It is within the real-time dynamic safety window allowed by the current predicted value, i.e. ;

[0030] Meanwhile, real-time prediction confidence level When the circuit breaker is closed, a closing command is sent.

[0031] For loads that start cold, the residual voltage phase of the load-side line is 0. When the grid-side voltage crosses zero, a closing command is sent to the circuit breaker.

[0032] A phase prediction-based black start load recovery control system for energy storage systems includes a phase acquisition module, a phase prediction module, and a dynamic evaluation module.

[0033] The phase acquisition module is used to monitor and acquire the grid-side voltage phase and the residual voltage phase of the load-side line of the circuit breaker, and calculate the real-time phase difference based on the grid-side voltage phase and the residual voltage phase of the load-side line.

[0034] The phase prediction module is used to calculate the phase difference change rate based on the real-time phase difference, and to calculate the phase difference prediction value based on the phase difference change rate and the classical kinematic displacement prediction model.

[0035] The dynamic evaluation module is used to calculate the real-time dynamic safety window and real-time prediction confidence based on the phase difference prediction value, preset the confidence threshold, determine whether the phase difference prediction value is within the real-time dynamic safety window, compare the real-time prediction confidence with the confidence threshold, and perform circuit breaker closing control based on the judgment result and comparison result.

[0036] Compared with the prior art, the present invention has the following beneficial effects:

[0037] This invention establishes a dynamic phase tracking and accurate prediction mechanism from the source by monitoring the voltage phase between the grid side and the load side, calculating the real-time phase difference, optimizing the phase difference change rate calculation by combining a frequency adaptive correction term, and deriving the predicted phase difference value at future moments based on a classical kinematic displacement prediction model. This effectively adapts to the dynamic relationship between grid voltage phase and load characteristics, solves the control deviation problem caused by the traditional strategy ignoring the correlation between the two, significantly reduces the probability of large excitation inrush current generation, avoids a series of faults such as grid oscillation and equipment damage caused by inrush current, and significantly improves the safety of load recovery for energy storage black start.

[0038] Furthermore, this invention possesses a powerful dynamic adjustment and risk warning and avoidance capability through a dual evaluation system of dynamic safety window and real-time prediction confidence. The real-time dynamic safety window adaptively adjusts based on multi-dimensional parameters such as system rated parameters, load characteristics, and inrush current coefficient, flexibly matching the safe closing range according to different operating conditions and adapting to the differentiated characteristics of weak power grids and various loads. The confidence assessment integrates recent prediction error statistics and phase difference change acceleration penalty terms, comprehensively evaluating the reliability of prediction results, accurately identifying system transient processes, and avoiding erroneous closing operations when linear prediction models fail. Compared with the shortcomings of traditional strategies lacking risk control, this significantly improves the reliability and stability of load restoration.

[0039] Furthermore, a dedicated optimization strategy was designed for cold-start loads. When there is no residual voltage phase on the load side, the circuit breaker automatically selects the zero-crossing point of the grid voltage for closing, minimizing transient flux surges and inrush currents in inductive loads. This fills the gap in existing technologies for differentiated adaptation to different load characteristics and expands the applicable scenarios of the method. Simultaneously, the closing control logic is precisely matched with the circuit breaker's operating time. By predicting the closing time that meets the conditions in advance and issuing the command, mechanical delay is used to ensure the contacts close at the moment of minimum phase difference, further enhancing closing accuracy, reducing the impact of closing operations on the grid and load equipment, and extending equipment lifespan.

[0040] The control system of this invention, through the coordinated design of three major modules—phase acquisition, phase prediction, and dynamic evaluation—has a clear structure and rigorous logic. It achieves fully automated operation from phase monitoring and predictive calculation to closing control, requiring no manual intervention. This solves the problems of low efficiency and large errors associated with traditional manual closing, improving the ease and efficiency of black-start load restoration. Compared to existing technologies, this invention achieves breakthroughs in control logic, adaptability, and risk management, effectively overcoming the core bottlenecks restricting the application effect and restoration reliability of energy storage black-start systems. It provides an efficient, safe, and universal technical solution for black-start load restoration of energy storage systems, possessing broad engineering application value. Attached Figure Description

[0041] Figure 1 is a schematic flowchart of a black-start load recovery control method for an energy storage system based on phase prediction in an embodiment of the present invention. Detailed Implementation

[0042] To further understand the content of this invention, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments are merely illustrative and not limiting of the invention.

