Intrinsic voltage estimation and parameter self-tuning method and system for low-voltage governance

Abstract

The application discloses an intrinsic voltage estimation and parameter self-setting method and system for low-voltage governance, and is applied to the technical field of smart grids, and the method comprises the following steps: acquiring electrical quantity data of a target governance device; correcting the current line equivalent resistance and line equivalent reactance according to the electrical quantity data; calculating the intrinsic power grid voltage according to the electrical quantity data, the corrected line equivalent resistance and the corrected line equivalent reactance; generating the proportional gain of a closed-loop controller through a preset inverse proportional relationship of the corrected line equivalent resistance, generating the integral gain and the differential gain according to the proportional gain according to a preset proportion; and updating the control parameters through the calculated proportional gain, integral gain and differential gain. According to the application, the line impedance does not need to be manually input, the line characteristics are self-learned through online active disturbance and passive observation, the deployment conditions of different transformer areas, different lengths and different loads are adapted, and the deployment cost and maintenance cost are significantly reduced.

Description

Technical Field

[0001] This invention relates to the field of intelligent voltage management technology, and more specifically to an intrinsic voltage estimation and parameter self-tuning method and system for low voltage management. Background Technology

[0002] Low-voltage distribution areas are prone to low-voltage problems for end users in rural power grids, mountainous areas, long-radius power supply areas, and areas with seasonal load fluctuations. Energy storage-type low-voltage end-user management devices inject or absorb active power into the grid through charging and discharging, thereby raising or lowering the local voltage to maintain a qualified voltage for end users. In this process, the device must rely on three core parameters to operate stably: The first category is the equivalent impedance parameters of the line, including the line resistance R and the line reactance X. The line impedance determines the magnitude of the rise or fall of the voltage at the local measurement point caused by the injection or absorption of a unit of active power, directly affecting the power command calculation, closed-loop response speed, and control stability.

[0003] The second category is intrinsic grid voltage, which is the voltage corresponding to the original power supply state of the upstream transformer area after the influence of the device's own charging and discharging current on the voltage at the measurement point has been removed. Intrinsic grid voltage serves as input data for higher-level algorithms such as historical learning, target voltage decision-making, and load identification.

[0004] The third category includes the proportional gain, integral gain, and derivative gain of closed-loop controllers. Energy storage devices track the target voltage through a PID closed-loop system. The controller gain selection needs to match the line impedance: when the impedance is high, the impact of unit power on voltage is large, requiring a smaller gain to avoid overshoot; when the impedance is low, the impact of unit power is small, requiring a larger gain to ensure response speed.

[0005] The above three types of parameters have the following problems in actual deployment: First, line impedance parameters are difficult to obtain. Low-voltage terminal lines have complex structures and varying lengths, and user load types differ greatly. It is impossible to pre-record impedance parameters for every device before it leaves the factory. During the deployment phase, manual measurement and entry of impedance is costly and prone to errors. Furthermore, line impedance will drift with changes in season, temperature, and user access, and manually entered values ​​will quickly become invalid.

[0006] Second, the intrinsic grid voltage cannot be directly measured. Existing low-voltage regulation devices generally use the measured voltage at the measurement point as historical samples and the basis for target decision-making. However, when the device injects or absorbs active power into the grid, the voltage at the measurement point relative to the original grid voltage will be raised or lowered by the device's charging and discharging; when there is capacitive reactive current in the device, the voltage at the measurement point will also be raised by the reactive current. Directly using the measured voltage for historical learning will misjudge the device's own regulation effect as the natural state of the grid, causing historical sample contamination, target voltage deviation, and even the abnormal phenomenon of "the more it is regulated, the more it deviates."

[0007] Third, controller parameters rely on empirical tuning. Existing solutions often use PID gain parameters that are fixed in the laboratory and then programmed into the code, making it impossible to adapt to different distribution area line conditions in the field. Using empirical gain when the line impedance is high can lead to overshoot oscillations; using empirical gain when the impedance is low can result in slow response and failure to keep up with the target voltage. Classical self-tuning methods such as the Ziegler-Nichols method rely on system identification or critical oscillation tests, which are neither safe nor economical in low-voltage control scenarios.

[0008] Fourth, the methods for identifying line impedance are limited, and it is difficult to simultaneously separate R and X. Existing impedance identification methods mostly use a single signal source (either active injection or passive observation), and often only identify one of R or R+X, which is difficult to meet the simultaneous requirement in low-voltage management devices that "R determines the voltage amplitude response and X determines the reactive current rise correction." When only R is used, the intrinsic voltage estimation has a systematic deviation under capacitive reactive current; when only X is used, it is impossible to generate accurate PID gain.

[0009] Fifth, active perturbations can disturb users, while passive observations lack observability. Actively injecting power perturbations can yield high-quality samples, but frequent execution can negatively impact users' electricity experience. Passive observations, while not interfering with users, suffer from small natural power variations, high correlation between active and reactive power, and inability to separate X when the matrix is ​​singular. Finding a balance between these two approaches is crucial for engineering implementation.

[0010] Sixth, there is a lack of a unified upper-level interface for parameters from different sources. Devices often have multiple sources of impedance, such as manually configured impedance, learning impedance, and factory default impedance. If the upper-level algorithm needs to perceive each source individually, the code will have many branches and be prone to errors. Summary of the Invention

[0011] In order to at least overcome the above-mentioned deficiencies in the prior art, the purpose of this application is to provide an intrinsic voltage estimation and parameter self-tuning method and system for low-voltage governance.

[0012] In a first aspect, embodiments of this application provide an intrinsic voltage estimation and parameter self-tuning method for low-voltage management, including: The electrical quantity data of the target treatment device is acquired periodically; the electrical quantity data is divided into first electrical quantity data acquired through active power perturbation and second electrical quantity data acquired through passive operation response. The current line equivalent resistance and line equivalent reactance are corrected based on the electrical quantity data using an exponentially weighted moving average algorithm; the correction weight of the first electrical quantity data is greater than the correction weight of the second electrical quantity data. Select the corresponding reverse strategy based on the current working state of the target governance device, and calculate the intrinsic grid voltage based on electrical quantity data, the corrected line equivalent resistance, and the corrected line equivalent reactance. The corrected equivalent resistance of the line is used to generate the proportional gain of the closed-loop controller through a preset inverse proportional relationship. The integral gain and derivative gain are then generated according to the preset ratio based on the proportional gain. The control parameters are updated by calculating the proportional gain, integral gain, and derivative gain.

