Full-effect electric energy optimization rapid detection system and method thereof

Abstract

The invention discloses a full-effect electric energy optimization rapid detection system and method, and particularly relates to the technical field of electric energy detection. The method comprises the following steps: acquiring voltage, current, frequency, harmonic component and three-phase unbalance degree of a power grid in real time, calculating a power grid impedance value and a load dynamic characteristic ratio, drawing an impedance frequency characteristic curve, analyzing a harmonic amplification abnormal degree and a load dynamic response characteristic, and evaluating the accuracy of a detection instruction generated by the full-effect electric energy optimization device. The error correction is carried out on the inaccuracy detection instruction, the duty ratio is dynamically adjusted by utilizing the PWM technology, the accurate matching of the amplitude and the phase of the output current is realized, the error correction effect is evaluated by comparing the real-time detection value with the theoretical calculation value, and the detection model and the compensation strategy are automatically optimized. The problems of detection precision reduction and error compensation caused by complex power grid conditions in the prior art are solved, harmonic amplification, reactive power over-compensation and three-phase balance damage are avoided, and efficient and accurate electric energy optimization and power grid stability improvement are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of electric energy detection, and in particular to a full-effect electric energy optimization rapid detection system and a method thereof. Background Art

[0002] With the rapid development of modern power systems and the increasing complexity of power loads, power quality issues are becoming increasingly prominent, manifested in insufficient reactive power, harmonic interference, three-phase imbalance, voltage fluctuations and flickers, etc. These problems not only affect the operating efficiency and reliability of the power grid, but also have a negative impact on the normal operation of user equipment, increase the energy loss of the distribution system, and may even cause equipment failure or shutdown.

[0003] To solve the above problems, a series of power optimization devices have gradually appeared on the market, among which KEML500 full-effect power optimization device has become an effective solution with its modular design, high reliability and excellent performance. The device can quickly compensate for reactive power, filter harmonics, improve three-phase imbalance, and suppress voltage fluctuations and flickers. Its technical characteristics have shown significant advantages in practical applications.

[0004] The prior art has the following deficiencies: Under complex grid conditions, such as multi-point grounding, excessive nonlinear loads, or sensor failures in the detection system, the device may lose accuracy in detecting reactive power, harmonic current, or voltage fluctuations, resulting in erroneous compensation instructions. This miscompensation may lead to harmonic amplification, reactive power overcompensation, or even destruction of three-phase balance, further deteriorating the power quality. In addition, miscompensation behaviors (such as harmonic amplification, reactive power overcompensation, or destruction of three-phase balance) will cause further deterioration of grid parameters, such as increased harmonic currents and reactive loads exceeding the processing capacity of the equipment. This deteriorated grid state may exceed the design range of the detection algorithm, resulting in the inability of the detection algorithm to accurately separate harmonic components, determine reactive demand, or three-phase load distribution, thus falling into a vicious feedback loop of "inaccurate detection-miscompensation-grid deterioration." Summary of the invention

[0005] The purpose of the present invention is to provide a full-effect electric energy optimization rapid detection system and method thereof to solve the deficiencies in the background technology.

[0006] In order to achieve the above object, the present invention provides the following technical solution: a full-effect electric energy optimization rapid detection method, comprising the following steps: S1: The electrical parameters of the power grid, including voltage, current, frequency, harmonic components and three-phase imbalance, are collected in real time through high-precision sensors installed on the power grid busbar; S2: Inject a small disturbance current into the power grid within the specified frequency range, measure the voltage response caused by the disturbance current, calculate the impedance value of the power grid at each frequency point, and calculate the ratio of the load current change rate to the voltage change rate by measuring the instantaneous change of the load current and voltage; S3: Draw the impedance frequency characteristic curve according to the calculated impedance value at each frequency point, analyze the abnormal degree of harmonic amplification of the power grid, and judge the dynamic response of the load under different working conditions according to the calculated ratio of the load current change rate to the voltage change rate; S4: According to the abnormal degree of harmonic amplification of the power grid and the dynamic response of the load under different working conditions, the accuracy of the detection instructions generated by the full-efficiency power optimization device is evaluated, and the detection instructions are divided into accurate detection instructions and inaccurate detection instructions according to the evaluation results; S5: Error correction is performed on the inaccurate detection instructions generated by the full-efficiency power optimization device, and the duty cycle is dynamically adjusted according to the error amplitude and phase deviation using PWM technology, so that the amplitude and phase of the output current match the target instruction; S6: By comparing the real-time values ​​of the detected harmonic current, reactive power and load current with the theoretical calculated values, the error correction effect is evaluated. If the error correction effect is poor, the detection model and compensation strategy are automatically adjusted.

