A safety control method for preventing reverse flow, overload and demand control of an energy storage system

Abstract

This invention belongs to the field of energy storage system safety control technology, and discloses a safety control method for backflow prevention, overload prevention, and demand management in energy storage systems, solving the problem of the lack of a unified coordination mechanism for the three types of safety constraints in energy storage systems. It includes: checking whether the meters at each control point are online, confirming that all meters are online; sequentially performing overload prevention checks, backflow prevention checks, and demand management checks; wherein, all three checks are executed using the PCC point power value after dynamic load tracker selection and replacement processing, and if an emergency response condition is triggered during execution, the system directly enters the emergency response phase; all three safety checks generate power upper limits through a limit accumulation mechanism; after the limit is closed through checks, a proportional power limit propagation method is used to issue the final power control command to the energy storage converter. This invention unifies and coordinates the three types of safety constraints, eliminating command conflicts.

Description

Technical Field

[0001] This invention belongs to the field of energy storage system safety control technology, and particularly relates to a safety control method for energy storage systems to prevent backflow, prevent overload, and manage demand. Background Technology

[0002] As a key energy conversion hub connecting distributed new energy sources (photovoltaics, wind power, etc.) with user loads and the power grid, the safe operation of energy storage systems involves multiple dimensions of power boundary constraints. In industrial and commercial energy storage application scenarios, energy storage systems are connected to the main power grid through the Point of Common Coupling (PCC), and need to meet the following three types of safety constraints simultaneously: (1) Anti-reverse flow constraint: In some application scenarios, the user side is not allowed to send power back to the grid (i.e., reverse flow is not allowed). When the distributed photovoltaic power generation is greater than the load power, the excess power is fed back to the grid through the PCC point, which may violate the grid connection agreement or cause protection malfunction. Existing technologies usually adopt a fixed threshold comparison method, and trigger a power limiting command when the power of the PCC point is lower than a certain threshold. (2) Anti-overload constraint: When the energy storage system is charging and discharging, the apparent power on the transformer or incoming cable of the PCC point must not exceed its rated capacity. Existing technologies usually only make overload judgments based on active power, without considering the influence of power factor, which can easily lead to misjudgment or omission in scenarios with large actual reactive power. (3) Demand control constraints: The maximum demand (MD) of a user is constrained by the power supply contract. Exceeding the declared demand will result in a demand charge penalty. The energy storage system needs to regulate the power at the PCC point through charging and discharging to keep the power within the demand limit. Existing technologies usually use simple threshold comparisons, which lack optimization for power allocation in multi-PCC scenarios.

[0003] The main defects of the existing technology are: (1) The three types of safety constraints lack a unified coordination mechanism. The existing technology usually treats anti-reverse current, anti-overload, and demand control as three independent modules and executes them separately, each issuing power commands independently, which leads to command conflicts or uncontrollable superposition effects. For example, the anti-reverse current module requires a reduction in discharge power, while the demand control module may simultaneously require an increase in charging power. The superposition of the two commands may cause the PCS (energy storage converter) to receive conflicting power limits.

[0004] (2) Overload judgment is based solely on active power, without considering apparent power and power factor. In actual operation, when the power factor is low, even if the active power does not exceed the limit, the apparent power may have already exceeded the rated capacity of the transformer. Using active power alone as the overload criterion poses a safety hazard.

[0005] (3) Demand reduction allocation lacks an adaptive mechanism in multi-PCC scenarios. When the system contains multiple PCC points (multiple transformers), the existing technology usually distributes the total excess demand evenly to each PCC without considering the overload degree and adjustable margin of each PCC, resulting in some PCCs being over-limited while some PCCs still have unused margin.

[0006] (4) Power measurement fluctuations cause control jitter. The power measurement value at the PCC point is affected by factors such as load fluctuations and sampling accuracy, resulting in instantaneous fluctuations. Directly using instantaneous values ​​for safety judgment can easily lead to frequent switching of control strategies (jitter), affecting equipment life and power grid quality.

[0007] (5) Delayed response to security incidents. Existing technologies use a fixed-period polling method for security detection. When emergency events such as backflow or overload occur between two detection cycles, the system cannot respond in a timely manner, posing a security risk.

[0008] (6) The power limit is not accurately issued. Existing technology usually issues absolute power values ​​directly to the PCS. When multiple PCS are running in parallel, it is difficult to ensure the rationality of power distribution among the PCS. Moreover, when the output status of the PCS changes, it needs to be recalculated. Summary of the Invention

[0009] The purpose of this invention is to provide a safety control method for preventing backflow, overload, and demand management in energy storage systems, effectively solving the problem of the lack of a unified coordination mechanism for the three types of safety constraints in energy storage systems: preventing backflow, overload, and demand management.

[0010] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a safety control method for preventing backflow, overload and demand management of an energy storage system, including the following steps: S1, check whether the electricity meters at each checkpoint are online; if any checkpoint electricity meter is offline, set the power limit to the minimum value; if all checkpoint electricity meters are online, proceed to step S2.

[0011] S2. Perform three layers of safety checks in sequence: the first layer is overload prevention check, the second layer is backflow prevention check, and the third layer is demand control check.

[0012] Among them, the three-layer safety checks are all performed by the PCC point power value after the dynamic load tracker selects the nearest replacement. If the edge triggers the emergency response condition during the execution, it will directly enter the emergency response stage and trigger the safety control.

[0013] All three safety checks generate power limits through a limit accumulation mechanism: the overload protection check first sets a basic limit based on the transformer's apparent power, and each subsequent check further tightens the limit based on the power limit set in the previous level. No subsequent check may relax the power limit set in the previous level.

[0014] S3. After the limit is checked and closed, the final power control command is issued to the energy storage converter using a proportional power limit propagation method.

[0015] Furthermore, the specific method for overload prevention checks is as follows: S11, obtain the active power, apparent power, and power factor of the PCC point; S12, use a dynamic load tracker to smooth the active power, apparent power, and power factor obtained in step S11 to eliminate instantaneous fluctuations; S13, use apparent power instead of active power as the basis for overload judgment, and calculate the overload amount in combination with the power factor; S14, set dual overload thresholds: a warning threshold and a hard limit threshold. When the overload amount is greater than the hard limit threshold, it is determined to be an overload state and overload counting begins, and power adjustment action is generated; the warning threshold is 95% of the rated capacity of the transformer at the PCC point, and the hard limit threshold is 100% of the rated capacity of the transformer at the PCC point.