[0013] In one possible implementation, acquiring the first electrical quantity data includes: During off-peak hours at night, the energy storage converter is instructed to enter standby mode, and after the energy storage converter is in standby mode, the measured voltage, active current component and reactive current component are collected as baseline electrical quantity data. Multiple sets of power pulses with different pulse powers are injected sequentially, and the corresponding measured voltage, active current component, and reactive current component are collected as steady-state electrical quantity data after each set of power pulses is injected and reaches a steady state. The multiple sets of power pulses with different pulse powers include forward pulses and reverse pulses. The injection process includes single converter injection and multi-converter parallel injection. The difference between each set of steady-state electrical quantity data and the corresponding baseline electrical quantity data is calculated as the first electrical quantity data; The acquisition of the second electrical quantity data includes: During off-peak hours outside of nighttime electricity consumption, electrical quantity data is continuously sampled at a preset cycle. The second electrical quantity data is obtained by differentiating the electrical quantity data of the current sampling point from the electrical quantity data of the previous sampling point.

[0014] In one possible implementation, the electrical quantity data includes voltage difference ΔV, active current difference ΔIp, and reactive current difference ΔIq; Correcting the current line equivalent resistance and line equivalent reactance based on the aforementioned electrical quantity data includes: Based on the electrical quantity data, obtain the total number of samples count, active square sum sip2=Σ(ΔIp²), reactive square sum siq2=Σ(ΔIq²), active and reactive product sum sipiq=Σ(ΔIp×ΔIq), active voltage product sum sipdv=Σ(ΔV×ΔIp), and reactive voltage product sum siqdv=Σ(ΔV×ΔIq); When reactance is observable, the resistance observation Rmeasured and the reactance observation Xmeasured are calculated according to the following formulas: Rmeasured=(siq2×sipdv−sipiq×siqdv) / det; Xmeasured=(sip2×siqdv−sipiq×sipdv) / det; det = sip2 × siq2 − sipiq²; In the formula, det is the determinant of the coefficient matrix; When reactance is unobservable, the observed resistance value Rmeasured is calculated using the following formula, without correcting the equivalent line reactance in this round: Rmeasured = sipdv / sip2; The current equivalent line resistance and equivalent line reactance are corrected using an exponentially weighted moving average algorithm based on the observed resistance and reactance values.

[0015] In one possible implementation, the determination of whether the reactance is observable includes: Calculate the average value of the sum of squares of reactive power avgiq2=siq2 / count. If avgiq2 is less than the first preset threshold, the reactance is determined to be unobservable. If the absolute value of the determinant of the coefficient matrix is ​​less than or equal to the second preset threshold, the matrix is ​​determined to be singular and the reactance is unobservable.

[0016] In one possible implementation, correcting the current equivalent line resistance and equivalent line reactance based on resistance and reactance observations using an exponentially weighted moving average algorithm includes: The corrected equivalent resistance and equivalent reactance of the line are calculated using the following formula: Rnew=Rold×(1-wR)+Rmeasured×wR; Xnew=Xold×(1-wX)+Xmeasured×wX; In the formula, Rnew is the corrected equivalent line resistance, Xnew is the corrected equivalent line reactance, Rold is the current equivalent line resistance, Xold is the current equivalent line reactance, wR is the correction weight of the equivalent resistance corresponding to the first electrical quantity data or the second electrical quantity data, and wX is the correction weight of the equivalent reactance corresponding to the first electrical quantity data or the second electrical quantity data.

[0017] In one possible implementation, calculating the intrinsic grid voltage includes: When the measured voltage is invalid, the intrinsic grid voltage of the previous round is used as the intrinsic grid voltage calculated in this round. When both the measured reactive current and the measured active current are less than the preset value, the measured voltage will be taken as the intrinsic grid voltage. When the target governance device is currently in standby mode, the intrinsic grid voltage is calculated according to the following formula: Vn = Vm − Iq × Xnew; In the formula, Vn is the intrinsic grid voltage, Vm is the measured voltage, Iq is the measured reactive current, and Xnew is the corrected line equivalent reactance. When the target treatment device is currently in a charging state, the intrinsic grid voltage is calculated according to the following formula: Vn = Vm + Ip × Rnew − Iq × Xnew; In the formula, Ip is the measured active current, and Rnew is the corrected line equivalent resistance; When the target treatment device is currently in a discharge state, the intrinsic grid voltage is calculated according to the following formula: Vn = Vm - Ip × Rnew − Iq × Xnew.

[0018] In one possible implementation, the calculation of the proportional gain, integral gain, and derivative gain, as well as the updating of the control parameters, include: Calculate the proportional gain using the following formula: Kp = α × 110 / Rnew; In the formula, Kp is the proportional gain, α is the safety margin, and Rnew is the corrected line equivalent resistance. The integral gain is calculated using the following formula: Ki = β × Kp; In the formula, Ki is the integral gain, and β is the integral coefficient; The differential gain is calculated using the following formula: Kd = γ × Kp; In the formula, Kd is the differential gain, and β is the differential coefficient; The calculated proportional gain Kp is clamped between a preset lower limit and an upper limit of the proportional gain; The final proportional gain, integral gain, and derivative gain are used as the current control parameters.

[0019] In one possible implementation, obtaining the control parameters further includes: Control parameters are obtained based on the following priority: First priority: When a non-zero control parameter is manually set, the manual value is used directly as the current control parameter; Second priority: When the device starts up, the learned value of the control parameter corresponding to the direction is loaded as the current control parameter; the direction includes the charging direction and the discharging direction; Third priority: Use the calculated proportional gain, integral gain, and derivative gain as the current control parameters; Fourth priority: When the device is powered on for the first time, the impedance is not calibrated, and the learning value has not been established, the default value is used as the current control parameter.

[0020] In one possible implementation, obtaining the learned values ​​includes: In each control cycle, read the actual effective proportional gain; When the target governance device is currently in a charging state, the current proportional gain is stored in the sliding window queue in the charging direction; when the target governance device is currently in a discharging state, the current proportional gain is stored in the sliding window queue in the discharging direction. When the number of data in the sliding window queue in the charging direction reaches the length of the preset window, the arithmetic mean of all proportional gains in the sliding window queue in the charging direction is calculated as the first sliding average. When the number of data in the sliding window queue in the discharge direction reaches the length of the preset window, the arithmetic mean of all proportional gains in the sliding window queue in the discharge direction is calculated as the second sliding average. Using a preset evaluation period, the first moving average value is compared with the learned value of the stored charging direction, and the second moving average value is compared with the learned value of the stored discharging direction. When the difference between the first moving average and the learned value of the stored charging direction exceeds a preset update threshold, the first moving average is used as the learned value of the new charging direction; when the difference between the second moving average and the learned value of the stored discharging direction exceeds a preset update threshold, the second moving average is used as the learned value of the new discharging direction.