[0007] Preferably, in S3, the impedance amplitude abnormality index is generated after analyzing the curve of impedance amplitude changing with frequency, and the impedance amplitude abnormality index is obtained by: Set the original signal of impedance amplitude to x(t), select wavelet function ψ(t) and decomposition level J, perform discrete wavelet transform on x(t), and the wavelet transform coefficients The calculation expression is: ;in, is the wavelet function after translation and scale transformation, a is the scale parameter, b is the translation parameter; the approximate coefficient and detail coefficient obtained by wavelet decomposition are expressed as: ;in, and are the scaling function and the wavelet function, respectively. is the approximate coefficient, is the detail coefficient; reconstruct the signal components at each level, analyze the frequency components of the reconstructed signal, extract the main frequency components and their amplitudes within the passband, ; In the formula, is the minimum frequency value of the frequency range, is the maximum frequency value of the frequency range; To reconstruct the signal, identify the frequency with the largest amplitude in the spectrum As the main frequency, it represents the frequency component with the largest amplitude in the signal spectrum, and records the corresponding amplitude Represents the amplitude of the main frequency, for each frequency component in the passband Amplitude Calculate the mean and standard deviation , Refers to the i-th frequency component in the signal spectrum, and calculates the impedance amplitude anomaly index. The expression is: ; HK is the impedance amplitude anomaly index.

[0008] Preferably, in S3, a load fluctuation index is generated after analyzing the change in the ratio of the load current change rate to the voltage change rate. The load fluctuation index is obtained by: Obtain current change rate from real-time monitoring of the load and voltage change rate The time series data is in the form of an N×2 matrix, where N is the number of time points and the first column is , the second column is : ; Standardize each column of data in X: ;in, is the mean of each column of data, is the standard deviation of each column of data, calculate the standardized data matrix, ; C is a 2×2 matrix that describes and The variance and covariance between them, T is the matrix transpose, calculate the eigenvalues ​​λ1, λ2 of the covariance matrix C and the corresponding eigenvectors v1, v2, the expression is: Cv=λv; the eigenvalue represents the variance of the data in the direction of the corresponding eigenvector; Sort the eigenvalues ​​so that λ1≥λ2, the corresponding eigenvectors are v1 and v2, and the standardized data Project to the principal component direction v1 to get the principal component score : is the change component of load fluctuation, and the load fluctuation index is defined as the absolute mean of the principal component scores: ; Where LFI is the load fluctuation index.

[0009] Preferably, in S4, the accuracy of the detection instructions generated by the full-effect electric energy optimization device is evaluated according to the abnormal degree of harmonic amplification in the power grid and the dynamic response of the load under different working conditions, specifically: the impedance amplitude abnormality index and the load fluctuation index are converted into a comprehensive feature vector, and the comprehensive feature vector is used as the input of the machine learning model. The machine learning model uses each group of comprehensive feature vectors to predict the accuracy value label of the detection instructions generated by the full-effect electric energy optimization device as the prediction target, and takes minimizing the sum of prediction errors of the accuracy value labels of the detection instructions generated by all full-effect electric energy optimization devices as the training target. The machine learning model is trained until the sum of prediction errors converges and the model training is stopped. The accuracy value of the detection instructions generated by the full-effect electric energy optimization device is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0010] Preferably, the acquired accuracy value of the detection instruction generated by the full-effect electric energy optimization device is compared with the accuracy reference threshold preset according to historical data. If the accuracy value of the detection instruction generated by the full-effect electric energy optimization device is greater than or equal to the preset accuracy reference threshold, it means that the accuracy of the detection instruction generated by the full-effect electric energy optimization device is high. In this case, no warning signal is generated, and the detection instruction is classified as an accuracy detection instruction. If the accuracy value of the detection instruction generated by the full-effect electric energy optimization device is less than the preset accuracy reference threshold, it means that the accuracy of the detection instruction generated by the full-effect electric energy optimization device is low. In this case, a warning signal is generated, and the detection instruction is classified as an inaccuracy detection instruction.

[0011] Preferably, in S5, error correction is performed on the inaccuracy detection instruction generated by the full-efficiency power optimization device, and the duty cycle is dynamically adjusted according to the error amplitude and phase deviation using PWM technology, so that the amplitude and phase of the output current match the target instruction, specifically: Detect the actual output current amplitude Ioutput and the target command amplitude Itarget to calculate the amplitude error : ; If eamp>0, the output current amplitude is insufficient and the compensation needs to be increased; if eamp<0, the output current amplitude is too large and the compensation needs to be reduced; Detecting the phase of output current and target command phase Calculating Phase Deviation : ;like , the output current lags behind the target command, and the phase advance needs to be adjusted; if , the output current is ahead of the target instruction, and the phase delay needs to be adjusted.

