A power distribution network energy storage optimal configuration method for distributed power source carrying capacity improvement

Abstract

The present application relates to power distribution network power flow regulation and energy storage configuration technical field, disclose a kind of power distribution network energy storage optimization configuration method for distributed power load bearing capacity improvement, comprising: collecting power distribution network operating state data, node voltage amplitude partial derivative to injection power is solved in distributed energy management system to construct voltage power sensitivity matrix;Monitoring grid-connected power flow distribution characteristics, matching energy storage converter output characteristics;Active regulation instruction is input into feeder loop, time-varying equivalent resistance characteristic quantity is obtained by solving, and the voltage variation nominal deviation amount caused by transient temperature drift of equivalent impedance is output, and the transient voltage stability convergence criterion is output, the present application constructs impedance feedforward hedging mechanism, realizes control depth and physical heat dissipation endogenous cooperation, improves power distribution network dynamic bearing capacity.

Description

Technical Field

[0001] This invention relates to a method for optimizing the configuration of energy storage in distribution networks to enhance the carrying capacity of distributed power sources, belonging to the field of power flow regulation and energy storage configuration technology in distribution networks. Background Technology

[0002] Currently, with the large-scale integration of distributed power sources into the smart grid, the distribution network structure is evolving towards an active distribution network. The mainstream approach to improving power flow regulation capability is to use a distributed energy management system to coordinate and schedule multi-point distributed energy storage systems and regulate the active and reactive power output of energy storage converters. This approach has clear engineering value in smoothing out random power fluctuations and ensuring power quality.

[0003] With the increasing penetration rate of new energy sources, frequent high-frequency and high-power throughput causes severe transient thermal effects in converter arms and local lines. This heat dissipation causes nonlinear temperature drift in the parasitic resistance of the lines, resulting in millisecond-level dynamic distortion of the equivalent impedance and power flow sensitivity at the grid connection point. Due to the time lag characteristic of heat dissipation, the currently sampled voltage gradient is deeply superimposed with the residual electrothermal effect left by the previous action, leading to control phase misalignment. When facing non-ideal network conditions such as random communication delays or packet loss, the transient phase difference caused by this thermal inertia hysteresis can cause false transient reversals in the locally sampled voltage characteristics, resulting in oversaturation of the control loop feedback gain. This induces force direction mismatch and high-frequency control oscillations in group regulation, causing accelerated aging of battery arrays and fundamental degradation of the dynamic carrying capacity of the distribution network. For this kind of oscillation problem caused by environmental dynamic distortion, if a linear improvement method with increased energy storage capacity is adopted, it will not only fail to eliminate the physical limitations of line impedance drift, but will also generate... More severe transient heat dissipation exacerbates loop divergence and leads to increased construction costs. If the approach of using external high-frequency impedance identification hardware is adopted, its dynamic convergence is uncertain in complex high-frequency interference environments. In software control algorithms, Chinese invention patent application CN113904353A discloses a distributed energy storage adaptive droop control method based on voltage sensitivity matrix. It predicts the node voltage after the droop control takes effect by constructing a closed-loop model that includes the power system response and adjusts the droop coefficient accordingly. However, the underlying preset relies on the static constant characteristics of the distribution network impedance. The sensitivity matrix update is difficult to map the transient temperature drift of the local feeder impedance caused by high-frequency throughput of energy storage, which severs the causal relationship between control command flow and physical heat dissipation. Under the condition of perception delay caused by wide-area non-ideal communication environment, due to the lack of causal memory of the historical flow characteristics of previous continuous adjustment commands, the transient phase difference caused by thermal inertia hysteresis cannot be offset, which easily leads to power flow resonance and control loop divergence.

[0004] Therefore, the technical problem to be solved by this invention is how to utilize the local data flow and logic operator reorganization of the controller to accurately sense and compensate for the impedance thermal inertia hysteresis effect induced by the control action while maintaining the physical structure topology of the grid, eliminate the voltage gradient pseudo-reversal noise caused by the superposition of signal transmission delay and temperature drift, and maintain the control convergence and transient absolute stability of the multi-point energy storage system. Summary of the Invention

[0005] To address the problems in the background art, the technical solution of the present invention is as follows: A method for optimizing the configuration of energy storage in distribution networks to enhance the carrying capacity of distributed power sources, comprising the following steps:

[0006] Step S1: Collect real-time network source operation status data of the active distribution network online, calculate the partial derivatives of the voltage amplitude of each grid node with respect to the injected active power and the injected reactive power in the distributed energy management system, so as to assemble and construct a multi-dimensional voltage-power sensitivity matrix and quantitatively establish the control correlation network between the dynamic response of the distribution network node voltage and the reactive four-quadrant balance of the multi-point energy storage system.

[0007] Step S2: Dynamically monitor the spatiotemporal distribution characteristics of reverse power flow caused by distributed power grid connection, use the multidimensional voltage-power sensitivity matrix to match the active and reactive power output characteristics of the energy storage converter online, and adaptively adjust the power transmission and distribution path of each feeder to reduce the power flow loss of the entire distribution network.

[0008] Step S3: The active power regulation command time sequence continuously cyclically generated within the pre-control cycle is input into the feeder temperature drift impedance time-varying feedforward regulation loop. The feeder time-varying equivalent resistance characteristic quantity, which characterizes the transient temperature drift of the feeder impedance caused by the conduction current heating, is analyzed and calculated. The current converter loop damping factor is adjusted through the feeder time-varying equivalent resistance characteristic quantity to constrain the dynamic control gain of the energy storage converter transient regulation loop, offset the nominal deviation of voltage variation caused by the transient temperature drift of the feeder equivalent impedance, generate a transient voltage stability convergence benchmark, smooth out node voltage over-limit and suppress energy storage transient oscillation.

[0009] Preferably, step S1 includes the following sub-steps: Step S11, online acquisition of the injected active power change, injected reactive power change, and node voltage amplitude change of each node in the active distribution network within a determined control period; Step S12, establishing the node power time-varying balance topology equation within the distributed energy management system, simultaneously solving the time-varying transmission sensitivity of the injected active power change and injected reactive power change with the node voltage amplitude change, and assembling a multidimensional voltage-power sensitivity matrix.

[0010] Preferably, step S2 includes the following sub-steps: Step S21, based on the multidimensional voltage-power sensitivity matrix, quantitatively calculate the limit constraint conditions for local feeder overheating caused by active and reactive power backfeed, and identify the voltage over-limit risk interval in the spatiotemporal distribution characteristics of reverse power flow caused by distributed power grid connection; Step S22, for the voltage over-limit risk interval, allocate the capacity ratio of active power and reactive power output by the energy storage converter in the four quadrants, limit local feeder overheating and balance the power flow distribution of the entire network.

[0011] Preferably, step S3 includes the following sub-steps: step S31, continuously sampling the active power regulation command time series of multiple historical time nodes according to level, and constructing the pre-cycle command regulation characteristic time series; step S32, inputting the pre-cycle command regulation characteristic time series into the first-order equivalent heat loss dissipation conduction model of the distribution network for impedance transient temperature drift calculation, and outputting the feeder time-varying equivalent resistance characteristic quantity.

[0012] Preferably, after adjusting the current converter loop damping factor through the feeder time-varying equivalent resistance characteristic in step S3 to constrain the dynamic control gain of the transient regulation loop of the energy storage converter, the following sub-steps are also included: Step S33, collecting the node voltage step change rate signal of each feeder, and extracting the voltage variation polarity indicator representing the direction of the node voltage step through the sign function operator; Step S34, performing nonlinear product mapping between the voltage variation polarity indicator and the converter loop damping factor to generate a reverse phase offset compensation amount to eliminate the voltage variation phase deviation, filtering out the in-phase sudden step disturbance from the transient regulation loop of the energy storage converter, and outputting the transient voltage stability convergence benchmark.