[0021] Secondly, this application also provides an intrinsic voltage estimation and parameter self-tuning system for low-voltage management, including: The acquisition unit is configured to periodically acquire electrical quantity data of the target treatment device; the electrical quantity data is divided into first electrical quantity data acquired through active power perturbation and second electrical quantity data acquired through passive operation response; The weighting unit is configured to correct the current line equivalent resistance and line equivalent reactance based on the electrical quantity data using an exponentially weighted moving average algorithm; the correction weight of the first electrical quantity data is greater than the correction weight of the second electrical quantity data; The reverse calculation unit is configured to select a corresponding reverse calculation strategy based on the current operating state of the target governance device, and calculate the intrinsic grid voltage based on electrical quantity data, the corrected line equivalent resistance, and the corrected line equivalent reactance. The gain unit is configured to generate the proportional gain of the closed-loop controller by using the corrected equivalent resistance of the line through a preset inverse proportional relationship, and to generate integral gain and derivative gain according to the proportional gain at a preset ratio. The update unit is configured to update the control parameters by calculating the proportional gain, integral gain, and derivative gain.

[0022] Compared with the prior art, the present invention has the following advantages and beneficial effects: This invention provides an intrinsic voltage estimation and parameter self-tuning method and system for low-voltage governance. It eliminates the need for manual input of line impedance, learning line characteristics through online active disturbance and passive observation. This adapts to deployment conditions of different transformer areas, lengths, and loads, significantly reducing deployment and maintenance costs. The three-state differentiated back-calculation covers all operating states of the device, ensuring the intrinsic grid voltage, stripped of the device's own influence, is obtained at any given time, avoiding historical sample contamination. When sample observability conditions are met, two-dimensional least squares are used to jointly solve for resistance and reactance; otherwise, it automatically downgrades to one-dimensional solution for resistance only, avoiding reactance contamination under weak observability conditions, thus balancing accuracy and robustness. Active pulses are of high quality, causing interference to users but occurring infrequently, while passive observations are of slightly lower quality, causing no interference to users but occurring frequently. The dual sources are fused with differentiated weights, ensuring sample quality while reducing the impact on user power consumption. Persistent learning values ​​allow the device to retain optimal parameters accumulated over long-term operation after power failure and restart, shortening the time to reach optimal control state after startup. Attached Figure Description

[0023] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram of the method steps in an embodiment of this application. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the accompanying drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0025] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0026] Please refer to the following: Figure 1 This is a flowchart illustrating the intrinsic voltage estimation and parameter self-tuning method for low-voltage governance provided in this embodiment of the invention. Further, the intrinsic voltage estimation and parameter self-tuning method for low-voltage governance may specifically include the contents described in steps S1-S5.

[0027] S1: Periodically acquire electrical quantity data of the target treatment device; the electrical quantity data is divided into first electrical quantity data acquired through active power injection disturbance and second electrical quantity data acquired through passive operation response; S2: Based on the electrical quantity data, the current line equivalent resistance and line equivalent reactance are corrected using an exponentially weighted moving average algorithm; the correction weight of the first electrical quantity data is greater than the correction weight of the second electrical quantity data; S3: Select the corresponding reverse strategy according to the current working state of the target governance device, and calculate the intrinsic grid voltage based on the electrical quantity data, the corrected line equivalent resistance and the corrected line equivalent reactance; S4: The corrected equivalent resistance of the line is used to generate the proportional gain of the closed-loop controller through a preset inverse proportional relationship, and the integral gain and derivative gain are generated according to the preset ratio based on the proportional gain. S5: Update the control parameters using the calculated proportional gain, integral gain, and derivative gain.

[0028] In implementing this application embodiment, it is necessary to first acquire dual-source physical foundation data, namely, first electrical quantity data and second electrical quantity data. The electrical quantity data consists of the measured voltage, active current component, and reactive current component of the device, serving as the fundamental input data for impedance identification and intrinsic voltage estimation. Active power disturbance injection involves actively injecting known power pulses into the grid and identifying impedance by measuring the voltage and current responses. This method is characterized by high sample quality but has a brief impact on users; therefore, it is typically performed during low-load periods at night, injecting power pulses of known amplitude into the grid through an energy storage converter. Passive operation response, on the other hand, involves collecting electrical quantity fluctuations generated by natural power changes during normal operation. This method is characterized by no user interference but a lower signal-to-noise ratio. Therefore, in the process of correcting resistance and reactance using electrical quantity data, the data detected by active disturbance is more accurate and has a higher signal-to-noise ratio, allowing for a larger weighting and more accurate correction of resistance and reactance. In calculating the intrinsic grid voltage, since the measured voltage is not the actual grid voltage, but rather the intrinsic grid voltage minus or plus the voltage drop across the line impedance caused by the device's charging and discharging current, this application embodiment sets different back-calculation strategies for charging, discharging, and standby scenarios to deduce the accurate intrinsic grid voltage. Similarly, the control parameters of the PID controller can be further calculated based on the equivalent resistance of the line. It should be understood that in the process of this application embodiment, each time new electrical quantity data is collected, a round of calculation is required to determine whether to update the parameters accordingly, and then the update is completed. This application embodiment has strong adaptability; it does not require manual input of impedance parameters, and the device can complete parameter tuning online. At the same time, by assigning high weights to high-quality active samples and low weights to passive samples, more accurate parameter updates can be performed daily, and the parameters can be updated more in real time.

[0029] In one possible implementation, acquiring the first electrical quantity data includes: During off-peak hours at night, the energy storage converter is instructed to enter standby mode, and after the energy storage converter is in standby mode, the measured voltage, active current component and reactive current component are collected as baseline electrical quantity data. Multiple sets of power pulses with different pulse powers are injected sequentially, and the corresponding measured voltage, active current component, and reactive current component are collected as steady-state electrical quantity data after each set of power pulses is injected and reaches a steady state. The multiple sets of power pulses with different pulse powers include forward pulses and reverse pulses. The injection process includes single converter injection and multi-converter parallel injection. The difference between each set of steady-state electrical quantity data and the corresponding baseline electrical quantity data is calculated as the first electrical quantity data; The acquisition of the second electrical quantity data includes: During off-peak hours outside of nighttime electricity consumption, electrical quantity data is continuously sampled at a preset cycle. The second electrical quantity data is obtained by differentiating the electrical quantity data of the current sampling point from the electrical quantity data of the previous sampling point.