[0016] Furthermore, the specific method for backflow prevention checks is as follows: S21, preset the backflow trigger threshold and backflow release threshold, and satisfy the condition that the backflow trigger threshold < the backflow release threshold; S22, obtain the active power at the PCC point; S23, when the transformer power at the PCC point is < the backflow trigger threshold, calculate the energy storage discharge power limit, and reduce the discharge power of the energy storage system according to the energy storage discharge power limit; when the transformer power at the PCC point is > the backflow release threshold, increase the discharge power of the energy storage system within the preset constraint range, and the increased discharge power does not exceed the power upper limit set by the overload prevention layer; when the backflow trigger threshold ≤ the transformer power at the PCC point ≤ the backflow release threshold, maintain the current discharge power of the energy storage system unchanged; S24, pass the discharge power limit to the subsequent layer through the limit accumulation mechanism, and execute the corresponding discharge power adjustment.

[0017] Furthermore, the specific method for demand control inspection is as follows: S31, determine the rated capacity, current active power and preset overload threshold of each PCC point transformer, calculate the overload margin of each transformer, and generate the allocation weight and reduction limit of each transformer.

[0018] S32. Obtain the total reduction required by the system, and use a two-round weighted allocation algorithm to reduce the upper limit of charging power and the upper limit of discharging power of the energy storage devices connected to each PCC point.

[0019] The first round of allocation is based on the allocation weights generated in step S31 to obtain the initial reduction amount of each transformer. It is then determined whether the initial reduction amount of each transformer exceeds the upper limit of the corresponding transformer's reduction. If so, the actual reduction amount of the transformer is limited to the upper limit of the transformer's reduction, and the excess portion is summarized as the remaining reduction amount.

[0020] The second round of allocation involves distributing the remaining reduction capacity to transformers that still have capacity that can be reduced after the first round of allocation, with the allocation ratio being proportional to the remaining capacity that can be reduced of the transformer.

[0021] S33. Demand check and allocation execution.

[0022] Get the total active power of all PCC points at present, and get the upper limit of charging direction demand (demond) and the upper limit of discharging direction demand (demond_d) agreed in the contract.

[0023] The total active power is compared with demond and demond_d respectively, and a safety factor of 0.99 is introduced to tighten the upper limit of demand. If the total active power is within the tightened upper limit of demand, the power limit on each transformer is lifted.

[0024] If the total active power exceeds the tightened demand limit, the excess is calculated and used as the total reduction. Then, the two-round weighted allocation algorithm in step S32 is called to distribute the reduction task to each transformer for execution.

[0025] Furthermore, in step S31, the overload margin is the difference between the transformer's overload threshold and the current apparent power. For transformers with smaller overload margins and larger current active power, the allocation weight is higher.

[0026] The upper limit for reduction of each transformer is determined based on the difference between the rated capacity and the current active power.

[0027] Determine if the allocation weight of all transformers is zero. If so, adjust the allocation weight of all transformers to be proportional to their current active power. If not, maintain the allocation weight generated based on the overload margin.

[0028] Furthermore, in step S3, the proportional power limit propagation method includes: obtaining the limited power and expected power of all energy storage devices connected to the target transformer, and calculating the total limit that the target transformer should have, the ratio of the total limit to the expected power is the power limit ratio.

[0029] If the power limit ratio is ≤0.01, charging and discharging of energy storage devices under the target transformer are not allowed; if the power limit ratio is ≥1, the output power of each energy storage device under the target transformer is allowed to reach 110% of the rated power; if 0.01 < power limit ratio <1, the upper limit of the power limit is calculated based on the power limit ratio and the rated capacity of the energy storage, and is issued to each energy storage device connected to the target transformer. Each energy storage device calculates the actual allowed output power based on its own rated power and the received upper limit of the power limit.

[0030] Furthermore, each PCC point corresponds to a dynamic load tracker, which maintains a fixed-length circular buffer. The power data of the most recently sampled PCC points are cached in the circular buffer. It has logic for exponential weighted regression, nearest replacement, 0.1kW deduplication, and 70kW degradation during cold start. It also extracts trend prediction information and statistical average information from the same cached data. The timing of cached data updates is decoupled from the query.

[0031] Furthermore, a reactive power compensation mechanism is introduced, which uses the reactive power of the energy storage converter to perform closed-loop adjustment of the power factor at the PCC point, thereby reducing apparent power and indirectly alleviating overload. Specifically, the active power and reactive power of the PCC point, as well as the active power and reactive power of the energy storage converter, are collected to calculate the current power factor.

[0032] The system obtains the user-configured reactive power compensation mode, which includes a target reactive power mode and a target power factor mode. If the system is in the target reactive power mode, the reactive power deficit is calculated based on the difference between the configured target reactive power and the current reactive power. If the system is in the target power factor mode, the system calculates the corresponding reactive power boundary based on the configured target power factor and compares the absolute value of the current power factor with the absolute value of the target power factor. If the current power factor is better than the target power factor, the reactive power of the energy storage converter is gradually recovered, and the recovered amount is used as the reactive power deficit to prevent overcompensation. If the current power factor is worse than the target power factor, the difference between the reactive power required to reach the target power factor and the current reactive power is calculated and used as the reactive power deficit.

[0033] Reactive power adjustment commands are generated based on the reactive power deficit and sent to the energy storage converter for execution. The actual reactive power at the PCC point is used as the controlled variable, and the reactive power deficit is used as the control deviation. By adjusting the reactive power output of the energy storage converter, the reactive power at the PCC point is brought to converge within the target value range, thus forming a reactive power compensation closed-loop control.

[0034] Furthermore, the reactive power Ql of each energy storage converter is jointly constrained by the rated capacity Se of the energy storage converter and the current active power Pl: Pl 2 + Ql 2 ≤Se 2 When multiple energy storage converters are connected in parallel, 80% of the sum of the reactive power of each energy storage converter, Qlmax, is the upper limit of the system's reactive power, and the lower limit of the system's reactive power is -80%Qlmax. Furthermore, the sum of the square of the active power and the square of the reactive power of each energy storage converter is not greater than the square of its rated capacity.

[0035] Obtain the current reactive power of the energy storage converter and calculate the reactive power deficit that needs to be compensated. Calculate the initial reactive power target value based on the formula: Initial reactive power target = Current reactive power + Reactive power deficit.