[0013] Preferably, the method further includes the following steps: Step S4, establishing a basic database of full-cycle multi-dimensional operation history evolution, and continuously online statistically analyzing the dynamic response time series of active and reactive power of the multi-point energy storage system under extremely high frequency output conditions, the frequency of charge and discharge cycle switching, and the cumulative time shift of the feeder time-varying equivalent resistance characteristic quantity.

[0014] Preferably, the method further includes the following steps: Step S5, extracting the rate of change and steady-state deviation of the feeder time-varying equivalent resistance characteristic from the full-cycle multi-dimensional operation history evolution database, and combining the charge-discharge cycle switching frequency to calculate the health degradation index characterizing the overall aging degree of the multi-point energy storage system, and quantitatively presenting the performance evolution state of the energy storage array during long-term operation.

[0015] Preferably, the method further includes the following steps: Step S6, when the health decay index is greater than or equal to the determined safe life threshold, automatically reduce the upper limit of the output of the current active power adjustment command sequence, and proportionally increase the adjustment weight of reactive power in the four-quadrant balance, thereby delaying the aging of the energy storage array through reactive power priority adjustment control.

[0016] Preferably, the method further includes the following steps: Step S7, under the condition that the sensing delay is caused by wide-area non-ideal communication in the active distribution network, the transient voltage stability convergence benchmark provides an adaptive transient voltage constraint boundary with a time-series causal relationship for the transient active power command, and dynamically limits the variation range of the transient voltage stability convergence benchmark, so as to maintain the transient stability control capability of the multi-point energy storage system in the wide-area asynchronous information environment.

[0017] Compared with the prior art, the beneficial effects of the present invention are:

[0018] 1. In the optimized configuration of energy storage in distribution networks, a dynamic virtual impedance mechanism is injected into the energy storage coordination control loop to capture the thermal inertia hysteresis effect of the equivalent impedance of the grid induced by continuous adjustment actions in real time. This transforms the second-level heat dissipation evolution trajectory and the microsecond-level electrical control flow into a cross-dimensional endogenous collaborative control system, correcting the impedance drift caused by temperature rise. This path overcomes the decoupling limitation of treating network impedance as a static constant in distribution network control, avoids local power flow oscillations in the distribution network caused by disordered output of multi-point distributed energy storage, and ensures that the electrical damping characteristics of the entire network are always within a safe range when high-penetration distributed power sources are connected, effectively improving the dynamic carrying capacity of the distribution network.

[0019] 2. This invention constructs a nonlinear strong cooperative control topology with temporal causal memory and introduces an electrothermal memory self-correction mechanism. The topology collects the historical flow characteristics of the continuously evolving active power commands in the pre-step process and transforms them into state variables that constrain the current grid damping weight. Compared with traditional methods that rely solely on transient electrical indicators, the topology can provide an accurate phase convergence benchmark for the energy storage control loop under high-frequency operating conditions, solve the problem of wide-area state perception lag caused by non-ideal communication conditions, eliminate the interference of large fluctuations in power flow in the distribution network on control convergence, and maintain the absolute stability of the transient process of the multi-point distributed energy storage system.

[0020] 3. This invention relies on the nonlinear mapping relationship between the dynamic thermal memory damping coefficient and the sign function to filter and remove voltage step phase misalignment noise caused by transient thermal drift of the equivalent conductance of the local area network. Without changing the original physical structure and topology of the network, the mapping relationship is used to avoid control signal malfunction when multiple distributed power sources feed back power, accurately locate the voltage over-limit risk of weak nodes in the network, and effectively smooth out abnormal voltage steps at nodes. This eliminates the need for the energy storage system to frequently cross safety boundaries to perform unnecessary power compensation charging and discharging operations, thereby delaying the aging of the energy storage array and ensuring the safety benefits of long-term operation of the smart grid. Attached Figure Description

[0021] Figure 1 This is a flowchart illustrating the execution logic of the distribution network energy storage optimization configuration method for improving the carrying capacity of distributed power sources, as described in this invention.

[0022] Figure 2 This diagram illustrates the technical dimensions of the distribution network energy storage optimization configuration method for improving the carrying capacity of distributed power sources, as described in this invention.

[0023] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0024] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0025] A method for optimizing the configuration of energy storage in distribution networks to enhance the carrying capacity of distributed power sources includes the following steps:

[0026] Step S1: Collect real-time network source operation status data of the active distribution network online, calculate the partial derivatives of the voltage amplitude of each grid node with respect to the injected active power and the injected reactive power in the distributed energy management system, so as to assemble and construct a multi-dimensional voltage-power sensitivity matrix and quantitatively establish the control correlation network between the dynamic response of the distribution network node voltage and the reactive four-quadrant balance of the multi-point energy storage system.

[0027] Step S2: Dynamically monitor the spatiotemporal distribution characteristics of reverse power flow caused by distributed power grid connection, use the multidimensional voltage-power sensitivity matrix to match the active and reactive power output characteristics of the energy storage converter online, and adaptively adjust the power transmission and distribution path of each feeder to reduce the power flow loss of the entire distribution network.

[0028] Step S3: The active power regulation command time sequence continuously cyclically generated within the pre-control cycle is input into the feeder temperature drift impedance time-varying feedforward regulation loop. The feeder time-varying equivalent resistance characteristic quantity, which characterizes the transient temperature drift of the feeder impedance caused by the conduction current heating, is analyzed and calculated. The current converter loop damping factor is adjusted through the feeder time-varying equivalent resistance characteristic quantity to constrain the dynamic control gain of the energy storage converter transient regulation loop, offset the nominal deviation of voltage variation caused by the transient temperature drift of the feeder equivalent impedance, generate a transient voltage stability convergence benchmark, smooth out node voltage over-limit and suppress energy storage transient oscillation.

[0029] Preferably, step S1 includes the following sub-steps: Step S11, online acquisition of the injected active power change, injected reactive power change, and node voltage amplitude change of each node in the active distribution network within a determined control period; Step S12, establishing the node power time-varying balance topology equation within the distributed energy management system, simultaneously solving the time-varying transmission sensitivity of the injected active power change and injected reactive power change with the node voltage amplitude change, and assembling a multidimensional voltage-power sensitivity matrix.

[0030] Preferably, step S2 includes the following sub-steps: Step S21, based on the multidimensional voltage-power sensitivity matrix, quantitatively calculate the limit constraint conditions for local feeder overheating caused by active and reactive power backfeed, and identify the voltage over-limit risk interval in the spatiotemporal distribution characteristics of reverse power flow caused by distributed power grid connection; Step S22, for the voltage over-limit risk interval, allocate the capacity ratio of active power and reactive power output by the energy storage converter in the four quadrants, limit local feeder overheating and balance the power flow distribution of the entire network.

[0031] Preferably, step S3 includes the following sub-steps: step S31, continuously sampling the active power regulation command time series of multiple historical time nodes according to level, and constructing the pre-cycle command regulation characteristic time series; step S32, inputting the pre-cycle command regulation characteristic time series into the first-order equivalent heat loss dissipation conduction model of the distribution network for impedance transient temperature drift calculation, and outputting the feeder time-varying equivalent resistance characteristic quantity.