[0030] In the implementation of this application embodiment, a specific scheme for acquiring first electrical quantity data and second electrical quantity data is provided. The first electrical quantity data is collected during off-peak electricity hours at night when user load is low and the power grid background is stable. The energy storage converter is instructed to enter standby mode. In standby mode, the device output is zero, and the collected electrical quantities reflect the operating state of the power grid itself; that is, the baseline voltage, baseline active current, and baseline reactive current constitute the zero-power reference point baseline electrical quantity data. Subsequently, multiple sets of power pulses with different powers are injected sequentially, and the corresponding electrical quantity data are collected after steady state. The voltage and current changes caused by the known power injected by the device are the sole factor, while the original background load of the power grid is subtracted from the baseline. Therefore, the first electrical quantity data after the difference only reflects the impact of the injected power on the system. Given the injected power value, the corresponding voltage and current responses are measured, and the system impedance can be calculated according to Ohm's law. In the embodiments of this application, a positive pulse causes the voltage to rise and a negative pulse causes the voltage to fall. The two pulses cover both ends of the voltage regulation range, which means that the energy storage device is in two states of charging and discharging. In the complementary combination of a single converter and multiple converters in parallel, the reactive component is smaller when a single converter is injected, and the reactive component increases significantly when multiple converters are in parallel. This switching makes the current change in the sample have a large range of variation, so that the reactance can be accurately identified.

[0031] In the embodiments of this application, during normal daytime operation, the second electrical quantity data obtained by differential sampling is continuously sampled at a preset period. The differential calculation process expresses that when the device power changes between two adjacent sampling periods, the grid connection point voltage will change slightly accordingly. By capturing this change value through differential sampling, it can be used for the calculation of the corresponding impedance. It utilizes natural power fluctuations, has no interference to users, but has a low signal-to-noise ratio.

[0032] In one possible implementation, the electrical quantity data includes voltage difference ΔV, active current difference ΔIp, and reactive current difference ΔIq; Correcting the current line equivalent resistance and line equivalent reactance based on the aforementioned electrical quantity data includes: Based on the electrical quantity data, obtain the total number of samples count, active square sum sip2=Σ(ΔIp²), reactive square sum siq2=Σ(ΔIq²), active and reactive product sum sipiq=Σ(ΔIp×ΔIq), active voltage product sum sipdv=Σ(ΔV×ΔIp), and reactive voltage product sum siqdv=Σ(ΔV×ΔIq); When reactance is observable, the resistance observation Rmeasured and the reactance observation Xmeasured are calculated according to the following formulas: Rmeasured=(siq2×sipdv−sipiq×siqdv) / det; Xmeasured=(sip2×siqdv−sipiq×sipdv) / det; det = sip2 × siq2 − sipiq²; In the formula, det is the determinant of the coefficient matrix; When reactance is unobservable, the observed resistance value Rmeasured is calculated using the following formula, without correcting the equivalent line reactance in this round: Rmeasured = sipdv / sip2; The current equivalent line resistance and equivalent line reactance are corrected using an exponentially weighted moving average algorithm based on the observed resistance and reactance values.

[0033] In the implementation of this application, the least squares method is used to calculate the resistance and reactance. For each difference sample point, the physical equation is: ΔVi = ΔIpi × R + ΔIqi × X, where ΔVi is the voltage change at the i-th time point, ΔIpi is the active current change at the i-th time point, ΔIqi is the reactive current change at the i-th time point, R is the resistance, and X is the reactance. This is an equation with two unknowns, and multiple sample points together form an overdetermined system of equations, which does not have an exact solution. The goal of the least squares method is to find a set of R and X that minimizes the sum of squared errors for all sample points. This is achieved by taking the partial derivatives of R and X and setting the partial derivatives to zero, thus obtaining the normal system of equations.

[0034] In this embodiment, storing the original data of all sample points would require storing 3n floating-point numbers, consuming a huge amount of memory in the embedded system. Therefore, the cumulative method used in this application compresses the data into six values ​​by accumulating six statistics in real time during the sampling process: total number of samples (count), active power sum of squares (sip2=Σ(ΔIp²), reactive power sum of squares (siq2=Σ(ΔIq²), active and reactive power product sum (sipiq=Σ(ΔIp×ΔIq), active voltage product sum (sipdv=Σ(ΔV×ΔIp), and reactive voltage product sum (siqdv=Σ(ΔV×ΔIq)). This eliminates the need to store the original data. When X is observable, the coefficient matrix is ​​non-singular, the normal equation system has a unique solution, and the two-dimensional equation system can be solved using the Cramer rule described above. When X is unobservable, the reactive power dimension is abandoned, and only the univariate relation ΔV=ΔIp×R is used to solve the equation using one-dimensional least squares. X is not updated, and historical values ​​are preserved. Then, the current resistance and reactance are corrected using an exponentially weighted moving average algorithm. It should be understood that if the reactance is determined to be unobservable, then no reactance correction is performed in this round.

[0035] In one possible implementation, the determination of whether the reactance is observable includes: Calculate the average value of the sum of squares of reactive power avgiq2=siq2 / count. If avgiq2 is less than the first preset threshold, the reactance is determined to be unobservable. If the absolute value of the determinant of the coefficient matrix is ​​less than or equal to the second preset threshold, the matrix is ​​determined to be singular and the reactance is unobservable.

[0036] In the implementation of this application, a specific condition is provided for judging whether the reactance is "obvious". This is the basic condition for deciding whether to use a two-dimensional or two-dimensional degraded solution. Since ΔIq is squared, summed, and then averaged, avgiq2 reflects the energy intensity of the reactive current change. However, if the average change of ΔIq is too small, such as less than 0.5A, this signal may be overwhelmed by measurement noise, resulting in a low signal-to-noise ratio. Therefore, in this case, attempting to solve X yields a noise value rather than the true reactance value. In this case, the first preset threshold of 0.25 corresponds to |ΔIq|≈0.5A. In the implementation of this application, matrix non-singularity determination is also required. When det is close to 0, the matrix is ​​close to singular, and the normal equations are almost degraded. Small measurement errors can cause huge numerical fluctuations in the solution results. In this case, the second preset threshold is generally 1×10. -6 .