[0036] The initial reactive power target value is compared with ±80%Qlmax, and a limiting process is performed: if the initial reactive power target value is greater than 80%Qlmax, it is clamped to 80%Qlmax, and the undigested reactive power deficit is recorded as the difference between the initial reactive power target value and 80%Qlmax; if the initial reactive power target value is less than -80%Qlmax, it is clamped to -80%Qlmax, and the undigested reactive power deficit is recorded as the difference between the initial reactive power target value and -80%Qlmax; if the initial reactive power target value is between 80%Qlmax and -80%Qlmax, it is determined that the reactive power deficit has been completely digested, and the undigested reactive power deficit is cleared to zero.

[0037] The initial reactive power target value after the limiting process is issued as the final reactive power command to the energy storage converter for execution, and the undigested reactive power deficit is returned to the upper-level system for decision-making.

[0038] Furthermore, an error counter hysteresis mechanism is introduced, which means that when a preset number of abnormal situations occur consecutively, the protection action is triggered, and the protection action is released only after a preset number of recovery signals occur consecutively.

[0039] Compared with the prior art, the beneficial technical effects of the present invention are: (1) Three types of safety constraints are unified and coordinated to eliminate command conflicts: By accumulating the power limits of the three safety modules of overload prevention, reverse flow prevention and demand control in an orderly manner according to priority (coverage → take minimum value → take minimum value and issue), the most stringent limit is ensured to take effect and there will be no command contradictions or superimposed loss of control.

[0040] (2) More accurate overload detection based on apparent power: Overload judgment is made by combining apparent power with power factor, rather than relying solely on active power. In scenarios with low power factor (such as a large number of motor loads), even if the active power does not exceed the limit, the apparent power exceeding the limit can be accurately detected, avoiding the risk of transformer overload. At the same time, dual thresholds of 95% warning and 100% hard limit are set to provide a warning buffer.

[0041] (3) Weighted allocation of multiple PCCs improves demand control efficiency: Demand reduction allocation is performed based on the overload margin of each PCC as a weight. PCCs with severe overload take on more reduction, while PCCs with sufficient margin maintain higher output. The first round allocates according to the weight ratio, and the second round redistributes the remaining amount of PCCs with overload to PCCs with margin, achieving globally optimal power allocation. Compared with average allocation, it reduces unnecessary power constraints.

[0042] (4) The proportional conversion algorithm is adapted to the scenario of multiple PCS in parallel, and the dual threshold hysteresis anti-reverse current avoids frequent switching: the power limit is propagated through the power limit ratio, rather than the absolute power value. Two parameters are set: the reverse current trigger threshold and the reverse current release threshold. When the PCC power is between the two thresholds, it is in the "safety belt" and does not take action.

[0043] (5) A circular cache serves two purposes with a single set of data, combining behavioral prediction and statistical smoothing: Based on the circular cache, the most recent N PCC power samples are cached. The same cached data is used to extract trend information through exponentially weighted linear regression and to extract mean information through weighted statistics, achieving dual reuse of "prediction" and "smoothing". Behavioral prediction allows the system to anticipate load trends and reserve margin in advance; the weighted average value is equivalent to a low-pass filter, replacing instantaneous values ​​in safety judgment, and reflects the real load level better than instantaneous values, avoiding false triggering caused by meter spikes and load switching spikes. The exponential decay weight makes new data contribute more, taking into account both smoothness and the speed of following load step changes. The result cache decouples the query and update timing, without affecting the real-time performance of the control cycle.

[0044] (6) Reactive power compensation mechanism improves power quality and releases transformer capacity: In addition to the three types of active power safety constraints, a reactive power compensation closed loop is introduced: It supports dual modes of target reactive power and target power factor, and uses PCC reactive power as the controlled variable to adjust the PCS reactive power output in a closed loop. The upper limit of reactive power capacity is dynamically calculated according to the P / Q circle constraint with active power output and derated by 20%. When multiple PCS are connected in parallel, the command is always within the system capacity range. Reactive power compensation reduces the PCC reactive power component, causing the apparent power to decrease, indirectly releasing the transformer capacity margin, and coordinating with the overload prevention direction; reactive power commands are issued through an independent channel, without interfering with active power limit control; the clamping recovery logic when the reactive power factor is already better than the target power factor avoids overcompensation, and reactive power exceeding the limit is included in the edge trigger emergency response, and the abnormal response is timely.

[0045] (7) Error counter hysteresis mechanism enhances stability: The hysteresis interval is designed to be 5-10-5. After the error condition lasts for 5 times, the limit is triggered and recovery requires going through 5 normal states; this avoids frequent actions caused by occasional fluctuations and ensures that there is a sufficient recovery buffer period after the safety condition is lifted.

[0046] (8) Edge-triggered emergency response shortens the response time of safety events: Real-time detection of edge changes of conditions such as reverse flow, overload, demand exceeding limits, and abnormal power factor during data update callbacks. When conditions change abruptly, an emergency message is sent immediately to trigger safety control without waiting for the next cycle. Attached Figure Description

[0047] Figure 1 This is a flowchart of the safety control method of the present invention. Detailed Implementation

[0048] Example 1: This example provides a safety control method for energy storage systems to prevent backflow, overload, and demand management. It establishes a three-level sequential coordination safety control architecture, achieves orderly accumulation of power limits by calling the sequence in stages, and combines overload detection based on apparent power and power factor, multi-PCC weighted demand allocation, dynamic load tracking and prediction, edge-triggered emergency response and other technical means to achieve unified and coordinated control of the three types of safety constraints.

[0049] like Figure 1 As shown, the process includes the following steps: S1, call the main entry function to check whether each gate meter is online; if any gate meter is offline, set the power limit to the minimum value; if all gate meters are online, proceed to step S2.

[0050] S2. Perform three layers of safety checks in sequence: the first layer is overload prevention check, the second layer is backflow prevention check, and the third layer is demand control check.

[0051] Among them, the three-layer safety checks are all performed by the PCC point power value after the dynamic load tracker selects the nearest replacement. If the edge triggers the emergency response condition during the execution, it will directly enter the emergency response stage and trigger the safety control.

[0052] All three safety checks generate power limits through a limit accumulation mechanism: the overload protection check first sets a basic limit based on the transformer's apparent power, and each subsequent check further tightens the limit based on the power limit set in the previous level. No subsequent check may relax the power limit set in the previous level.

[0053] S3. After the limit is checked and closed, the final power control command is issued to the energy storage converter using a proportional power limit propagation method.

[0054] (a) Overload prevention check.

[0055] In step S2, the specific method for overload prevention check is as follows: S11, obtain the active power, apparent power and power factor of PCC point.