[0032] Preferably, after adjusting the current converter loop damping factor through the feeder time-varying equivalent resistance characteristic in step S3 to constrain the dynamic control gain of the transient regulation loop of the energy storage converter, the following sub-steps are also included: Step S33, collecting the node voltage step change rate signal of each feeder, and extracting the voltage variation polarity indicator representing the direction of the node voltage step through the sign function operator; Step S34, performing nonlinear product mapping between the voltage variation polarity indicator and the converter loop damping factor to generate a reverse phase offset compensation amount to eliminate the voltage variation phase deviation, filtering out the in-phase sudden step disturbance from the transient regulation loop of the energy storage converter, and outputting the transient voltage stability convergence benchmark.

[0033] Preferably, the method further includes the following steps: Step S4, establishing a basic database of full-cycle multi-dimensional operation history evolution, and continuously online statistically analyzing the dynamic response time series of active and reactive power of the multi-point energy storage system under extremely high frequency output conditions, the frequency of charge and discharge cycle switching, and the cumulative time shift of the feeder time-varying equivalent resistance characteristic quantity.

[0034] Preferably, the method further includes the following steps: Step S5, extracting the rate of change and steady-state deviation of the feeder time-varying equivalent resistance characteristic from the full-cycle multi-dimensional operation history evolution database, and combining the charge-discharge cycle switching frequency to calculate the health degradation index characterizing the overall aging degree of the multi-point energy storage system, and quantitatively presenting the performance evolution state of the energy storage array during long-term operation.

[0035] Preferably, the method further includes the following steps: Step S6, when the health decay index is greater than or equal to the determined safe life threshold, automatically reduce the upper limit of the output of the current active power adjustment command sequence, and proportionally increase the adjustment weight of reactive power in the four-quadrant balance, thereby delaying the aging of the energy storage array through reactive power priority adjustment control.

[0036] Preferably, the method further includes the following steps: Step S7, under the condition that the sensing delay is caused by wide-area non-ideal communication in the active distribution network, the transient voltage stability convergence benchmark provides an adaptive transient voltage constraint boundary with a time-series causal relationship for the transient active power command, and dynamically limits the variation range of the transient voltage stability convergence benchmark, so as to maintain the transient stability control capability of the multi-point energy storage system in the wide-area asynchronous information environment.

[0037] Example 1: When the system faces the spatiotemporal fluctuations in reverse power flow caused by high-penetration distributed photovoltaic grid connection, in the 10kV active distribution network operating environment of digital evolution, the total grid-connected capacity of distributed photovoltaic reaches 8.5MW and the new energy penetration rate is 85%. The 2MW / 4MWh distributed energy storage system connected to 3 power flow sensitive nodes is frequently in a state of microsecond-level and high-density power throughput, which induces transient temperature drift of impedance in the converter bridge arm and local feeder lines, causing millisecond-level nonlinear changes in the equivalent dynamic conductance at the grid connection point. Under network conditions of 120ms transient random delay and packet loss in wired communication networks, the residual effect of electrothermal history with thermal inertia hysteresis characteristics and the lagging electrical response cause phase misalignment, resulting in pseudo-transient flips in the locally sampled voltage characteristics. Traditional centralized optimization scheduling methods, due to computational convergence lag, lead to oversaturation of controller feedback gain, resulting in output direction mismatch and high-frequency control oscillations. Ultimately, this leads to accelerated aging of the battery array and degradation of the dynamic carrying capacity of the distribution network. To address the apparent contradiction between the aforementioned millisecond-level dynamic distortion and the overall electrothermal conduction timescale, this electrothermal meter... The transition from layer to whole is achieved through the superposition of the skin effect caused by high-frequency alternating current and the local micro-region thermal resistance. Under the transient power surges of high density and frequent throughput, the conduction current is mainly concentrated in the skin depth layer within 20 micrometers of the surface of the local feeder conductor and in the micrometer-level contact point micro-region of the physical contact surface of the energy storage converter. This current accumulation at the surface level causes a sudden surge in the local current density. Within a millisecond timescale, the resulting transient high Joule heat is constrained by the surface thermal resistance of the metal conductor and cannot be immediately dissipated to the overall bulk heat capacity at the center of the conductor. The extremely high transient temperature rise rate at the millisecond level is induced in the surface layer and contact micro-region of the conductor, directly causing millisecond-level nonlinear changes in the equivalent dynamic conductance at the grid connection point; while the overall volume temperature of the conductor accumulates slowly on the scale of seconds or even minutes. The superposition of these two across scales constitutes the transient temperature drift and thermal inertia hysteresis characteristics of the impedance. The main control chip of the general distributed energy management system and the digital signal processor of the energy storage converter run the method of this invention in real time. The processor collects the three-phase power frequency AC voltage at the grid connection point in real time with a sampling period of 10μs, and outputs the fundamental positive sequence voltage amplitude through the positive sequence component extraction operator. And according to the preset reference voltage nominal value Calculate the voltage deviation at the current moment. The timing differential operator with a differential step size of 2ms is used to calculate the voltage timing deviation rate of the current control node. Simultaneously, the energy storage throughput status bit tags of adjacent electrical nodes are obtained through the low-frequency carrier communication interface of the distribution network. The voltage timing deviation variation rate With feature labels Input a preset discrete arbitration matrix operator, output a globally consistent coordination scalar. The digital signal processor inputs this scalar into the built-in second-order dynamic inertial time-delay operator, smooths the scalar in the time domain, and calls the power flow sensitivity rules of the distribution network nodes to decouple and calculate the initial target command of the active power of the energy storage system at the current moment. With reactive power initial target command Specifically, after decoupling and calculating the initial target commands for active and reactive power, the system dynamically reconstructs the output ratio of the multi-point distributed energy storage system to physically achieve adaptive adjustment of the power transmission and distribution paths of each feeder. The main control chip diverts power flow between multiple parallel local feeders based on the sensitivity coefficient of each node voltage to injected power in the multi-dimensional voltage and power sensitivity matrix. When it is identified that a local feeder has reverse power flow caused by distributed power grid connection and has the risk of overheating, the control algorithm proportionally reduces the charging power of the energy storage converter connected to the front end of the overheated feeder and proportionally increases the charging throughput of the energy storage converter at the end of the adjacent redundant feeder with thermal margin. Through the output peak-shifting coordination and power throughput transfer of the energy storage array, the power transmission and distribution path of the entire network is physically changed while maintaining the original physical topology of the distribution network, thereby adaptively reducing the overall power flow loss of the entire network.