[0037] In one possible implementation, correcting the current equivalent line resistance and equivalent line reactance based on resistance and reactance observations using an exponentially weighted moving average algorithm includes: The corrected equivalent resistance and equivalent reactance of the line are calculated using the following formula: Rnew=Rold×(1-wR)+Rmeasured×wR; Xnew=Xold×(1-wX)+Xmeasured×wX; In the formula, Rnew is the corrected equivalent line resistance, Xnew is the corrected equivalent line reactance, Rold is the current equivalent line resistance, Xold is the current equivalent line reactance, wR is the correction weight of the equivalent resistance corresponding to the first electrical quantity data or the second electrical quantity data, and wX is the correction weight of the equivalent reactance corresponding to the first electrical quantity data or the second electrical quantity data.

[0038] In the implementation of this application, a specific process for correcting an exponentially weighted moving average algorithm is provided. This algorithm is essentially a weighted sum of all historical measurements. During this continuous process, the weight of each historical measurement decays exponentially over time, allowing recent measurements to better reflect the current line status than longer-term measurements. Generally, the wR for active disturbances is 0.3, and the wR for passive responses is 0.15; the wX for active disturbances is 0.05 or 0.02, and the wX for passive responses is 0.02 or 0.01.

[0039] In one possible implementation, calculating the intrinsic grid voltage includes: When the measured voltage is invalid, the intrinsic grid voltage of the previous round is used as the intrinsic grid voltage calculated in this round. When both the measured reactive current and the measured active current are less than the preset value, the measured voltage will be taken as the intrinsic grid voltage. When the target governance device is currently in standby mode, the intrinsic grid voltage is calculated according to the following formula: Vn = Vm − Iq × Xnew; In the formula, Vn is the intrinsic grid voltage, Vm is the measured voltage, Iq is the measured reactive current, and Xnew is the corrected line equivalent reactance. When the target treatment device is currently in a charging state, the intrinsic grid voltage is calculated according to the following formula: Vn = Vm + Ip × Rnew − Iq × Xnew; In the formula, Ip is the measured active current, and Rnew is the corrected line equivalent resistance; When the target treatment device is currently in a discharge state, the intrinsic grid voltage is calculated according to the following formula: Vn = Vm - Ip × Rnew − Iq × Xnew.

[0040] In the implementation of this application, three differentiated back-calculation methods for intrinsic voltage under different operating conditions are provided. The intrinsic grid voltage is the original voltage of the upstream grid of the device. The charging and discharging current of the device generates a voltage drop between the measurement point and the upstream grid. Therefore, the measured voltage is the result of the intrinsic grid voltage superimposed with this voltage drop. The essence of intrinsic voltage back-calculation is to remove this voltage drop from the measured voltage. It should be understood that this voltage drop has positive and negative directions.

[0041] In this embodiment of the application, when the measured voltage is lower than 10V, it is likely that the measurement circuit is faulty. At this time, any calculation based on the measured voltage is unreliable. In this case, the intrinsic voltage of the previous round should be used as the result of the current round of calculation, which is equivalent to freezing the effective value of the previous moment and waiting for the measurement to recover.

[0042] When both the measured reactive current and the measured active current are less than the preset values, such as when the measured active current is less than 0.5A and the measured reactive current is less than 0.5A, the influence of the device itself on the voltage at the measurement point can be ignored. In this case, the measured voltage can be directly taken as the intrinsic grid voltage. If the voltage drop compensation formula is forcibly substituted, the voltage drop value under small current is very small and its proportion is less than the measurement noise, which may introduce errors.

[0043] In this application embodiment, the calculation process for three different operating conditions is expressed. In the standby state, the device outputs zero power and the active current is basically 0. However, reactive current still flows through the reactance of internal filter capacitors and other components. The capacitive reactive current raises the measured voltage, so Iq×X needs to be deducted. In the charging state, the device absorbs active power from the grid, and the current flows from the grid to the device. The measured voltage is pulled down, so Ip×R needs to be added back. At the same time, the reactive current raises the measured voltage, so Iq×X needs to be deducted. In the discharging state, the device injects active power into the grid, and the current flows from the device to the grid. The measured voltage is raised, so Ip×R needs to be deducted. At the same time, the reactive current raises the measured voltage, so Iq×X needs to be deducted.

[0044] In one possible implementation, the calculation of the proportional gain, integral gain, and derivative gain, as well as the updating of the control parameters, include: Calculate the proportional gain using the following formula: Kp = α × 110 / Rnew; In the formula, Kp is the proportional gain, α is the safety margin, and Rnew is the corrected line equivalent resistance. The integral gain is calculated using the following formula: Ki = β × Kp; In the formula, Ki is the integral gain, and β is the integral coefficient; The differential gain is calculated using the following formula: Kd = γ × Kp; In the formula, Kd is the differential gain, and β is the differential coefficient; The calculated proportional gain Kp is clamped between a preset lower limit and an upper limit of the proportional gain; The final proportional gain, integral gain, and derivative gain are used as the current control parameters.

[0045] When implemented in this application embodiment, the injected power P in a purely resistive circuit device will cause a voltage change ΔV: ΔV = (P / Vnominal) × R; Where Vnominal is the rated voltage of the power grid, the power command generated in closed-loop control is Pp = Kp × ΔV_error, where ΔV_error is the difference between the target voltage and the actual voltage. When Kp × R / Vnominal = 1, the deviation can be eliminated in one cycle (single control cycle), and the system is at the critical oscillation boundary. Therefore, the critical gain is Kpcritical = Vnominal / R. For a 220V system, considering the relationship between the effective voltage value and the peak voltage value and the definition of bipolar error, the theoretical critical gain is 220 / R. To prevent the voltage from exceeding the rated value during the adjustment process, half of this value (i.e., 110 / R) is taken. Based on this, a safety margin of 60% is taken to cope with line impedance error and control delay. Finally, Kp = 0.6 × 110 / Rnew.

[0046] In this embodiment, the integral term is used to eliminate steady-state error, and the derivative term is used to predict the trend of deviation changes and suppress overshoot and oscillation. In engineering experience, the integral time constant is approximately 3 to 4 times the proportional term time constant, therefore Ki≈Kp / 3≈0.3×Kp. The derivative time constant is approximately 0.1 to 0.2 times the proportional term time constant, therefore Kd≈0.1×Kp. This proportional relationship ensures that the relative weights of the three terms of the controller remain constant under different line impedance conditions, resulting in consistent dynamic characteristics of the control system.