[0056] S12. Use a dynamic load tracker to smooth the active power, apparent power and power factor obtained in step S11 to eliminate instantaneous fluctuations.

[0057] S13. Apparent power is used instead of active power as the basis for overload judgment, and the overload is calculated in combination with the power factor. Power systems generally have requirements for the power factor. If the power factor is too low, it is judged as abnormal data, and the power factor is treated as 1.

[0058] S14. Set dual overload thresholds: a warning threshold and a hard limit threshold. When the overload exceeds the hard limit threshold, an overload state is determined, overload counting begins, and power regulation is initiated. The warning threshold is 95% of the rated capacity of the transformer at the PCC point, and the hard limit threshold is 100% of the rated capacity of the transformer at the PCC point.

[0059] (ii) Backflow prevention inspection.

[0060] In step S2, the specific method for backflow prevention check is as follows: S21, preset backflow trigger threshold and backflow release threshold, and satisfy backflow trigger threshold < backflow release threshold.

[0061] S22. Obtain the active power of PCC point.

[0062] S23. When the transformer power at point PCC is less than the reverse current trigger threshold, calculate the energy storage discharge power limit and reduce the discharge power of the energy storage system according to the energy storage discharge power limit.

[0063] When the transformer power at the PCC point is greater than the reverse current release threshold, the discharge power of the energy storage system is increased within the preset constraint range, and the increased discharge power does not exceed the power limit set by the overload protection layer; when the reverse current trigger threshold is less than or equal to the transformer power at the PCC point and less than or equal to the reverse current release threshold, the current discharge power of the energy storage system is maintained unchanged.

[0064] The preset constraint range refers to the following: after the reverse current restriction is lifted, the discharge power is allowed to be increased, but the increased discharge power is still subject to the limit written in the overload protection layer of the three-layer architecture and other strategy constraints (such as the maximum output of SOC and PCS).

[0065] S24. The discharge power limit is passed to the subsequent layer through the limit accumulation mechanism and the corresponding discharge power adjustment is performed.

[0066] (iii) Demand control inspection.

[0067] In step S2, the specific method for demand control inspection is as follows: S31, determine the rated capacity, current active power and preset overload threshold of each PCC point transformer, calculate the overload margin of each transformer, and generate the allocation weight and reduction limit of each transformer.

[0068] The overload margin is the difference between the overload threshold of the transformer and the current apparent power. (1) For transformers with smaller overload margins and larger current active power, the allocation weight is higher. (2) Determine the upper limit of reduction for each transformer based on the difference between the rated capacity and the current active power. (3) Determine whether the allocation weight of all transformers is zero (when all transformers are not overloaded, the overload margin is large, and the allocation weight may be all zero or very small). If yes, adjust the allocation weight of all transformers to be proportional to their current active power; if no, maintain the allocation weight generated based on the overload margin.

[0069] S32. Obtain the total reduction required by the system, and use a two-round weighted allocation algorithm to reduce the upper limit of charging power and the upper limit of discharging power of the energy storage devices connected to each PCC point.

[0070] The first round of allocation is based on the allocation weights generated in step S31 to obtain the initial reduction amount of each transformer. It is then determined whether the initial reduction amount of each transformer exceeds the upper limit of the corresponding transformer's reduction. If so, the actual reduction amount of the transformer is limited to the upper limit of the transformer's reduction, and the excess portion is summarized as the remaining reduction amount.

[0071] The second round of allocation involves distributing the remaining reduction capacity to transformers that still have capacity that can be reduced after the first round of allocation, with the allocation ratio being proportional to the remaining capacity that can be reduced of the transformer.

[0072] S33, Demand check and allocation execution: (1) Obtain the total active power of all PCC points at present, and obtain the upper limit of demand in the charging direction (demond) and the upper limit of demand in the discharging direction (demond_d) as agreed in the contract.

[0073] (2) Compare the total active power with demond and demond_d respectively, and introduce a safety factor of 0.99 to tighten the upper limit of demand. If the total active power is within the tightened upper limit of demand, the power limit on each transformer is lifted; if the total active power exceeds the tightened upper limit of demand, the excess amount is calculated and the excess amount is used as the total reduction amount. Then, the two-round weighted allocation algorithm in step S32 is called to send the reduction task to each transformer for execution.

[0074] Excess amount = Total active power at the current PCC point − tightened upper limit of demand (i.e., diamond × 0.99 or diamond_d × 0.99).

[0075] (iv) Proportional power limit propagation method (executed symmetrically in the charging and discharging directions).

[0076] In step S3, the proportional power limit propagation method includes: (1) obtaining the limited power and expected power of all energy storage devices connected to the target transformer, and calculating the total limit of the target transformer. The ratio of the total limit to the expected power is the power limit ratio.

[0077] The total limit should be equal to the expected total power of the energy storage devices connected to the PCC point minus the excess.

[0078] (2) If the power limit ratio is ≤0.01, the energy storage devices under the target transformer are not allowed to charge and discharge.

[0079] If the power limitation ratio is ≥1, then the output power of each energy storage device under the target transformer is allowed to reach 110% of the rated power.

[0080] If 0.01 < power limit ratio < 1, then the upper limit of the power limit is calculated based on the power limit ratio and the rated capacity of the energy storage, and is sent to each energy storage device connected to the target transformer. Each energy storage device calculates the actual allowable output power based on its own rated power and the received upper limit of power.

[0081] (v) A multifunctional dynamic load tracking method based on a ring buffer.

[0082] Each PCC point corresponds to a dynamic load tracker. The dynamic load tracker maintains a fixed-length circular buffer, which caches the power data of the most recently sampled PCC points. It has logic such as exponential weighted regression (λ=0.85, about 10-cycle half-life), nearest replacement, 0.1kW deduplication, and 70kW degradation during cold start. It also extracts trend prediction information and statistical average information from the same cached data.

[0083] (1) Ring cache data structure.

[0084] ① Circular cache write: After new sampled data is written to the cache array, the pointer moves forward in a circular fashion. When the cache is full, the new data naturally overwrites the oldest data, without the need to perform any memory block relocation operations.

[0085] ② Result caching: All derived values ​​are immediately recalculated and cached each time the data is updated. The cached value is returned directly during the query, with a time complexity of O(1) and no impact on the real-time performance of the control cycle.

[0086] ③ Lazy allocation mechanism: The tracker table grows dynamically according to the number of PCC points, and an instance is created only when the first data arrives.