[0038] During the dynamic offsetting of operational costs and capacity truncation steps before the final control command is issued, the digital signal processor acquires the current state of charge of the energy storage battery array. Operating temperature at the highest point inside the array The heat dissipation resistance compensation parameters under the current operating conditions are calculated using the conductivity temperature drift trajectory of the line parasitic resistance. and at operating temperature When the temperature exceeds the preset normal operating boundary of 45°C and continues to accumulate, the dynamic capacity cutoff rule is triggered, and the cutoff coefficient is adjusted. Switching to 0.3, to counteract the control phase misalignment noise caused by impedance thermal inertia hysteresis, the processor uses a built-in discretization time series operator to read the initial active power target instructions generated in the preceding 20 consecutive sampling periods from the memory register. It then calculates the square-weighted discrete sum of the initial active power target instructions in the preceding 20 consecutive sampling periods using a square convolution operator, outputting a dynamic thermal inertia state variable that characterizes the residual heat accumulation caused by continuous current scouring of the parasitic impedance of the local area network frame at the current grid connection point. Based on the numerical range of the dynamic thermal inertia state variable, the preset adaptive self-locking damping gain constant is corrected in situ in real time, outputting the dynamic thermal memory damping coefficient. When the dynamic thermal inertia state variable does not exceed the preset safe thermal power consumption threshold, the dynamic thermal memory damping coefficient... Basic nominal value When the dynamic thermal inertia state variable crosses the safe thermal power consumption threshold and the voltage timing deviation rate of change at the current moment... When the absolute value of [value] shows a continuous increasing trend, the processor will dynamically thermally remember the damping coefficient. Switch to high damping step value However, when the initial target active power command is affected by cross-flow interference in the distribution network, With voltage timing deviation variation rate When the product of and is negative and the dynamic thermal inertia state variable resides at the preset high-level overload boundary, the processor will use the dynamic thermal memory damping coefficient. Forced to converge at a preset lower bound value In practice, the first-order equivalent heat loss dissipation and conduction model of the distribution network is implemented in the digital signal processor through discrete difference equations. The input data of this model is the aforementioned active power initial target command time sequence. Its dynamic processing is as follows: the main control chip uses Joule's law to calculate the transient heat generated by the current flowing through the feeder line in the current control cycle, and deducts the heat dissipation of the feeder to the external environment, thereby dynamically iteratively updating and outputting the current feeder time-varying equivalent resistance characteristic. Specifically, the feeder time-varying equivalent resistance characteristic at the current moment is equal to... The equivalent resistance characteristic value of the previous moment is added to the product of the square of the initial target active power command of the current moment and the preset thermal resistance coefficient, and then the discrete correction value of the ratio of the equivalent resistance characteristic value of the previous moment to the preset heat dissipation time constant is subtracted; where the thermal resistance coefficient is determined by dynamic calibration based on the resistivity of the feeder metal conductor at a reference temperature of 25℃, and the heat dissipation time constant is set to 15 seconds; through differential iteration of this model, the nominal value of the transient temperature drift of the impedance caused by the heating of the feeder due to the conduction current is continuously output and transmitted to the adaptive correction program of the converter loop damping factor.

[0039] The digital signal processor will calculate the dynamic thermal memory damping coefficient. Substituting the instructions into the instruction reconstruction equation, combined with the initial target instruction for active power... Voltage timing deviation variation rate Cutoff coefficient and heat dissipation resistance compensation parameters The nonlinear coupling calculation outputs the final active power control command, and its specific algebraic expression formula is as follows: ,in, This is the final active power control command. This is the initial target command for active power. This refers to the truncation coefficient corresponding to the capacity dynamic truncation rule. These are parameters for compensating for heat dissipation resistance. This represents the voltage timing deviation rate of variation at the current control node. The dynamic thermal memory damping coefficient, It is a standard symbolic function, and the subscript is... For electrical node index, superscript The initial state identifier is indicated by the superscript, and the final reconstructed state identifier is indicated by the subscript. This is a dynamic adaptive self-locking regulation indicator; when the energy storage system is charging and the local voltage is rising, or when the energy storage system is discharging and the local voltage is falling, the output value of the standard sign function is positive, and the dynamic virtual impedance damping term in the denominator, composed of the adaptive self-locking damping gain constant and the absolute value of the voltage timing deviation rate of change, takes effect, actively lowering the final active power control command. The amplitude of the converter feedback gain is reduced to mitigate oversaturation caused by transient impedance temperature drift. However, when the implicit interference from cross-current power flow in the distribution network causes a false transient reversal in the direction of the locally acquired voltage gradient, the initial target command is affected. With voltage timing deviation variation rate When the product is negative, the output of the standard sign function is negative, and the damping term in the denominator is transformed into a reverse compensation factor to maintain the convergence stability of the control drive software under power flow noise. Finally, the processor calculates the final active power control command. The signal is converted into a pulse width modulation signal and drives the gate trigger circuit of the power semiconductor switch in the energy storage converter to physically regulate the power flow of the distribution network. Before running the multiplication and division operators of the active power control command reconstruction equation, the digital signal processor starts the denominator safety monitoring and protection program. The algebraic calculation value of the dynamic virtual impedance damping denominator term, which is composed of the adaptive self-locking damping gain constant, the absolute value of the voltage timing deviation rate of change, and the standard sign function, is transmitted to the limiting register. It is compared in real time with the preset lower limit absolute value dead zone threshold of 0.2. If the dynamic power flow is affected by the cross-current interference of the distribution network, the protection program will be activated. When the absolute value of the denominator term of the virtual impedance damping is less than 0.2, the limiting register automatically triggers logic self-locking and rigidly limits the denominator value participating in the division operation to a fixed value of 0.2. This eliminates the divergence of active power control commands and system closed-loop instability caused by the denominator approaching zero within the arithmetic logic unit of the control chip. In actual operation, to avoid the denominator term of the command reconstruction equation approaching zero or even generating a mathematical singularity due to negative term offsetting when the product of the initial active power target command and the voltage timing deviation rate is negative and the standard sign function output is negative, the underlying logic of the control software... The system incorporates adaptive limiting self-locking and extreme value lock-up protection. Specifically, before executing multiplication and division operators, the digital signal processor monitors in real time the calculation result of the dynamic virtual impedance damping term in the denominator, which consists of the adaptive self-locking damping gain constant and the absolute value of the voltage timing deviation rate of change. It also forcibly sets the minimum absolute value dead zone threshold of the overall denominator value to be no less than 0.2. If strong interference from cross-current power flow in the distribution network causes the algebraic calculation value of the denominator to fall below this dead zone threshold, the processor will directly trigger logic self-locking, rigidly limiting and forcibly locking the input value of the denominator to a fixed value. The value of 0.2 completely eliminates the physical risks of overshooting and infinite divergence in the final active power control command caused by the denominator approaching zero, ensuring the transient absolute convergence of the control drive software and the closed-loop stability of the system under extreme power flow noise interference. Under the dynamic regulation effect of the feedforward defense mechanism, the phase misalignment noise caused by the heat accumulation of the preceding continuous power output action and the current electrical response is stripped online. The voltage gradient pseudo-reversal interference under the intermittent delay environment of the wireless communication link is filtered out, and the fundamental positive sequence voltage amplitude of each power flow sensitive node is reduced. After 24 hours of continuous operation testing, the system remained stable within the safe operating window. The upper limit of the dynamic carrying capacity of distributed power sources across the entire network increased from 6.2MW to 8.2MW, representing an improvement of 32.25%. At the same time, the cumulative duration of node voltage over-limit was reduced to zero, and the highest daily operating temperature of individual energy storage battery array cells decreased from 51.2℃ to 41.5℃, avoiding array aging caused by frequent nonlinear overcharging and over-discharging. Ultimately, the overall power flow loss rate of the entire network decreased from 4.8% to 3.1%, and the dynamic carrying capacity boundary of the distribution network changed from qualitative smoothing to quantitative control.