[0047] In this embodiment, when the resistance is extremely small, such as 0.3Ω, Kp=66 / 0.3=220, which far exceeds the stability margin. In this case, the output power is clamped to 150 to limit the maximum output power and prevent oscillation. However, when R is extremely large, such as 10Ω, Kp=66 / 10=6.6. Below this value, the controller response is too sluggish, so the clamping is set to 6.6 to ensure minimum response capability. The clamping is essentially a soft safety constraint set according to the physical limitations of the actuator and the stability requirements of the control system.

[0048] In one possible implementation, obtaining the control parameters further includes: Control parameters are obtained based on the following priority: First priority: When a non-zero control parameter is manually set, the manual value is used directly as the current control parameter; Second priority: When the device starts up, the learned value of the control parameter corresponding to the direction is loaded as the current control parameter; the direction includes the charging direction and the discharging direction; Third priority: Use the calculated proportional gain, integral gain, and derivative gain as the current control parameters; Fourth priority: When the device is powered on for the first time, the impedance is not calibrated, and the learning value has not been established, the default value is used as the current control parameter.

[0049] In the implementation of this application embodiment, a priority decision chain containing four control parameter sources is provided. The first priority is the manually set value. When the user sets a non-zero PID parameter through the DLT645 protocol or debugging interface, the manual value is used directly. This ensures that on-site debugging personnel can override the automatic tuning result at any time, which is suitable for emergency situations or special operating conditions. During the manual mode, the automatic tuning continues to be calculated in the background but is not output. It will automatically resume after the manual flag is cleared.

[0050] The second priority is to persist the learned values, which are loaded from memory when the device starts up. This ensures that even if impedance identification is not yet complete, the device can use the optimal parameters accumulated in the previous operating cycle, thus shortening the transition time from startup to the optimal control state.

[0051] The third priority is the online identification value, which is the PID parameters derived online based on the currently identified R. This is the default operating mode of the device, and the parameters are automatically updated as the line impedance drifts, without the need for manual interaction.

[0052] The fourth priority is the cold start default value, i.e., the factory preset parameters; this ensures that the controller has available parameters when it is first powered on, and will not fail to start due to lack of parameters. Typically, it is Kp=15, Ki=4.5, Kd=1.5, corresponding to a medium impedance circuit with R=4.4Ω.

[0053] In one possible implementation, obtaining the learned values ​​includes: In each control cycle, read the actual effective proportional gain; When the target governance device is currently in a charging state, the current proportional gain is stored in the sliding window queue in the charging direction; when the target governance device is currently in a discharging state, the current proportional gain is stored in the sliding window queue in the discharging direction. When the number of data in the sliding window queue in the charging direction reaches the length of the preset window, the arithmetic mean of all proportional gains in the sliding window queue in the charging direction is calculated as the first sliding average. When the number of data in the sliding window queue in the discharge direction reaches the length of the preset window, the arithmetic mean of all proportional gains in the sliding window queue in the discharge direction is calculated as the second sliding average. Using a preset evaluation period, the first moving average value is compared with the learned value of the stored charging direction, and the second moving average value is compared with the learned value of the stored discharging direction. When the difference between the first moving average and the learned value of the stored charging direction exceeds a preset update threshold, the first moving average is used as the learned value of the new charging direction; when the difference between the second moving average and the learned value of the stored discharging direction exceeds a preset update threshold, the second moving average is used as the learned value of the new discharging direction.

[0054] In the implementation of this application embodiment, a specific process for generating learning values ​​is provided, wherein each control cycle reads the actual effective proportional gain Kp, and then, according to the current charging or discharging state of the device, stores the Kp value in the corresponding direction's sliding window queue, and then performs a moving average calculation. If the difference between the moving average and the previous persistent value exceeds a preset threshold, such as 3.0, then the corresponding direction's learning value in the persistent storage is updated.

[0055] Based on the same inventive concept, this application also provides an intrinsic voltage estimation and parameter self-tuning system for low-voltage management, comprising: The acquisition unit is configured to periodically acquire electrical quantity data of the target treatment device; the electrical quantity data is divided into first electrical quantity data acquired through active power perturbation and second electrical quantity data acquired through passive operation response; The weighting unit is configured to correct the current line equivalent resistance and line equivalent reactance based on the electrical quantity data using an exponentially weighted moving average algorithm; the correction weight of the first electrical quantity data is greater than the correction weight of the second electrical quantity data; The reverse calculation unit is configured to select a corresponding reverse calculation strategy based on the current operating state of the target governance device, and calculate the intrinsic grid voltage based on electrical quantity data, the corrected line equivalent resistance, and the corrected line equivalent reactance. The gain unit is configured to generate the proportional gain of the closed-loop controller by using the corrected equivalent resistance of the line through a preset inverse proportional relationship, and to generate integral gain and derivative gain according to the proportional gain at a preset ratio. The update unit is configured to update the control parameters by calculating the proportional gain, integral gain, and derivative gain.

[0056] For example, a specific example of an active pulse calibration process is provided: A low-voltage control device is deployed at the end of a rural power grid. The factory default impedance is R=4.0Ω and X=0. On the first day of operation, the device detected stable voltage at 3:00 AM: the standard deviation of the voltage was less than 0.5V for 60 consecutive seconds, automatically triggering active pulse calibration.

[0057] After calibration is initiated, the state machine executes the following steps in sequence: 1. PCS enters standby mode, confirmation complete in 2 seconds; 2. A baseline of zero samples was acquired for 2 seconds, with an average Vm≈217.3V; 3. Eight sets of charging pulses are injected sequentially, with power values ​​of 2000W, 1600W, 1000W, 2200W, 1600W, 2000W, 1000W, and 2200W respectively. Each pulse lasts 1.5 seconds. The first four pulses use a single converter with a reactive current component of approximately 1.2A, while the latter four pulses use a dual converter with a reactive current component of approximately 2.4A, creating a ΔIq difference. 4. Recover for 1 second after each pulse; 5. During the settlement phase, the cumulative sample values ​​were sip2=142.3, siq2=18.5, sipiq=21.4, sipdv=312.7, siqdv=38.2, and count=8. 6. Calculate avgiq2 = 2.31 ≥ 0.25, X is observable; 7. Calculate det = 142.3 × 18.5 − 21.4² ≈ 2174.6, the matrix is ​​non-singular; 8. Solving with two-dimensional least squares yields Rmeasured = 2.16Ω and Xmeasured = 0.18Ω; 9. Merge update with default value: Rnew=4.0×0.7+2.16×0.3≈3.45Ω, Xnew=0×0.95+0.18×0.05=0.009Ω.