[0087] ④ Deduplication mechanism: When the deviation between the new sampled value and the previous value is less than 0.1kW, it is discarded to avoid polluting the window with duplicate data.

[0088] In summary, the data update process of the dynamic load tracker is as follows: Receive the current power sample value corresponding to the target transformer. Determine if a tracker instance corresponding to the target transformer exists; if not, perform a lazy allocation operation to create a new tracker instance and assign it a preset conservative default prediction value (70kW) as the initial power value; if it exists, directly call the already created tracker instance. Calculate the absolute value of the difference between the current power sample value and the previous power sample value recorded in the tracker instance, and compare this absolute value with a preset change threshold (0.1kW); if the absolute value of the difference is less than the preset change threshold, determine that the current power sample value is duplicate data and discard it to avoid polluting the records with duplicate values ​​generated by polling, and end the update process.

[0089] (2) Reuse of cached data: Behavioral prediction based on weighted regression.

[0090] Perform exponentially weighted linear regression on historical data in the circular cache to extract load behavior trends and achieve predictive functionality: ① Weight design: Sample weights ,in, The number of occurrences is indicated; the latest sample has a weight of 1.0, and the weight of older data decays exponentially (λ=0.85, approximately 10 half-lives), making the prediction focus on recent behavior. ② Regression solution: The slope is calculated using the least squares method. and intercept This yields the load change trend. ③ Forecast output: Forecast value for the next cycle. (Linear extrapolation); ④ Predicting changes : ,in This represents the maximum weighted residual (recent sudden fluctuations). =3.0 (noise amplification factor, physically meaning to reserve 3 times the recent noise standard deviation), meaning that the predicted change simultaneously covers both the trend term and the noise term. ⑤ Goodness of fit: Weighted R² is used to evaluate the reliability of the regression; R² is low when the data is messy.

[0091] Specifically, the historical samples in the circular buffer are processed as follows: ① The samples in the buffer are sorted in chronological order, and each sample is assigned a weight that decays exponentially with time, where the weight of the newest sample is 1.0, and so on. Sample weights , The attenuation factor is a preset value, and 0.8 ≤ With a value of ≤0.9, the weight of historical data decays exponentially with its age, achieving a half-life of approximately 10 sampling periods.

[0092] ② The least squares method is used to perform linear regression on the weighted sample to obtain the slope characterizing the trend of load change. and intercept .

[0093] ③ Based on the slope and intercept, perform linear extrapolation to obtain the predicted value for the next period. .

[0094] ④ Calculate the predicted change : , This is the preset noise amplification factor, and ≥ 2.5 is used to characterize the magnitude of recent sudden fluctuations and to reserve a safety margin.

[0095] ⑤ Calculate the weighted coefficient of determination R², which is used to evaluate the reliability of the regression model. When R² is lower than the preset threshold, the corresponding prediction results are downweighted or ignored.

[0096] ⑥ The slope, predicted value, predicted change and determination coefficient mentioned above are cached as derived calculation results in the tracker instance, so that the external module can directly return the cached value when querying, and the time complexity of the query operation is O(1).

[0097] When the circular buffer is insufficient, it is handled according to degradation logic: When =0 / 1, the default predicted change amount (70kW, corresponding to the traditional manual margin) is used. When =2, Ensure that the tracker provides conservative yet usable output from the very first moment of startup.

[0098] (3) Reuse of cached data: smooth replacement based on weighted average.

[0099] The same circular buffer data is used to calculate the window-weighted average, serving as a more comprehensive basis for judgment than the instantaneous value: ① The principle that the average value is superior to the instantaneous value: Instantaneous power is affected by load switching, meter sampling refresh, communication delays, etc., resulting in spikes, and single-point judgment is prone to false triggering; the weighted average is mathematically equivalent to a low-pass filter, suppressing high-frequency noise while retaining the true low-frequency changes. ② The significance of exponential weighting (rather than simple averaging): Simple moving average treats all samples equally, and it takes the entire window period to catch up after load step changes; exponential weighting makes new samples contribute more and old samples contribute exponentially, balancing smoothness and tracking speed. ③ Application method (nearest replacement principle): When the deviation between the weighted average and the instantaneous value is within the reliable range (50kW), the average value is used to replace the instantaneous value for safety judgment; when the deviation is too large, the instantaneous value is retained to prevent the tracker from masking the true power change when it malfunctions.

[0100] (4) Decoupling the timing of cached data updates from queries.

[0101] The tracker's updates and queries are decoupled: ① Update timing: It is only triggered in the data update callback (emergency channel) - load_trk_update is called when the PCC point _pApp / \_pAct / \_pRact data is sent, ensuring that the cached data is truly "fresh" rather than the duplicate values ​​from periodic polling. ② Query timing: Each link of the security control reads the cached results in O(1) complexity through load_trk_get, decoupling the computational cost from the call frequency. Even if the security control is executed frequently, it will not introduce regression calculation delay.

[0102] (vi) Reactive power compensation mechanism.

[0103] Reactive power compensation supports three operating modes, which can be switched through parameter configuration: mode value 0 means "disabled", which means no reactive power adjustment is performed; mode value 1 means "target reactive power", which means the actual reactive power at the PCC point is adjusted to the given target reactive power; mode value 2 means target power factor, which means the power factor at the PCC point is adjusted to the given target power factor.

[0104] In this embodiment, the power factor of the PCC point is adjusted in a closed loop by the reactive power of the energy storage converter, thereby reducing the apparent power and indirectly alleviating the overload; specifically: (1) collect the active power and reactive power of the PCC point and the active power and reactive power of the energy storage converter, and calculate the current power factor.

[0105] (2) Obtain the reactive power compensation mode configured by the user, wherein the reactive power compensation mode includes a target reactive power mode and a target power factor mode: if it is in the target reactive power mode, then calculate the reactive power deficit based on the difference between the configured target reactive power and the current reactive power; if it is in the target power factor mode, then calculate the corresponding reactive power boundary based on the configured target power factor, and compare the absolute value of the current power factor with the absolute value of the target power factor: if the current power factor is better than the target power factor, then gradually recover the reactive power of the energy storage converter, and use the recovered amount as the reactive power deficit to prevent overcompensation; if the current power factor is worse than the target power factor, then calculate the difference between the reactive power required to reach the target power factor and the current reactive power, and use it as the reactive power deficit.

[0106] (3) Generate reactive power adjustment instructions based on reactive power deficit and issue them to the energy storage converter for execution; wherein, the actual reactive power at PCC point is used as the controlled variable and the reactive power deficit is used as the control deviation. By adjusting the reactive power output of the energy storage converter, the reactive power at PCC point is brought to converge within the target value range, thus forming a reactive power compensation closed-loop control.