[0040] Example 2: When the system faces high-frequency alternating disturbances caused by reverse power flow from distributed sources, a grid hardware-in-the-loop simulation test bench serves as the experimental platform. This bench includes a digital full-physical simulation converter frame with a rated voltage of 10kV and microsecond-level real-time calculation capabilities. This frame is used to construct a semi-physical feedback loop for verifying the active and reactive power flow control of the smart grid. Regarding data acquisition, the experiment collects actual active and reactive power time-series data output by a real distributed solar grid-connected inverter during 24 hours of continuous operation. Simultaneously, to reconstruct the complex electromagnetic interference environment within the distribution network and simulate predetermined industrial measurement and control boundaries, Gaussian white noise with a signal-to-noise ratio of 20dB and 50Hz low-order power frequency interference harmonics sampled by the power quality analysis unit are actively superimposed onto the active and reactive power time-series data as test benchmarks. Before the test begins… The processor applies a numerical determination to the historical sampling window length involved in the preceding control loop. The historical sampling window length is jointly constrained by the heat dissipation rate of the local feeder line and the operation cycle of the converter digital signal processor. Its setting is used to balance the transient thermal memory tracking accuracy of the active power control command and the data processing energy consumption of the main control chip. When the high-frequency fluctuation of active power causes the absolute value of the timing variation rate of the equivalent dynamic conductance of the local line to be greater than 0.05 Siemens per second, in order to avoid tracking aliasing due to thermal inertia hysteresis under the Nyquist sampling limit, the historical sampling window length tends to the lower limit of its operating range. Under this operating condition, by substituting the voltage timing deviation variation rate and the reverse power flow fluctuation frequency into the constraint balance relationship, the historical sampling window length is calculated to be 20 sampling cycles, thereby establishing the initial operating state benchmark of the test group under high noise injection.

[0041] During the active power control command issuance phase, four parallel comparison processes were set up. The present invention sample group employed control logic including a complete command reconstruction equation and a dynamic thermal memory damping coefficient. The control group used a traditional control method based on a static fixed impedance constant. For the partially missing control group, the dynamic thermal memory damping coefficient was forcibly set to a constant of zero in the command reconstruction equation to eliminate time-series causal memory terms. The out-of-range control group set the cutoff coefficient corresponding to the capacity dynamic cutoff rule to 0.85, exceeding the reasonable window boundary. When the test bench injected an active power adjustment command into the feeder loop due to a drastic jump in photovoltaic output, different intermediate data evolution states were generated in the local cache of the digital signal processors of each group. The relaxation time of the present invention sample group after pulse excitation release remained at 12μs, and its differential operator with a differential step size of 2ms calculated the voltage timing deviation variation rate in real time. The corresponding global consistency coordination scalar is 2.53V per second. The temperature stabilized at 0.72. Due to the accumulation of thermal resistance in the local area network structure caused by the preceding continuous power surge, the highest operating temperature of the battery array rose from 35℃ to 42.3℃. The processor internally calculated that the dynamic thermal inertia state variable exceeded the preset safe thermal power consumption threshold, triggering the in-situ real-time correction of the adaptive self-locking damping gain constant, thus improving the dynamic thermal memory damping coefficient. The voltage spontaneously jumps from the base nominal value of 1.24 to the high-damping step value of 3.50, exhibiting a self-locking suppression state. Conversely, when the cutoff coefficient of the out-of-range control group deviates from the reasonable operating window due to over-setting, the damping adjustment term in the denominator saturates due to the overload effect, resulting in a deep overshoot in the final active power control command and triggering the overcurrent protection program. This is consistent with the physical law of over-temperature degradation caused by inverter arm thermal instability. Furthermore, due to the lack of thermal memory offsetting causal chain, the coordinated scalar of the partially missing control group exhibits high-frequency phase misalignment, affecting the amplitude of the locally sampled fundamental positive sequence voltage. It oscillates violently between 10.45V and 11.68V and cannot maintain control loop convergence stability under a network latency of 120ms.

[0042] When the test continued for a 24-hour control cycle and long-term stability tracking of the final control output was performed, the experimental group and the control groups showed performance gradient differences in the dimension of distributed power supply capacity improvement. In the control group that did not use the method of this invention, when the system experienced reverse power flow overload due to local impedance thermal drift at the 4.2-hour mark, its upper limit of distributed power supply capacity improvement remained at 6.21MW, and the cumulative duration of node voltage over-limit reached 145.6 minutes. In contrast, the experimental group of this invention, employing a complete feedforward defense mechanism, demonstrated an active offsetting capability against environmental physical characteristic distortions and calculated the final active power control command. Maintaining within the safe control gain range, its fundamental positive sequence voltage amplitude The maximum fluctuation deviation is less than 3.5% of the preset safety threshold. The upper limit of the dynamic carrying capacity of the distributed power supply in the whole network has been stably expanded to 8.24MW. The cumulative duration of voltage over-limit has been reduced to zero. Due to the elimination of the output direction mismatch oscillation induced by the control phase misalignment, the highest daily operating temperature of the single cell of the energy storage battery array has decreased from 51.2℃ to 41.4℃, which slows down the nonlinear high-temperature aging rate of the battery array. At the same time, the overall power flow loss rate of the whole network has decreased from 4.83% under conventional control to 3.12%. This reflects that the nonlinear coupling and hedging of the dynamic thermal memory damping coefficient and the standard sign function can effectively solve the technical problem of transient conductance nonlinear temperature drift. The gradient distribution law of this data confirms that the range of protected parameter values ​​can provide a stable power flow regulation damping window for the distribution network. The dynamic carrying capacity boundary of the distribution network has changed from the original state to the digital control state.

[0043] Example 3: When the system faces the spatiotemporal fluctuation of reverse power flow caused by high-penetration distributed photovoltaic grid connection, in the 10kV active distribution network operation environment of digital evolution, the total grid-connected capacity of distributed photovoltaic reaches 8.5MW and the new energy penetration rate is 85%. The 2MW / 4MWh distributed energy storage system connected to 3 power flow sensitive nodes is frequently in microsecond-level and high-density power throughput state. As a result, transient temperature drift of impedance is induced in the converter bridge arm and local feeder lines, causing millisecond-level nonlinear changes in the equivalent dynamic conductance at the grid connection point. Under the network conditions of 120ms transient random delay and packet loss in the wireless communication network, the residual effect of electrothermal history causality with thermal inertia hysteresis characteristics and the lagging electrical response produce phase misalignment, causing false transient reversal of the voltage characteristics sampled locally. The traditional centralized optimization scheduling method causes the controller feedback gain to be oversaturated due to the lag in calculation convergence, which in turn causes output direction mismatch and high-frequency control oscillation, ultimately leading to accelerated aging of the battery array and degradation of the dynamic carrying capacity of the distribution network.

[0044] The main control chip of the general-purpose distributed energy management system and the digital signal processor of the energy storage converter run the method of this invention in real time. The processor collects the three-phase power frequency AC voltage at the grid connection point in real time with a sampling period of 10μs, and outputs the fundamental positive sequence voltage amplitude through the positive sequence component extraction operator. And according to the preset reference voltage nominal value Calculate the voltage deviation at the current moment. The timing differential operator with a differential step size of 2ms is used to calculate the voltage timing deviation rate of the current control node. Simultaneously, the energy storage throughput status bit tags of adjacent electrical nodes are obtained through the low-frequency carrier communication interface of the distribution network. The voltage timing deviation variation rate With feature labels Input a preset discrete arbitration matrix operator, output a globally consistent coordination scalar. The digital signal processor inputs this scalar into the built-in second-order dynamic inertial time-delay operator, smooths the scalar in the time domain, and calls the power flow sensitivity rules of the distribution network nodes to decouple and calculate the initial target command of the active power of the energy storage system at the current moment. With reactive power initial target command In this process, the node power time-varying balance topology equation is used to solve and assemble the multidimensional voltage and power sensitivity matrix in real time within the distributed energy management system. When solving the multidimensional voltage and power sensitivity matrix, the distributed energy management system establishes the node power time-varying balance topology equation based on Kirchhoff's current law and power conservation relationship of the current operating section of the active distribution network. It obtains the changes in injected active power and injected reactive power of each network node online, eliminates the phase angle changes of high-frequency alternation using the admittance matrix of the network topology, and simultaneously solves the real-time time-varying transmission sensitivity partial derivatives of the voltage amplitude changes of each node with respect to the changes in injected active power and injected reactive power. The real-time time-varying transmission sensitivity partial derivatives are arranged in a multidimensional manner according to the network node topology index order, and assembled into a multidimensional voltage and power sensitivity matrix containing active coupling terms and reactive coupling terms. This quantitatively establishes the control correlation network between the dynamic response of the distribution network node voltage and the power balance of the multi-point energy storage system. The topological equations are constructed based on Kirchhoff's current law and power conservation relationship at the current operating section of the active distribution network. Specifically, they are expressed as changes in injected active power and injected reactive power at each node, which are represented as a combination of linear partial derivative matrices of voltage amplitude changes and phase angle changes at all nodes in the network. Within a defined control period, the distributed energy management system acquires transient power change data of each electrical node online. It uses the admittance matrix of the network topology to eliminate the phase angle changes of high-frequency alternation, thereby simultaneously solving for the real-time time-varying transmission sensitivity partial derivatives of voltage amplitude changes at each node with respect to changes in injected active power and injected reactive power. These partial derivatives are arranged in a multidimensional matrix according to the topological index order of the network nodes, and assembled into a multidimensional voltage and power sensitivity matrix containing active and reactive coupling terms, thereby quantitatively establishing the correlation network between the dynamic voltage response of the distribution network nodes and the power balance of the multi-point energy storage system.