[0058] Subsequent calibrations continued to push R towards 2.16Ω and X towards 0.18Ω, gradually establishing a long-term memory of the deployed circuitry.

[0059] For example, here is a specific example of passive charging ramp-up acquisition of supplementary impedance: At 06:30 one day, the device entered the charging backup state before the morning rush hour, and the power naturally climbed from 0W to 4000W.

[0060] Within 5 seconds, 10 difference points were accumulated, indicating weak sample observability (avgiq2 = 0.18 < 0.25). Therefore, the result was automatically downgraded to one-dimensional least squares, yielding Rmeasured = 2.32Ω. X was not updated, maintaining the historical value of 0.18Ω.

[0061] The passive sample is fused and updated with a weight of 0.15: Rnew = 2.16 × 0.85 + 2.32 × 0.15 ≈ 2.18Ω.

[0062] For example, a specific example of the influence of the intrinsic voltage stripping device itself is provided: The device operates in charging mode, with measured voltage Vm = 215.3V, active current Ip = 8.2A, reactive current Iq = 0.8A, effective impedance R = 2.18Ω, and X = 0.18Ω. The charging formula can be deduced by working backwards: Vn=215.3+8.2×2.18−0.8×0.18≈233.03V The actual voltage of the upstream intrinsic grid is 233.03V, while the voltage at the measurement point is pulled down to 215.3V by the device absorbing 8.2A of active current. If 215.3V is used directly as the historical sample, it will be mistakenly judged that the grid voltage is "too low and needs to be addressed" at that moment, resulting in historical learning pollution.

[0063] Switching to discharge mode, Vm=218.0V, Ip=6.5A, Iq=0.6A, deducing backwards: Vn=218.0−6.5×2.18−0.6×0.18≈203.72V That is, the upstream intrinsic voltage is actually 203.72V, and the device injects 6.5A of active current to raise the measurement point to 218.0V. This example shows that the same device can obtain an accurate intrinsic voltage in both charging and discharging states.

[0064] For example, a specific example of impedance-driven PID self-tuning and learning is provided: Calculated based on R=2.18Ω: Kp = 66 / 2.18 ≈ 30.3, which is within the range of [6.6, 150], so clamping is not required; Ki = 30.3 × 0.3 ≈ 9.1; Kd = 30.3 × 0.1 ≈ 3.0.

[0065] The controller continues to run with these parameters, saving the Kp moving average every 5 minutes. At the 30th minute, the accumulated Kp moving average over 6 times is 30.5, which deviates from the persistent learning value of 26.0 by 4.5, exceeding the threshold of 3.0, triggering a persistent update, and the new value of 30.5 is written to non-volatile storage.

[0066] After the device restarts the next day, it restores 30.5 from persistent storage. In the early stages before impedance identification is completed, the learned value is used first, shortening the time from startup to optimal control state.

[0067] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0068] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices or units, or may be electrical, mechanical or other forms of connection.

[0069] The units described as separate components may or may not be physically separate. As will be apparent to those skilled in the art, the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0070] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0071] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or grid device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0072] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for estimating intrinsic voltage and self-tuning parameters for low-voltage management, characterized in that, include: The electrical quantity data of the target treatment device is acquired periodically; the electrical quantity data is divided into first electrical quantity data acquired through active power perturbation and second electrical quantity data acquired through passive operation response. The current line equivalent resistance and line equivalent reactance are corrected based on the electrical quantity data using an exponentially weighted moving average algorithm; the correction weight of the first electrical quantity data is greater than the correction weight of the second electrical quantity data. Select the corresponding reverse strategy based on the current working state of the target governance device, and calculate the intrinsic grid voltage based on electrical quantity data, the corrected line equivalent resistance, and the corrected line equivalent reactance. The corrected equivalent resistance of the line is used to generate the proportional gain of the closed-loop controller through a preset inverse proportional relationship. The integral gain and derivative gain are then generated according to the preset ratio based on the proportional gain. The control parameters are updated by calculating the proportional gain, integral gain, and derivative gain.

2. The intrinsic voltage estimation and parameter self-tuning method for low-voltage management according to claim 1, characterized in that, The acquisition of the first electrical quantity data includes: During off-peak hours at night, the energy storage converter is instructed to enter standby mode, and after the energy storage converter is in standby mode, the measured voltage, active current component and reactive current component are collected as baseline electrical quantity data. Multiple sets of power pulses with different pulse powers are injected sequentially, and the corresponding measured voltage, active current component, and reactive current component are collected as steady-state electrical quantity data after each set of power pulses is injected and reaches a steady state. The multiple sets of power pulses with different pulse powers include forward pulses and reverse pulses. The injection process includes single converter injection and multi-converter parallel injection. The difference between each set of steady-state electrical quantity data and the corresponding baseline electrical quantity data is calculated as the first electrical quantity data; The acquisition of the second electrical quantity data includes: During off-peak hours outside of nighttime electricity consumption, electrical quantity data is continuously sampled at a preset cycle. The second electrical quantity data is obtained by differentiating the electrical quantity data of the current sampling point from the electrical quantity data of the previous sampling point.

3. The intrinsic voltage estimation and parameter self-tuning method for low-voltage management according to claim 1, characterized in that, The electrical quantity data includes voltage difference ΔV, active current difference ΔIp, and reactive current difference ΔIq; Correcting the current line equivalent resistance and line equivalent reactance based on the aforementioned electrical quantity data includes: Based on the electrical quantity data, obtain the total number of samples count, active square sum sip2=Σ(ΔIp²), reactive square sum siq2=Σ(ΔIq²), active and reactive product sum sipiq=Σ(ΔIp×ΔIq), active voltage product sum sipdv=Σ(ΔV×ΔIp), and reactive voltage product sum siqdv=Σ(ΔV×ΔIq); When reactance is observable, the resistance observation Rmeasured and the reactance observation Xmeasured are calculated according to the following formulas: Rmeasured=(siq2×sipdv−sipiq×siqdv) / det; Xmeasured=(sip2×siqdv−sipiq×sipdv) / det; det = sip2 × siq2 − sipiq²; In the formula, det is the determinant of the coefficient matrix; When reactance is unobservable, the observed resistance value Rmeasured is calculated using the following formula, without correcting the equivalent line reactance in this round: Rmeasured = sipdv / sip2; The current equivalent line resistance and equivalent line reactance are corrected using an exponentially weighted moving average algorithm based on the observed resistance and reactance values.