[0107] The reactive power command symmetrical limiting and issuance process is as follows: (1) The reactive power Ql of each energy storage converter is constrained by the rated capacity Se of the energy storage converter and the current active power Pl: Pl 2 + Ql 2 ≤Se 2 When multiple energy storage converters are connected in parallel, 80% of the sum of the reactive power of each energy storage converter, Qlmax, is the upper limit of the system's reactive power, and the lower limit of the system's reactive power is -80%Qlmax. Furthermore, the sum of the square of the active power and the square of the reactive power of each energy storage converter is not greater than the square of its rated capacity. (2) Obtain the current reactive power of the energy storage converter and calculate the reactive power deficit that needs to be compensated. Calculate the initial reactive power target value based on the formula: Initial reactive power target = Current reactive power + Reactive power deficit. (3) Compare the initial reactive power target value with ±80%Qlmax and perform limiting processing: If the initial reactive power target value is greater than 80%Qlmax, clamp it to 80%Qlmax and record the undigested reactive power deficit as the difference between the initial reactive power target value and 80%Qlmax; If the initial reactive power target value is less than -80%Qlmax, clamp it to -80%Qlmax and record the undigested reactive power deficit as the difference between the initial reactive power target value and -80%Qlmax; If the initial reactive power target value is between 80%Qlmax and -80%Qlmax, it is determined that the reactive power deficit has been completely digested and the undigested reactive power deficit is cleared to zero.

[0108] (4) The initial reactive target value after the limiting process is sent to the energy storage converter as the final reactive command for execution, and the undigested reactive power deficit is returned to the upper system for decision-making.

[0109] (vii) Error counter hysteresis mechanism.

[0110] This embodiment designs an error counter hysteresis mechanism to control the continuity of safety conditions and the recovery process, avoiding frequent actions caused by occasional fluctuations.

[0111] When five consecutive abnormal conditions occur, a protection action is triggered. Once the protection action is triggered, the counter will immediately be set to 10, entering the "recovery period." Afterward, five consecutive normal conditions must occur, and the counter will gradually decrease from 10 back to 5, and then to 0, before full recovery is achieved and the protection action is released. This forms a "5-10-5" hysteresis interval.

[0112] (viii) Edge-triggered emergency response mechanism.

[0113] This embodiment is designed with an edge-triggered emergency response mechanism, which immediately triggers safety control when a sudden change occurs in safety conditions (such as reverse current, overload, excessive demand, or abnormal power factor), without waiting for the next cycle.

[0114] The edge-triggered emergency response mechanism specifically includes: (1) receiving the current cycle power data sent in real time from the gate table; based on the current cycle power data, judging each hazard in the preset emergency check table, the hazard includes at least one of reverse flow, overload, demand exceeding limit and power factor abnormality. (2) obtaining the historical state of each hazard in the previous processing cycle, and comparing the historical state with the current state obtained by the current judgment. (3) for each hazard, executing edge detection logic: ① if a jump from "no hazard" to "hazard" is detected, i.e., a rising edge is identified, an emergency message is immediately generated and sent to wake up the safety control thread to perform immediate processing without waiting for the next periodic safety check moment; ② if a jump from "hazard" to "no hazard" is detected, the corresponding hazard flag is automatically reset; ③ if the hazard state does not change, the existing state is maintained to avoid repeated triggering. Through the edge-triggered emergency response mechanism, the hazard response time is reduced from seconds to about 200 milliseconds.

[0115] (ix) Strategy parameter configuration method.

[0116] The security control method in this embodiment supports multiple operating strategies, and the strategy parameters are configured and persistently stored through a unified strategy structure.

[0117] Upon startup, the device first reads policy parameters from persistent storage. If found, they are applied immediately; otherwise, a set of default parameters is written and saved: the policy mode defaults to "EMS specified," and the demand threshold and anti-backflow threshold default to 1250kW (or, if a gate table is configured on-site, the parameters are taken from the gate table). These parameters are then used for demand checking and anti-backflow checking, respectively.

[0118] (x) A method for parsing the device-PCC correspondence based on JSON configuration files.

[0119] The safety control method in this embodiment is based on a topology mapping of "equipment belonging to PCC": power limits are issued in groups by PCC, demand reduction is weighted and allocated by PCC, and overload protection is checked on a per-unit basis by PCC. All of these rely on the accurate correspondence between each energy storage device (PCS) and gateway table (PCC master table) and its corresponding PCC point (transformer). This embodiment declaratively describes this correspondence using a JSON configuration file and automatically parses and establishes the mapping when the device starts up, adapting to different site electrical topologies without requiring code modification.

[0120] (1) Overall structure of the configuration file.

[0121] Configuration files (such as safe_control.json) consist of two main top-level nodes: { "VER_DESC": "MGCC-20210601-01-es104_para", "devlist": { / * Node 1: Device list and topology relationship * / "DevSN": "317140500062204101000030", "body": [ ...device array... ] }, "t104cfg": { / * Node 2: Telemetry / Telecommunications / Remote Control Point Table * / "dataset": [ ... YC / YK / YX / PARAM arrays... ] } } The parsing process (json_parse_root function) is as follows: First, the "devlist" node is parsed to complete device registration and establish the binding between the device instance and the channel / extended IO; then, the "t104cfg" node is parsed to complete the point table association, binding each telemetry / telecommunication point to its respective device. After the two-stage parsing, the three-layer relationship chain of "device-measuring point-PCC" is fully established.

[0122] (2) Binding mechanism between key fields of equipment list and PCC.

[0123] Each element in the “devlist.body” array describes a field device, and the key fields directly related to the PCC correspondence are shown in Table 1.

[0124] Table 1. Key fields directly related to the PCC correspondence. When the system reads a device element, it first parses two basic fields: device number and collector number (converting from 1-based to 0-based). Then it checks if there is a PCCNo field: if there is, it means that this device is connected to a transformer, and the number is subtracted by 1 to get the internal index; if not, it is marked as 255, indicating "not connected to any transformer". Next, the string model and type are translated into an internal enumeration, and parameters such as rated power are parsed out (125 is used as a fallback by default).

[0125] Then, the processing is divided according to the equipment type: if it is an energy storage (PCS), load_pcscfg is called to record its PCC affiliation into the PCS runtime structure; if it is a meter / load, load_loadcfg is called to load the load structure, and the type is LOAD_PCC will be registered as the monitoring instance of that PCC, and its rated capacity p_rating will be used as the capacity benchmark of the transformer.