[0045] During the dynamic offsetting of operational costs and capacity truncation steps before the final control command is issued, the digital signal processor acquires the current state of charge of the energy storage battery array. Operating temperature at the highest point inside the array The heat dissipation resistance compensation parameters under the current operating conditions are calculated using the conductivity temperature drift trajectory of the line parasitic resistance. and at operating temperature When the temperature exceeds the preset normal operating boundary of 45°C and continues to accumulate, the dynamic capacity cutoff rule is triggered, and the cutoff coefficient is adjusted. Switching to 0.3, to counteract the control phase misalignment noise caused by impedance thermal inertia hysteresis, the processor uses a built-in discretization time series operator to read the initial active power target instructions generated in the preceding 20 consecutive sampling periods from the memory register. It then calculates the square-weighted discrete sum of the initial active power target instructions in the preceding 20 consecutive sampling periods using a square convolution operator, outputting a dynamic thermal inertia state variable that characterizes the residual heat accumulation caused by continuous current scouring of the parasitic impedance of the local area network frame at the current grid connection point. Based on the numerical range of the dynamic thermal inertia state variable, the preset adaptive self-locking damping gain constant is corrected in situ in real time, outputting the dynamic thermal memory damping coefficient. When the dynamic thermal inertia state variable does not exceed the preset safe thermal power consumption threshold, the dynamic thermal memory damping coefficient... Basic nominal value When the dynamic thermal inertia state variable crosses the safe thermal power consumption threshold and the voltage timing deviation rate of change at the current moment... When the absolute value of [value] shows a continuous increasing trend, the processor will dynamically thermally remember the damping coefficient. Switch to high damping step value However, when the initial target active power command is affected by cross-flow interference in the distribution network, With voltage timing deviation variation rate When the product of and is negative and the dynamic thermal inertia state variable resides at the preset high-level overload boundary, the processor will use the dynamic thermal memory damping coefficient. Forced to converge at a preset lower bound value The digital signal processor will calculate the dynamic thermal memory damping coefficient. Substituting the instructions into the instruction reconstruction equation, combined with the initial target instruction for active power... Voltage timing deviation variation rate Cutoff coefficient and heat dissipation resistance compensation parameters The nonlinear coupling calculation outputs the final active power control command, and its specific algebraic expression formula is as follows: ,in, This is the final active power control command. This is the initial target command for active power. This refers to the truncation coefficient corresponding to the capacity dynamic truncation rule. These are parameters for compensating for heat dissipation resistance. This represents the voltage timing deviation rate of variation at the current control node. The dynamic thermal memory damping coefficient, It is a standard symbolic function, and the subscript is... For electrical node index, superscript The initial state identifier is indicated by the superscript, and the final reconstructed state identifier is indicated by the subscript. This is a dynamic adaptive self-locking regulation indicator; when the energy storage system is charging and the local voltage is rising, or when the energy storage system is discharging and the local voltage is falling, the output value of the standard sign function is positive, and the dynamic virtual impedance damping term in the denominator, composed of the adaptive self-locking damping gain constant and the absolute value of the voltage timing deviation rate of change, takes effect, actively lowering the final active power control command. The amplitude of the converter feedback gain is reduced to mitigate oversaturation caused by transient impedance temperature drift. However, when the implicit interference from cross-current power flow in the distribution network causes a false transient reversal in the direction of the locally acquired voltage gradient, the initial target command is affected. With voltage timing deviation variation rate When the product is negative, the output of the standard sign function is negative, and the damping term in the denominator is transformed into a reverse compensation factor to maintain the convergence stability of the control drive software under power flow noise. Finally, the processor calculates the final active power control command. The signal is converted into a pulse-width modulated signal and drives the gate trigger circuit of the power semiconductor switching transistor in the energy storage converter to physically regulate the power flow of the distribution network. Under non-ideal conditions of 120 millisecond transient random delay and packet loss in the wireless communication network, the adaptive transient voltage constraint boundary is realized through an adaptive dynamic amplitude limiting criterion set internally by the processor. Specifically, the digital signal processor extends the nominal voltage of the grid-connected AC reference at the power frequency reference point upward and downward by 5% based on the currently locally generated transient voltage stability convergence reference value, respectively, as a safe voltage boundary closed loop for active power variation within the current control cycle. If the processor detects that the locally acquired voltage gradient deviates from this 5% safe voltage boundary due to the causal residual effect caused by the communication delay, the processor will take action accordingly. When the deviation lasts for more than a preset 10-millisecond dead zone, the system locks the step size of the current final active power control command, forcibly limiting the absolute value of the maximum change in a single dynamic change to no more than 50 kilowatts per second. This rigid adaptive boundary of the time-series causality replaces the failed wide-area network synchronization information, thereby maintaining transient stability control capability in a wide-area asynchronous environment during the transition period of perceived delay. Under the dynamic regulation effect constituted by the feedforward defense mechanism, the high-level heat accumulation caused by the preceding continuous power output action and the phase misalignment noise caused by the current electrical response are stripped online. The voltage gradient false reversal interference under the intermittent delay environment of the wireless communication link is eliminated, and the fundamental positive-sequence voltage amplitude of each power flow sensitive node is reduced. After 24 hours of continuous operation testing, the system remained stable within the safe operating window. The upper limit of the dynamic carrying capacity of distributed power sources across the entire network increased from 6.2MW to 8.2MW, representing an improvement of 32.25%. At the same time, the cumulative duration of node voltage over-limit was reduced to zero, and the highest daily operating temperature of individual energy storage battery array cells decreased from 51.2℃ to 41.5℃, avoiding array aging caused by frequent nonlinear overcharging and over-discharging. Ultimately, the overall power flow loss rate of the entire network decreased from 4.8% to 3.1%, and the dynamic carrying capacity boundary of the distribution network changed from the original state to a digitally controlled state.

[0046] Example 4: When the system faces the deployment and commissioning of a 10kV active distribution network under digital evolution conditions, the basic nominal values ​​are established for the main control chip. As the initial benchmark, the digital signal chip initiates a calibration program under stable operating conditions at the grid connection point, acquiring the voltage amplitude of the feeder network and specifying a sampling step size of 10μs to continuously read 1000 data points. The main control chip evaluates the local area network noise power by calculating the variance of the 1000 data points. When the output variance is within the range of 0.01 to 0.05, the main control chip retrieves the preset baseline table and writes it into the address space of the memory register to determine the heat dissipation resistance compensation parameters. The initial zero point is used to establish the initial state boundary between the external environment and the algorithm logic.