[0043] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0044] This embodiment proposes a black-start load recovery control method for energy storage systems based on phase prediction, as shown in Figure 1, including the following steps:

[0045] The grid-side voltage phase and the residual voltage phase of the load-side line of the circuit breaker are monitored and acquired. Based on the grid-side voltage phase and the residual voltage phase of the load-side line, the real-time phase difference is calculated.

[0046] Based on the real-time phase difference, calculate the rate of change of the phase difference; based on the rate of change of the phase difference, calculate the predicted value of the phase difference using a classical kinematic displacement prediction model.

[0047] The real-time dynamic safety window and real-time prediction confidence level are calculated based on the phase difference prediction value. A confidence level threshold is preset, and it is determined whether the phase difference prediction value is within the real-time dynamic safety window. The real-time prediction confidence level is compared with the confidence level threshold, and the circuit breaker closing control is performed based on the judgment result and the comparison result.

[0048] Based on the above method steps, this application is used in the black-start load recovery phase after the establishment of an islanded microgrid in an energy storage system. For circuit breaker closing control, the precise closing method of this application is implemented. Relying on the millisecond-level precise control capability of the energy storage converter, real-time acquisition of circuit breaker phase data is performed through voltage transformers on both sides of the circuit breaker. Specifically, before the circuit breaker closing operation, the voltage phase of the grid side is synchronously monitored and acquired through voltage transformers on both sides of the circuit breaker. and the residual voltage phase of the load-side line For a line that has just been de-energized, there is a decaying residual voltage on the load side of the line caused by distributed capacitance and residual magnetic field. This method ignores the magnitude of the residual voltage and only locks its phase precisely. Even if the residual voltage amplitude is less than 10% of the rated voltage, it still maintains accurate phase detection.

[0049] Based on the grid-side voltage phase and the residual voltage phase of the load-side line, the real-time phase difference between the grid-side voltage phase and the residual voltage phase of the load-side line is calculated. By calculating the real-time phase difference Search The moment when the voltage on the grid side and the load side are in phase (=0), the circuit breaker is closed at this moment. Closing the circuit when the phase difference is 0 can fundamentally avoid instantaneous voltage differences caused by phase differences, suppress inrush current to the greatest extent, and solve the drawbacks of random closing.

[0050] As a mechanical and electrical device, the circuit breaker has a fixed mechanical action delay from receiving the closing command to the actual closing of the contacts. If the closing command is only issued when the real-time phase difference is equal to 0, after a mechanical delay of t, the voltage phase between the grid side and the load side has changed dynamically and deviated from the same point, resulting in the actual closing phase difference not being 0. The optimal closing timing fails, and inrush current will still be generated. Therefore, it is necessary to control by predicting in advance to solve the timing matching problem between the mechanical action delay of the circuit breaker and the rapid change of the grid phase, so as to achieve precise "advance" control.

[0051] Based on the real-time phase difference at time t The phase difference change rate is calculated using a frequency-adaptive rate-of-change correction term. The formula for calculating the phase difference change rate is as follows:

[0052]

[0053] in, The grid-side voltage frequency at time t is in Hz and is obtained in real time through grid-side phase-locked loop measurement. The frequency of the residual voltage on the load side at time t is expressed in Hz and is measured in real time by the load-side phase-locked loop. If there is no residual voltage, the value is 0. is the frequency change rate weighting coefficient, in units of S, determined through system simulation, with a typical value of 0.1~0.5; 360 is the core conversion coefficient, converting frequency-related quantities into angle-related quantities. In the formula, The main term is the rate of change of phase difference. This is a frequency adaptive rate of change correction term, used to take into account the frequency's own changing trend. If the grid frequency is deviating from the load frequency at an accelerating rate, this term will increase the predicted rate of change, thereby compensating in advance and making the prediction more accurate.

[0054] Based on the phase difference change rate obtained above and the real-time phase difference with time t Based on the kinematic displacement prediction model, the phase difference prediction model is derived, and the future time is calculated. The predicted phase difference value, and the kinematic displacement prediction model are as follows:

[0055] ,in, For the predicted displacement at future time T, The instantaneous displacement at the initial moment. Let be the instantaneous velocity at the initial moment, and T be the prediction time interval. Based on the kinematic displacement prediction model, the future moments are derived. The formula for calculating the phase difference prediction value is as follows:

[0056]

[0057] In the formula, For the future moment The predicted phase difference value; the future time This refers to the circuit breaker closing time; it is the real-time phase difference between the grid-side and load-side voltages measured at time t. This provides an initial reference point for prediction, and then the phase difference change rate is calculated by real-time phase difference. It quantifies the change in phase difference during the circuit breaker's operating delay, enabling high-precision prediction of the phase difference after the circuit breaker's mechanical delay.