[0126] (3) Three types of query paths grouped by PCC at runtime.

[0127] After the mapping is established, the three links of security control are all grouped using pccno as the key. The unified filtering mode for the three paths is as follows: first, check if the target PCC number matches (pccno=-1 means unlimited, all matches), then check if the device is online. Only if both conditions are met will it be included in this statistics or control; otherwise, it will be skipped.

[0128] Path 1: PCC instance discovery (overload prevention, demand weight calculation) - Traverse the load table. Any device of type PCC gate table that is online is a PCC monitoring point: Call pcc_is_ovld to determine overload according to the p_rating of the PCC, and then call handle_power_limit to send the limit value to the device under the pccno.

[0129] Path 2: PCS group statistics and limit issuance by PCC.

[0130] (4) Point table association and measurement point addressing.

[0131] The “t104cfg.dataset” binds telemetry / telecommunication points to devices. Each point is associated with a device via devNo and its semantics are defined via paradesc / name. At runtime, it is addressed by parameter name (PCC1_pApp / PCC1_pAct, etc.), as shown in Table 2. Among them, the apparent power pApp measurement point is configured with over-limit linkage telecommunication. Over-limit information on the meter side is directly sent up through the telecommunication channel, forming a dual verification with the overload detection. When emergency data points are sent up, the edge emergency response mechanism is triggered.

[0132] Table 2. Measurement points and their uses (5) Advantages and features of configuration parsing.

[0133] ① Declarative topology description: The field electrical topology (which equipment is connected to which transformer) is completely described by JSON configuration. The device software is not aware of the specific field. When changing the site, only the configuration file needs to be changed.

[0134] ② Default fault tolerance: p_rating defaults to 125, PCCNo defaults to 255 (no connection), loadtype defaults to LD_TYPE_ERR and an alarm is triggered. A single missing field will not cause parsing crash.

[0135] ③ Automatic conversion between 0-based and 1-based: PCCNo / exioNo are configured as 1-based natural numbers (in accordance with engineering practice), and are automatically converted to 0-based indexes for software use during parsing.

[0136] ④ Dual-channel consistency: The same _104dev_t parsing result is assembled into the uplink (com_up point table) and downlink (com_dn control) configurations respectively, ensuring that monitoring data and control commands act on the same topology view.

[0137] This embodiment declaratively describes the correspondence between energy storage devices, distributed power sources, gateway tables, and PCC points (transformers) using the PCCNo field in a JSON configuration file. This description is automatically parsed and integrated into the runtime structure of each device management module upon device startup. All three aspects of safety control (overload protection, demand allocation, and backflow prevention) are grouped and operated using PCCNo as the key. Complex field topologies with multiple PCCs and PCSs can be adapted simply by modifying the configuration file. Field defaults are fault-tolerant, automatic 0-based / 1-based conversion is implemented, and uplink and downlink dual channels share the same parsing result, ensuring that monitoring and control operate on a consistent topology view and avoiding errors caused by manual configuration of control objects.

[0138] Of course, the above description is not intended to limit the present invention, and the present invention is not limited to the examples given above. Any changes, modifications, additions or substitutions made by those skilled in the art within the scope of the present invention should also fall within the protection scope of the present invention.

Claims

1. A safety control method for energy storage systems to prevent backflow, overload, and demand management, characterized in that, Includes the following steps: S1. Check if the electricity meters at each checkpoint are online; if any checkpoint electricity meter is offline, set the power limit to the minimum value; if all checkpoint electricity meters are online, proceed to step S2. S2. Perform three layers of safety checks in sequence: the first layer is overload prevention check, the second layer is backflow prevention check, and the third layer is demand control check. Among them, the three-layer safety checks are all performed by the PCC point power value after the dynamic load tracker selects the nearest replacement. If the edge triggers the emergency response condition during the execution, it will directly enter the emergency response stage and trigger the safety control. All three safety checks generate power limits through a limit accumulation mechanism: the overload protection check first sets a basic limit based on the transformer's apparent power, and each subsequent check further tightens the limit based on the power limit set in the previous level. No subsequent check may relax the power limit set in the previous level. S3. After the limit is checked and closed, the final power control command is issued to the energy storage converter using a proportional power limit propagation method.

2. The safety control method for preventing backflow, overload, and demand management in an energy storage system according to claim 1, characterized in that, The specific methods for overload protection checks are as follows: S11. Obtain the active power, apparent power, and power factor of the PCC point; S12. Use a dynamic load tracker to smooth the active power, apparent power and power factor obtained in step S11 to eliminate instantaneous fluctuations. S13. Use apparent power instead of active power as the basis for overload judgment, and calculate the overload amount in combination with the power factor; S14. Set dual overload thresholds: warning threshold and hard limit threshold. When the overload exceeds the hard limit threshold, it is determined to be an overload state and overload counting starts, and power adjustment action is generated. The warning threshold is 95% of the rated capacity of the transformer at the PCC point, and the hard limit threshold is 100% of the rated capacity of the transformer at the PCC point.

3. The safety control method for preventing backflow, overload, and demand management in an energy storage system according to claim 1, characterized in that, The specific methods for backflow prevention inspection are as follows: S21. Preset the backflow trigger threshold and backflow release threshold, and satisfy the condition that the backflow trigger threshold < the backflow release threshold; S22. Obtain the active power at point PCC; S23. When the transformer power at the PCC point is less than the reverse current trigger threshold, calculate the energy storage discharge power limit and reduce the discharge power of the energy storage system according to the energy storage discharge power limit. When the transformer power at the PCC point is greater than the reverse current release threshold, the discharge power of the energy storage system is increased within the preset constraint range, and the increased discharge power does not exceed the power limit set by the overload protection layer. When the reverse current trigger threshold is less than or equal to the transformer power at point PCC and less than or equal to the reverse current release threshold, the current discharge power of the energy storage system remains unchanged. S24. The discharge power limit is passed to the subsequent layer through the limit accumulation mechanism and the corresponding discharge power adjustment is performed.