[0047] When the commissioning condition transitions to a high-frequency fluctuation condition in active power and the fundamental positive sequence voltage amplitude... When deviating from the steady-state boundary, the main control chip activates the thermal memory self-correction loop for monitoring, and the digital signal chip tracks the dynamic thermal inertia state variables based on the initial state boundary. The upward drift, when the dynamic thermal inertia state variable Exceeding the safe thermal power threshold At that time, the high-damping constraint branch starts, among which, It is the square-weighted discrete sum of the initial target active power command within the preceding 20 consecutive sampling periods, obtained by solving the square convolution operator at the current moment. Active power control command for preset safe thermal power consumption threshold Nonlinear step suppression is generated, the gate trigger circuit of the converter semiconductor switch outputs an adjustment signal, the impedance transient temperature drift is offset, and the fundamental positive sequence voltage amplitude is reduced. Returning to a stable state.

[0048] Example 5: When the system faces the operating condition of transient temperature drift of the equivalent impedance at the grid connection point of the distribution network caused by high-frequency random fluctuations of distributed power sources, the main control chip establishes a standardized pre-calibration procedure during the deployment and commissioning phase of the 10kV active distribution network under digital evolution conditions. Through the orderly calibration of multi-dimensional dynamic parameters, a convergence boundary for the control loop under high noise injection is constructed for the distributed energy storage system. The main control chip acquires the basic electrothermal test data of the active distribution network under steady-state conditions. When the highest operating temperature of the energy storage battery array is monitored to be stable at 35℃ and lower than the preset operating boundary of 45℃, the cutoff coefficient is set. The value is set to 0. 1000 AC voltage data points are continuously read at a sampling period of 10μs through the signal acquisition interface of the grid connection point, and the variance of this set of AC voltage data points is calculated. When the calculated variance is within the range of 0.01 to 0.05, the local area network is determined to be in a low-disturbance state, and the dynamic thermal memory damping coefficient is set to... Configuration is based on nominal value The basic nominal value is gradually fine-tuned in conjunction with the pulse excitation input from the test bench. The value is set to ensure that the initial active power target command output by the digital signal processor maintains a relaxation time of 12μs after the pulse excitation is released, and to satisfy the basic nominal value of the relaxation time constraint. The value 1.24 is stored in the preset address space in the memory register. Before the system is put into high-frequency alternating power dispatch, the digital signal processor starts the pre-calibration program under the steady-state condition of the active distribution network. It continuously reads 1000 sampling point data with a sampling period of 10 microseconds through the voltage and current sensors at the grid connection point, calculates the variance of the 1000 sampling point data to evaluate the background noise power of the local area network, and determines that the network is in a low disturbance state when the value is in the range of 0.01 to 0.05. It then injects an initial heat dissipation resistance compensation with a value of zero into the feeder temperature drift impedance time-varying feedforward adjustment loop. The parameters are retrieved, and the temperature drift curve of the reference conductivity of the metal wire material as a function of temperature is obtained. The cold resistance value at an ambient temperature of 25 degrees Celsius is used as the initial algebraic iteration reference for the time-varying equivalent resistance characteristic. An unambiguous initial state boundary is established between the external physical network environment and the internal control variables of the controller. The main control chip monitors the energy storage battery array through external thermocouple sensors to match the continuous high-power throughput caused by the high-frequency jump of the distributed power source. When the highest operating temperature of the array rises to 42.3°C and continues to accumulate to the normal operating boundary of 45°C, the capacity dynamic cutoff rule is triggered and the cutoff coefficient is set. The value is adjusted from 0 to 0.3. The built-in discretization time series operator is called to read the initial active power target instructions generated in the preceding 20 consecutive sampling periods from the memory register. The processor is assigned to solve the square-weighted discrete sum of the initial active power target instructions in the preceding 20 consecutive sampling periods through the square convolution operator to output the dynamic thermal inertia state variable. When the dynamic thermal inertia state variable crosses the preset safe thermal power consumption threshold and the absolute value of the current node voltage timing deviation change rate calculated by the timing difference operator with a differential step size of 2ms shows a continuous increasing trend, the high-damping constraint branch is activated to set the dynamic thermal memory damping coefficient. The value has been adjusted from the basic nominal value of 1.24 to the high-damping step value. By continuously injecting high-frequency fluctuation data of new energy sources into the test bench, the high-damping step value is finely adjusted. The size, and in the high damping step value When the maximum fluctuation deviation of the access point voltage is converged to within 3.5% of the preset safety threshold, the high-damping step value of 3.50 is fixed as the current control benchmark. The core control parameters fixed in the calibration procedure all have clear theoretical limits and engineering upper and lower bounds. Specifically, the capacity dynamic cutoff coefficient is set to 0.3, with an allowable engineering variation range of 0.15 to 0.45. A value below 0.15 will fail to effectively suppress battery heating under overheating conditions, while a value above 0.45 will lead to excessive reduction in the active power regulation capacity of energy storage, reducing its capacity to support new energy sources. The basic nominal value of the dynamic thermal memory damping coefficient is 1.24, with a theoretical boundary of 0.50 to 2.00. A value below 0.50 will cause voltage relaxation of the system under small disturbance conditions. A response time exceeding 20 microseconds will cause hysteresis, while a value above 2.00 will cause excessive gain attenuation in the control loop under normal conditions. The high-damping step value is 3.50, with a physical boundary range of 2.50 to 5.00. If it is below 2.50, it will not be able to generate sufficient self-locking damping to suppress feedback gain saturation during transient temperature drift outbreaks. If it is above 5.00, it will induce reverse overshoot control oscillations in the denominator. The lower limit convergence value is 0.15, with a defined engineering safety range of 0.05 to 0.40. If it is below 0.05, the system will lose the necessary reverse bridging damping compensation capability during pseudo-transient flips. If it is above 0.40, it will cause the feedforward gain to diverge at the moment of communication recovery. Through these value range windows determined by limit calculations, the overall stability of the loop is ensured during extreme operating condition switching.