4. The intrinsic voltage estimation and parameter self-tuning method for low-voltage management according to claim 3, characterized in that, The determination of whether reactance is observable includes: Calculate the average value of the sum of squares of reactive power avgiq2=siq2 / count. If avgiq2 is less than the first preset threshold, the reactance is determined to be unobservable. If the absolute value of the determinant of the coefficient matrix is ​​less than or equal to the second preset threshold, the matrix is ​​determined to be singular and the reactance is unobservable.

5. The intrinsic voltage estimation and parameter self-tuning method for low-voltage management according to claim 3, characterized in that, The current equivalent line resistance and equivalent line reactance are corrected using an exponentially weighted moving average algorithm based on resistance and reactance observations, including: The corrected equivalent resistance and equivalent reactance of the line are calculated using the following formula: Rnew=Rold×(1-wR)+Rmeasured×wR; Xnew=Xold×(1-wX)+Xmeasured×wX; In the formula, Rnew is the corrected equivalent line resistance, Xnew is the corrected equivalent line reactance, Rold is the current equivalent line resistance, Xold is the current equivalent line reactance, wR is the correction weight of the equivalent resistance corresponding to the first electrical quantity data or the second electrical quantity data, and wX is the correction weight of the equivalent reactance corresponding to the first electrical quantity data or the second electrical quantity data.

6. The intrinsic voltage estimation and parameter self-tuning method for low-voltage management according to claim 1, characterized in that, Calculating the intrinsic grid voltage includes: When the measured voltage is invalid, the intrinsic grid voltage of the previous round is used as the intrinsic grid voltage calculated in this round. When both the measured reactive current and the measured active current are less than the preset value, the measured voltage will be taken as the intrinsic grid voltage. When the target governance device is currently in standby mode, the intrinsic grid voltage is calculated according to the following formula: Vn = Vm − Iq × Xnew; In the formula, Vn is the intrinsic grid voltage, Vm is the measured voltage, Iq is the measured reactive current, and Xnew is the corrected line equivalent reactance. When the target treatment device is currently in a charging state, the intrinsic grid voltage is calculated according to the following formula: Vn = Vm + Ip × Rnew − Iq × Xnew; In the formula, Ip is the measured active current, and Rnew is the corrected line equivalent resistance; When the target treatment device is currently in a discharge state, the intrinsic grid voltage is calculated according to the following formula: Vn = Vm - Ip × Rnew − Iq × Xnew.

7. The intrinsic voltage estimation and parameter self-tuning method for low-voltage management according to claim 1, characterized in that, The calculation of proportional gain, integral gain, and derivative gain, as well as the updating of control parameters, include: Calculate the proportional gain using the following formula: Kp = α × 110 / Rnew; In the formula, Kp is the proportional gain, α is the safety margin, and Rnew is the corrected line equivalent resistance. The integral gain is calculated using the following formula: Ki = β × Kp; In the formula, Ki is the integral gain, and β is the integral coefficient; The differential gain is calculated using the following formula: Kd = γ × Kp; In the formula, Kd is the differential gain, and β is the differential coefficient; The calculated proportional gain Kp is clamped between a preset lower limit and an upper limit of the proportional gain; The final proportional gain, integral gain, and derivative gain are used as the current control parameters.

8. The intrinsic voltage estimation and parameter self-tuning method for low-voltage management according to claim 1, characterized in that, The acquisition of the control parameters also includes: Control parameters are obtained based on the following priority: First priority: When a non-zero control parameter is manually set, the manual value is used directly as the current control parameter; Second priority: When the device starts up, the learned value of the control parameter corresponding to the direction is loaded as the current control parameter; the direction includes the charging direction and the discharging direction; Third priority: Use the calculated proportional gain, integral gain, and derivative gain as the current control parameters; Fourth priority: When the device is powered on for the first time, the impedance is not calibrated, and the learning value has not been established, the default value is used as the current control parameter.

9. The intrinsic voltage estimation and parameter self-tuning method for low-voltage management according to claim 8, characterized in that, The acquisition of the learning values ​​includes: In each control cycle, read the actual effective proportional gain; When the target governance device is currently in a charging state, the current proportional gain is stored in the sliding window queue in the charging direction; when the target governance device is currently in a discharging state, the current proportional gain is stored in the sliding window queue in the discharging direction. When the number of data in the sliding window queue in the charging direction reaches the length of the preset window, the arithmetic mean of all proportional gains in the sliding window queue in the charging direction is calculated as the first sliding average. When the number of data in the sliding window queue in the discharge direction reaches the length of the preset window, the arithmetic mean of all proportional gains in the sliding window queue in the discharge direction is calculated as the second sliding average. Using a preset evaluation period, the first moving average value is compared with the learned value of the stored charging direction, and the second moving average value is compared with the learned value of the stored discharging direction. When the difference between the first moving average and the learned value of the stored charging direction exceeds a preset update threshold, the first moving average is used as the learned value of the new charging direction; when the difference between the second moving average and the learned value of the stored discharging direction exceeds a preset update threshold, the second moving average is used as the learned value of the new discharging direction.

10. An intrinsic voltage estimation and parameter self-tuning system for low-voltage management using the method of any one of claims 1 to 9, characterized in that, include: The acquisition unit is configured to periodically acquire electrical quantity data of the target treatment device; The electrical quantity data is divided into first electrical quantity data obtained through active power perturbation and second electrical quantity data obtained through passive operation response. The weighting unit is configured to correct the current line equivalent resistance and line equivalent reactance based on the electrical quantity data using an exponentially weighted moving average algorithm; The correction weight of the first electrical quantity data is greater than the correction weight of the second electrical quantity data; The reverse calculation unit is configured to select a corresponding reverse calculation strategy based on the current operating state of the target governance device, and calculate the intrinsic grid voltage based on electrical quantity data, the corrected line equivalent resistance, and the corrected line equivalent reactance. The gain unit is configured to generate the proportional gain of the closed-loop controller by using the corrected equivalent resistance of the line through a preset inverse proportional relationship, and to generate integral gain and derivative gain according to the proportional gain at a preset ratio. The update unit is configured to update the control parameters by calculating the proportional gain, integral gain, and derivative gain.

Images