4. The safety control method for preventing backflow, overload, and demand management in an energy storage system according to claim 1, characterized in that, The specific methods for demand control inspection are as follows: S31. Determine the rated capacity, current active power and preset overload threshold of each PCC point transformer, calculate the overload margin of each transformer, and generate the allocation weight and reduction limit of each transformer. S32. Obtain the total reduction required by the system, and use a two-round weighted allocation algorithm to reduce the upper limit of charging power and the upper limit of discharging power of the energy storage devices connected to each PCC point; The first round of allocation is based on the allocation weight generated in step S31 to obtain the initial reduction amount of each transformer, and to determine whether the initial reduction amount of each transformer exceeds the upper limit of the corresponding transformer's reduction. If so, the actual reduction amount of the transformer is limited to the upper limit of the transformer's reduction, and the excess part is summarized as the remaining reduction amount. The second round of allocation is to allocate the remaining reduction amount to transformers that still have capacity that can be reduced after the first round of allocation, and the allocation ratio is proportional to the remaining capacity that can be reduced of the transformer. S33. Demand checking and allocation execution; Get the total active power of all PCC points at present, and get the upper limit of charging direction demand (demond) and the upper limit of discharging direction demand (demond_d) agreed in the contract; The total active power is compared with demond and demond_d respectively, and a safety factor of 0.99 is introduced to tighten the upper limit of demand. If the total active power is within the tightened upper limit of demand, the power limit on each transformer is lifted. If the total active power exceeds the tightened demand limit, the excess is calculated and used as the total reduction. Then, the two-round weighted allocation algorithm in step S32 is called to distribute the reduction task to each transformer for execution.

5. The safety control method for preventing backflow, overload, and demand management in an energy storage system according to claim 4, characterized in that, In step S31, the overload margin is the difference between the transformer's overload threshold and the current apparent power. For transformers with smaller overload margins and larger current active power, the allocation weight is higher. The upper limit for reduction of each transformer is determined based on the difference between the rated capacity and the current active power. Determine if the allocation weight of all transformers is zero. If so, adjust the allocation weight of all transformers to be proportional to their current active power. If not, maintain the allocation weight generated based on the overload margin.

6. The safety control method for preventing backflow, overload, and demand management in an energy storage system according to claim 5, characterized in that, In step S3, the proportional power limit propagation method includes: Obtain the limited power and expected power of all energy storage devices connected to the target transformer, and calculate the total limit value to be imposed on the target transformer. The ratio of the total limit value to the expected power is the power limitation ratio. If the power limit ratio is ≤0.01, then charging and discharging of energy storage devices under the target transformer are not allowed. If the power limitation ratio is ≥1, then the output power of each energy storage device under the target transformer is allowed to reach 110% of the rated power; If 0.01 < power limit ratio < 1, then the upper limit of the power limit is calculated based on the power limit ratio and the rated capacity of the energy storage, and is sent to each energy storage device connected to the target transformer. Each energy storage device calculates the actual allowable output power based on its own rated power and the received power limit.

7. The safety control method for preventing backflow, overload, and demand management in an energy storage system according to claim 6, characterized in that, Each PCC point corresponds to a dynamic load tracker. The dynamic load tracker maintains a fixed-length circular buffer, which caches the power data of the most recently sampled PCC points. It has logic for exponential weighted regression, nearest replacement, 0.1kW deduplication, and 70kW cold start degradation. It also extracts trend prediction information and statistical average information from the same cached data. The timing of cached data updates is decoupled from the query.

8. The safety control method for preventing backflow, overload, and demand management in an energy storage system according to claim 1, characterized in that, A reactive power compensation mechanism is introduced, which uses the reactive power of the energy storage converter to perform closed-loop regulation of the power factor at the PCC point, thereby reducing apparent power and indirectly alleviating overload; specifically: Collect the active and reactive power of the PCC point and the active and reactive power of the energy storage converter, and calculate the current power factor; The system obtains the user-configured reactive power compensation mode, which includes a target reactive power mode and a target power factor mode. If it is in the target reactive power mode, the reactive power deficit is calculated based on the difference between the configured target reactive power and the current reactive power. If it is in the target power factor mode, the system calculates the corresponding reactive power boundary based on the configured target power factor and compares the absolute value of the current power factor with the absolute value of the target power factor. If the current power factor is better than the target power factor, the reactive power of the energy storage converter is gradually recovered, and the recovered amount is used as the reactive power deficit to prevent overcompensation. If the current power factor is worse than the target power factor, the difference between the reactive power required to reach the target power factor and the current reactive power is calculated and used as the reactive power deficit. Reactive power adjustment commands are generated based on the reactive power deficit and sent to the energy storage converter for execution. The actual reactive power at the PCC point is used as the controlled variable, and the reactive power deficit is used as the control deviation. By adjusting the reactive power output of the energy storage converter, the reactive power at the PCC point is brought to converge within the target value range, thus forming a reactive power compensation closed-loop control.

9. The safety control method for preventing backflow, overload, and demand management in an energy storage system according to claim 8, characterized in that, The reactive power Ql of each energy storage converter is jointly constrained by the rated capacity Se of the energy storage converter and the current active power Pl: Pl 2 + Ql 2 ≤Se 2 When multiple energy storage converters are connected in parallel, 80% of the sum of the reactive power of each energy storage converter, Qlmax, is the upper limit of the system's reactive power, and the lower limit of the system's reactive power is -80%Qlmax. Furthermore, the sum of the square of the active power and the square of the reactive power of each energy storage converter is not greater than the square of its rated capacity. Obtain the current reactive power of the energy storage converter and calculate the reactive power deficit that needs to be compensated. Calculate the initial reactive power target value based on the formula: Initial reactive power target = Current reactive power + Reactive power deficit. The initial reactive power target value is compared with ±80%Qlmax, and a limiting process is performed: if the initial reactive power target value is greater than 80%Qlmax, it is clamped to 80%Qlmax, and the undigested reactive power deficit is recorded as the difference between the initial reactive power target value and 80%Qlmax; if the initial reactive power target value is less than -80%Qlmax, it is clamped to -80%Qlmax, and the undigested reactive power deficit is recorded as the difference between the initial reactive power target value and -80%Qlmax; if the initial reactive power target value is between 80%Qlmax and -80%Qlmax, it is determined that the reactive power deficit has been completely digested, and the undigested reactive power deficit is cleared to zero. The initial reactive power target value after the limiting process is issued as the final reactive power command to the energy storage converter for execution, and the undigested reactive power deficit is returned to the upper-level system for decision-making.

10. The safety control method for preventing backflow, overload, and demand management in an energy storage system according to claim 9, characterized in that, An error counter hysteresis mechanism is introduced, which means that the protection action is triggered when a preset number of abnormal situations occur consecutively, and the protection action is released only after a preset number of recovery signals have occurred consecutively.

Images