[0049] The system introduces a cross-current power flow interference condition in the distribution network to calibrate the convergence boundary under extreme network noise. When a 120ms transient random delay occurs in the wireless communication network and the absolute value of the time-series variation rate of the equivalent dynamic conductance of the local area network exceeds 0.05 Siemens per second, causing a false transient reversal in the direction of the locally acquired voltage gradient, this reversal makes the product of the initial target command for active power and the voltage time-series deviation variation rate negative. When the main control chip detects that the dynamic thermal inertia state variable resides at the high-level overload boundary, it forcibly starts the extreme value convergence logic and sets the dynamic thermal memory damping coefficient. Forced to converge at a preset lower bound value The lower limit convergence value is adjusted in conjunction with the comparison results of the control loop stability. The value of is determined by the convergence of the lower bound. When configured to 0.15, the digital signal processor maintains the convergence of the control drive software under a communication network delay of 120ms to eliminate high-frequency oscillations in the control action. The main control chip substitutes the determined variables into the active power control command reconstruction equation, where the final active power control command calculation formula is as follows: ,in, This is the final active power control command. This is the initial target command for active power. This refers to the truncation coefficient corresponding to the capacity dynamic truncation rule. These are parameters for compensating for heat dissipation resistance. This represents the voltage timing deviation rate of variation at the current control node. The dynamic thermal memory damping coefficient, For standard symbolic functions, For time variables, As an electrical node index, the digital signal processor outputs the final control result according to the calculation formula. Under the digital evolution condition, the fixed parameter configuration is sent to the 2MW / 4MWh distributed energy storage system mounted on three power flow sensitive nodes and continuously tested for 24 hours. When the grid-connected capacity of new energy photovoltaic reaches 8.5MW and the new energy penetration rate reaches 85%, the calibration system collects reproducible data kernels characterizing the steady-state convergence of the control flow. Among them, the upper limit of the dynamic carrying capacity of the distributed power supply of the whole network is stably expanded to 8.24MW, which is an improvement compared to the upper limit of 6.21MW of the carrying capacity corresponding to the control system without this parameter configuration. The cumulative duration of node voltage over-limit in the active distribution network is reduced from 145.6 minutes to 0 minutes. The highest daily operating temperature of the single cell of the energy storage battery array is reduced from 51.2℃ to 41.4℃. The overall power flow loss rate of the whole network is reduced from 4.83% to 3.12%. When the dynamic thermal memory damping coefficient As the residual electric heating increases with accumulation, the final active power control command... The amplitude is suppressed by in-situ nonlinearity, so that the fundamental positive sequence voltage amplitude of each power flow sensitive node is maintained within the steady-state boundary under the transient random delay condition of the 120ms wireless communication network.

[0050] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0051] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for optimizing the configuration of energy storage in distribution networks to enhance the carrying capacity of distributed power sources, characterized in that, Includes the following steps: Step S1: Collect real-time network source operation status data of the active distribution network online, calculate the partial derivatives of the voltage amplitude of each grid node with respect to the injected active power and the injected reactive power in the distributed energy management system, so as to assemble and construct a multi-dimensional voltage-power sensitivity matrix and quantitatively establish the control correlation network between the dynamic response of the distribution network node voltage and the reactive four-quadrant balance of the multi-point energy storage system. Step S2: Dynamically monitor the spatiotemporal distribution characteristics of reverse power flow caused by distributed power grid connection, use the multidimensional voltage-power sensitivity matrix to match the active and reactive power output characteristics of the energy storage converter online, and adaptively adjust the power transmission and distribution path of each feeder to reduce the power flow loss of the entire distribution network. Step S3: The active power regulation command time sequence continuously cyclically generated within the pre-control cycle is input into the feeder temperature drift impedance time-varying feedforward regulation loop. The feeder time-varying equivalent resistance characteristic quantity, which characterizes the transient temperature drift of the feeder impedance caused by the conduction current heating, is analyzed and calculated. The current converter loop damping factor is adjusted through the feeder time-varying equivalent resistance characteristic quantity to constrain the dynamic control gain of the energy storage converter transient regulation loop, offset the nominal deviation of voltage variation caused by the transient temperature drift of the feeder equivalent impedance, generate a transient voltage stability convergence benchmark, smooth out node voltage over-limit and suppress energy storage transient oscillation.

2. The method for optimizing the configuration of energy storage in distribution networks to enhance the carrying capacity of distributed power sources according to claim 1, characterized in that, Step S1 includes the following sub-steps: Step S11, online acquisition of the changes in injected active power, injected reactive power, and node voltage amplitude of each node in the active distribution network within a determined control period; Step S12, establishment of the node power time-varying balance topology equations within the distributed energy management system, simultaneous solution of the time-varying transmission sensitivity of the changes in injected active power and injected reactive power with the changes in node voltage amplitude, and assembly of the multidimensional voltage-power sensitivity matrix.

3. The method for optimizing the configuration of energy storage in distribution networks to enhance the carrying capacity of distributed power sources according to claim 1, characterized in that, Step S2 includes the following sub-steps: Step S21, based on the multidimensional voltage-power sensitivity matrix, quantitatively calculate the local feeder overheating limit constraint conditions caused by active and reactive power backfeed, and identify the voltage over-limit risk range in the spatiotemporal distribution characteristics of reverse power flow caused by distributed power grid connection. Step S22: For the voltage over-limit risk range, allocate the capacity ratio of active power and reactive power output from the energy storage converter in the four quadrants to limit local feeder overheating and balance the power flow distribution of the entire network.

4. The method for optimizing the configuration of energy storage in distribution networks to enhance the carrying capacity of distributed power sources according to claim 1, characterized in that, Step S3 includes the following sub-steps: Step S31, continuously sample the active power regulation command time series of multiple historical time nodes according to level, and construct the pre-cycle command regulation characteristic time series; Step S32, input the pre-cycle command regulation characteristic time series into the first-order equivalent heat loss dissipation conduction model of the distribution network to perform impedance transient temperature drift calculation, and output the feeder time-varying equivalent resistance characteristic quantity.

5. The method for optimizing the configuration of energy storage in distribution networks to enhance the carrying capacity of distributed power sources according to claim 1, characterized in that, After adjusting the current converter loop damping factor to constrain the dynamic control gain of the energy storage converter transient regulation loop by using the feeder time-varying equivalent resistance characteristic in step S3, the following sub-steps are also included: Step S33, collect the node voltage step change rate signal of each feeder, and extract the voltage variation polarity indicator that characterizes the direction of the node voltage step through the sign function operator; Step S34, perform nonlinear product mapping between the voltage variation polarity indicator and the converter loop damping factor to generate a reverse phase offset compensation amount to eliminate the voltage variation phase deviation, filter out the in-phase sudden step disturbance from the transient regulation loop of the energy storage converter, and output the transient voltage stability convergence benchmark.

6. The method for optimizing the configuration of energy storage in distribution networks to enhance the carrying capacity of distributed power sources according to claim 1, characterized in that, The method also includes the following steps: Step S4, establish a basic database of full-cycle multi-dimensional operation history evolution, and continuously and online statistically analyze the dynamic response time series of active and reactive power of multi-point energy storage system under extremely high frequency output conditions, the frequency of charge and discharge cycle switching, and the cumulative time shift of feeder time-varying equivalent resistance characteristics.

7. A method for optimizing the configuration of energy storage in a distribution network to enhance the carrying capacity of distributed power sources, as described in claim 6, is characterized in that... The method also includes the following steps: Step S5, extract the rate of change and steady-state deviation of the feeder time-varying equivalent resistance characteristic from the full-cycle multi-dimensional operation history evolution database, and combine the charge and discharge cycle switching frequency to calculate the health degradation index characterizing the overall aging degree of the multi-point energy storage system, and quantitatively present the performance evolution state of the energy storage array during long-term operation.

8. The method for optimizing the configuration of energy storage in a distribution network to enhance the carrying capacity of distributed power sources, as described in claim 7, is characterized in that... The method also includes the following steps: Step S6, when the health degradation index is greater than or equal to the determined safe life threshold, automatically reduce the upper limit of the output of the current active power adjustment command sequence, and proportionally increase the adjustment weight of reactive power in the four-quadrant balance, so as to delay the aging of the energy storage array through reactive power priority adjustment control.

9. A method for optimizing the configuration of energy storage in distribution networks to enhance the carrying capacity of distributed power sources, as described in claim 1, is characterized in that... The method also includes the following steps: Step S7, under the condition that the sensing delay is caused by wide-area non-ideal communication in the active distribution network, the transient voltage stability convergence benchmark provides an adaptive transient voltage constraint boundary with a time-series causal relationship for the transient active power command, and dynamically limits the variation range of the transient voltage stability convergence benchmark in order to maintain the transient stability control capability of the multi-point energy storage system in a wide-area asynchronous information environment.

Images