[0058] In future moments Phase difference prediction value Using the system rated voltage, circuit breaker rated current, load equivalent impedance, and empirical coefficient parameters of inrush current as the core input, the real-time dynamic safety window allowed for the current phase difference prediction value is calculated. The system determines the safe allowable range corresponding to the current phase difference prediction value, enabling the closing safety allowable range to automatically adapt and adjust as the phase difference prediction value changes. It records the phase difference prediction and actual values ​​for the past N cycles, calculates the standard deviation and average value of the recent phase difference prediction error, and performs differential calculation on the phase difference change rate to obtain the acceleration of the phase difference change. Combining the standard deviation, average value, and acceleration of the recent phase difference prediction error with the acceleration penalty coefficient and smoothing constant, the system calculates the real-time prediction confidence level. According to the real-time dynamic security window and real-time prediction confidence Regarding the future The phase difference prediction value at each moment is comprehensively evaluated; the real-time dynamic safety window The calculation formula is as follows:

[0059]

[0060] In the formula, This is the basic safety window, set according to system impedance and withstand capability; a typical value is 15°. The surge current sensitivity coefficient was determined through simulation and experimentation, with typical values ​​ranging from 1.0 to 3.0. This is an estimated value of the impact current based on the predicted phase difference. ,in, The inrush current empirical coefficient is determined according to the load type: large transformers: 2.5 ~ 6.0, large motors: 4.0 ~ 8.0, resistive loads: 1.0 ~ 1.2; The system's rated voltage. The predicted phase difference at the closing time. Converting from degrees to radians This is the rated current of the circuit breaker. The load equivalent impedance. In the formula, the basic safety window... Defined as the closing window under the most ideal conditions. This constitutes a dynamic adjustment factor to realize dynamic safety window calculation. When the predicted inrush current is larger, the real-time dynamic safety window becomes smaller, thereby dynamically tightening the real-time dynamic safety window. This requires a more precise closing phase. When the predicted inrush current is small, the safety window is widened to improve the closing success rate. This embodiment upgrades the closing criterion from a fixed threshold to an adaptive, dynamically changing intelligent threshold related to the predicted inrush intensity, achieving a balance between safety and success rate.

[0061] The real-time prediction confidence The calculation formula is as follows:

[0062]

[0063] In the formula, This represents the standard deviation of the recent phase difference prediction error. This represents the average value of recent phase difference prediction errors. The smoothing constant is used to prevent the denominator from being zero. In this embodiment, the smoothing constant is... Set to a minimum value of 0.1°; The acceleration due to the phase difference change. This is the acceleration penalty coefficient, used to quantify the negative impact of acceleration on confidence. This embodiment constructs a drastic change penalty term. Combined with the recent phase difference prediction error stability term used for predicting stability A comprehensive confidence assessment model was constructed. In the confidence assessment model, the phase difference prediction error stability term... The confidence level is reflected by historical prediction errors. The larger the historical prediction error, the stronger the stability term of the phase difference prediction error. The larger the value of the term, the lower the confidence level. (This refers to a term related to drastic change penalties.) In this paper, the phase difference change acceleration is innovatively introduced as a criterion. When the system is in a transient state and the phase difference changes rapidly, the confidence level will be significantly reduced, avoiding the execution of erroneous closing commands when the linear prediction model fails, and ultimately achieving a comprehensive assessment of prediction stability and system transient risk.

[0064] Preset reliability threshold If and only if the future moments are predicted in real time Phase difference prediction value It is within the real-time dynamic safety window allowed by the current predicted value, i.e. Meanwhile, real-time prediction confidence level When the circuit breaker receives the closing command, it sends a closing command to the circuit breaker. The confidence threshold is typically set to 0.8 to ensure the reliability of the prediction results. After receiving the closing command, the circuit breaker, after its own mechanical delay time... Its contacts will close near the moment when the phase difference is the smallest, thereby achieving precise phase reclosing and reducing the impact of closing shock on the system. If any condition is not met, such as the phase difference exceeding the safety window or the real-time prediction confidence level being lower than the threshold, a closing command will not be issued, and real-time monitoring and calculation will continue until both conditions are met simultaneously.

[0065] In a preferred embodiment, for a load that has undergone a cold start, the residual voltage phase of the load-side line is 0. When the voltage on the grid side crosses zero, the transient magnetic flux impact generated by closing the circuit breaker is minimized. This effectively suppresses inrush current and electromagnetic torque impact, preventing damage to the motor windings due to excessive transient current and reducing disturbances to the power grid. Therefore, closing the circuit breaker is triggered when the grid side voltage crosses zero.

[0066] To verify the performance differences and technical advantages of the intelligent predictive closing method of this invention compared with the traditional random closing method, a comparative experiment was conducted in this embodiment. The test system parameters were set as follows: rated voltage 10kV, rated frequency 50 Hz, and the test load was a 10MVA unloaded transformer (typical impact coefficient K). i =5.0), circuit breaker operating time = 80ms, each method performs 20 closing operations. The parameters of the method of this invention are set as follows: basic safety window α = 15°, confidence threshold is 0.8, and impulse current sensitivity coefficient β = 2.0; the test results are shown in Table 1 below.

[0067] Table 1

[0068]

[0069] The test results show that the success rate of traditional random closing is 45% (9 / 20), while the success rate of intelligent predictive closing of this invention is 95% (19 / 20), an improvement of +111%; the average inrush current of traditional random closing is 3,250A, which is reduced to 850A by the method of this invention, a reduction of -74%; the maximum inrush current is reduced from 6,800A to 1,450A, a reduction of -79%; the voltage sag is reduced from 28% to 8%, a reduction of -71%; the system oscillation duration is shortened from 1.2s to 0.3s, a reduction of -75%; and the number of misoperations is reduced from 11 to 1, a reduction of -91%. The above experimental data show that, compared with traditional random closing, the intelligent predictive closing method proposed in this invention has significant advantages in safety, reliability and control accuracy. Its core superiority lies in the fact that it realizes a fundamental transformation from "blind closing" to "precise control" through phase difference prediction value, dynamic safety window and real-time prediction confidence assessment, effectively solving the problem of impact risk and high failure rate of load restoration in black start.

[0070] Based on the phase prediction-based black start load recovery control method for energy storage systems proposed in the above embodiments, this embodiment also proposes a phase prediction-based black start load recovery control system for energy storage systems, including a phase acquisition module, a phase prediction module, and a dynamic evaluation module.

[0071] The phase acquisition module is used to monitor and acquire the grid-side voltage phase and the residual voltage phase of the load-side line of the circuit breaker, and calculate the real-time phase difference based on the grid-side voltage phase and the residual voltage phase of the load-side line.

[0072] The phase prediction module is used to calculate the phase difference change rate based on the real-time phase difference, and to calculate the phase difference prediction value based on the phase difference change rate and the classical kinematic displacement prediction model.

[0073] The dynamic evaluation module is used to calculate the real-time dynamic safety window and real-time prediction confidence based on the phase difference prediction value, preset the confidence threshold, determine whether the phase difference prediction value is within the real-time dynamic safety window, compare the real-time prediction confidence with the confidence threshold, and perform circuit breaker closing control based on the judgment result and comparison result.

[0074] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can be appropriately combined to form other embodiments that can be understood by those skilled in the art. The above content is only for illustrating the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made based on the technical concept proposed in this invention shall fall within the scope of protection of the claims of this invention.

Claims

1. A black-start load recovery control method for an energy storage system based on phase prediction, characterized in that, Includes the following steps: The grid-side voltage phase and the residual voltage phase of the load-side line of the circuit breaker are monitored and acquired. Based on the grid-side voltage phase and the residual voltage phase of the load-side line, the real-time phase difference is calculated. Based on the real-time phase difference, a frequency-adaptive rate-of-change correction term is used to calculate the phase difference change rate. Based on this rate of change and a kinematic displacement prediction model, the phase difference prediction model is derived, and the predicted phase difference value is calculated. The formula for calculating the phase difference change rate is as follows: in, The rate of change of phase difference Let be the real-time phase difference at time t. Let be the grid-side voltage frequency at time t. Let be the frequency of the residual voltage on the load side at time t. The frequency change rate is the weighting coefficient, and 360 is the core conversion coefficient. The rate of change correction term is for frequency-adaptive operation; the kinematic displacement prediction model is as follows: in, For the predicted displacement at future time T, The instantaneous displacement at the initial moment. Let be the instantaneous velocity at the initial moment, and T be the prediction time interval; the formula for calculating the predicted phase difference is as follows: In the formula, For the future moment The phase difference prediction value is calculated; the real-time dynamic safety window and real-time prediction confidence are calculated based on the phase difference prediction value, a confidence threshold is preset, it is determined whether the phase difference prediction value is within the real-time dynamic safety window, and the real-time prediction confidence is compared with the confidence threshold. The circuit breaker closing control is performed based on the judgment result and the comparison result.

2. The black-start load recovery control method for an energy storage system based on phase prediction according to claim 1, characterized in that, In future moments Phase difference prediction value Using the system rated voltage, circuit breaker rated current, load equivalent impedance, and empirical coefficients for inrush current as inputs, the real-time dynamic safety window allowed for the current phase difference prediction value is calculated. Simultaneously, by recording the predicted and actual phase difference values ​​over the past N periods, the standard deviation and average value of the recent phase difference prediction error are calculated. The acceleration of the phase difference change is obtained by differential calculation of the phase difference change rate. Combining the standard deviation, average value, and acceleration of the phase difference change of the recent phase difference prediction error, an acceleration penalty coefficient and a smoothing constant are introduced to calculate the real-time prediction confidence level. According to the real-time dynamic security window and real-time prediction confidence For future moments A comprehensive evaluation is performed on the predicted phase difference values.

3. The black-start load recovery control method for an energy storage system based on phase prediction according to claim 2, characterized in that, The real-time dynamic security window The calculation formula is as follows: In the formula, Basic security window, The surge current sensitivity coefficient, This is an estimated value of the impact current based on the predicted phase difference. ,in, This is an empirical coefficient for the impact current. Converting from degrees to radians This is the rated current of the circuit breaker. The system's rated voltage. This is the equivalent impedance of the load.

4. The black-start load recovery control method for an energy storage system based on phase prediction according to claim 2, characterized in that, The real-time prediction confidence The calculation formula is as follows: In the formula, This represents the standard deviation of the recent phase difference prediction error. This represents the average value of recent phase difference prediction errors. This is the smoothing constant.

5. The black-start load recovery control method for an energy storage system based on phase prediction according to claim 2, characterized in that, Preset reliability threshold Predicting future moments Phase difference prediction value Is it in a real-time dynamic security window? Within this context, the real-time prediction confidence level is compared with a confidence threshold; if and only if the real-time predicted future time... Phase difference prediction value It is within the real-time dynamic safety window allowed by the current predicted value, i.e. Simultaneously, real-time prediction confidence level When the circuit breaker is closed, a closing command is sent.

6. The black-start load recovery control method for an energy storage system based on phase prediction according to claim 1, characterized in that, For loads that start cold, the residual voltage phase of the load-side line is 0. When the grid-side voltage crosses zero, a closing command is sent to the circuit breaker.

7. A phase prediction-based black-start load recovery control system for an energy storage system, based on the phase prediction-based black-start load recovery control method for an energy storage system according to any one of claims 1 to 6, characterized in that, It includes a phase acquisition module, a phase prediction module, and a dynamic evaluation module. The phase acquisition module monitors and acquires the grid-side voltage phase and the residual voltage phase of the load-side line of the circuit breaker, and calculates the real-time phase difference based on these phase phases. The phase prediction module calculates the phase difference change rate using a frequency-adaptive rate-of-change correction term based on the real-time phase difference, and derives a phase difference prediction model based on a classical kinematic displacement prediction model to calculate the predicted phase difference value. The formula for calculating the phase difference change rate is as follows: in, The rate of change of phase difference Let be the real-time phase difference at time t. Let be the grid-side voltage frequency at time t. Let be the frequency of the residual voltage on the load side at time t. The frequency change rate is the weighting coefficient, and 360 is the core conversion coefficient. The rate of change correction term is for frequency-adaptive operation; the kinematic displacement prediction model is as follows: in, For the predicted displacement at future time T, The instantaneous displacement at the initial moment. Let be the instantaneous velocity at the initial moment, and T be the prediction time interval; the formula for calculating the predicted phase difference is as follows: In the formula, For the future moment The phase difference prediction value; the dynamic evaluation module is used to calculate the real-time dynamic safety window and the real-time prediction confidence based on the phase difference prediction value, preset the confidence threshold, determine whether the phase difference prediction value is within the real-time dynamic safety window, compare the real-time prediction confidence with the confidence threshold, and perform circuit breaker closing control based on the judgment result and the comparison result.

Images