[0012] Preferably, the PWM technique changes the amplitude and phase of the output current by adjusting the duty cycle D, where the duty cycle is defined as the ratio of the duration of the high level in one cycle: ; Where: Ton is the high level duration of the PWM signal, T is the total period of the PWM signal; According to the amplitude error eamp, adjust the PWM duty cycle Damp, the expression is: ; is the initial duty cycle; kamp is the amplitude adjustment coefficient, which is set according to the device response characteristics; the adjusted duty cycle directly affects the output current amplitude: ; According to the phase deviation , adjust the PWM switching delay time , the expression is: :like , turn on the PWM switch in advance; if , delay turning on the PWM switch; Use the current sensor to collect the adjusted output current signal in real time to obtain the corrected amplitude Ioutput,new and phase ,new; recalculate the amplitude error eamp,new and phase deviation : ; ; Adjust the PWM duty cycle and delay time again according to the new error value to form a closed-loop control until the error converges to the set threshold.

[0013] Preferably, in S6, sensors and sampling devices are used to collect grid parameters in real time: harmonic current Iharmonic, detected: real-time amplitude of harmonic components in the grid; reactive power Qdetected: real-time reactive power of the system; load current Iload, detected: real-time current value at the load end; According to the system operating status and compensation strategy, calculate the theoretical value of harmonic current Iharmonic,theoretical, the theoretical value of reactive power Qtheoretical and the theoretical value of load current Iload,theoretical; Compare the real-time value with the theoretical value, calculate the error, and define the allowable error range; if all errors meet the threshold, the correction effect is good and there is no need to adjust the model and compensation strategy; if any error exceeds the threshold, the correction effect is poor and the detection model and compensation strategy need to be automatically adjusted.

[0014] Preferably, if the harmonic error is large, check whether the harmonic compensation instruction is accurate, optimize the compensation current amplitude and phase; dynamically adjust the compensation frequency band and harmonic order of the harmonic filter; If the reactive power error is large, check whether the inductive or capacitive reactive compensation is over-compensated or under-compensated; adjust the real-time amount of reactive current injection; If the load current error is large, check the three-phase current balance, optimize the compensation current distribution ratio, and dynamically adjust the compensation response time.

[0015] The present invention also provides a full-efficiency electric energy optimization rapid detection system, including a real-time data acquisition module, a power grid characteristic analysis module, a harmonic and load analysis module, a detection instruction accuracy evaluation module, an error correction module and an optimization module: Real-time data acquisition module: through the high-precision sensors installed on the power grid bus, the electrical parameters of the power grid are collected in real time, including voltage, current, frequency, harmonic components and three-phase imbalance; Grid characteristic analysis module: Injects a small disturbance current into the grid within a specified frequency range, measures the voltage response caused by the disturbance current, and calculates the impedance value of the grid at each frequency point, and calculates the ratio of the load current change rate to the voltage change rate by measuring the instantaneous change of the load current and voltage; Harmonic and load analysis module: Draw the impedance frequency characteristic curve according to the calculated impedance value at each frequency point, analyze the abnormal degree of harmonic amplification of the power grid, and judge the dynamic response of the load under different working conditions according to the calculated ratio of the load current change rate to the voltage change rate; Detection instruction accuracy evaluation module: evaluates the accuracy of the detection instructions generated by the full-efficiency power optimization device according to the abnormal degree of grid harmonic amplification and the dynamic response of the load under different working conditions, and divides the detection instructions into accurate detection instructions and inaccurate detection instructions according to the evaluation results; Error correction module: performs error correction on the inaccurate detection instructions generated by the full-effect power optimization device, and uses PWM technology to dynamically adjust the duty cycle according to the error amplitude and phase deviation, so that the amplitude and phase of the output current match the target instruction; Optimization module: By comparing the real-time values ​​of the detected harmonic current, reactive power and load current with the theoretical calculated values, the error correction effect is evaluated. If the error correction effect is poor, the detection model and compensation strategy are automatically adjusted.

[0016] In the above technical solution, the technical effects and advantages provided by the present invention are: 1. The present invention realizes high-precision detection and compensation of power quality problems under complex power grid conditions through a series of optimization methods, overcoming the problems of miscompensation and feedback deterioration in the prior art. By real-time acquisition of voltage, current, frequency, harmonic components and three-phase imbalance, combining wavelet transform and principal component analysis to extract impedance amplitude anomaly index and load fluctuation index, and using machine learning to evaluate the accuracy of detection instructions, the detection accuracy and reliability of instruction execution are significantly improved. In addition, PWM technology is used to perform dynamic error correction on inaccuracy detection instructions to achieve accurate matching of the amplitude and phase of the output current, thereby effectively solving the problems of harmonic amplification, reactive power overcompensation and three-phase balance destruction.

[0017] 2. The present invention automatically optimizes the detection model and compensation strategy. When the error correction effect is poor, the present invention can make adaptive adjustments to ensure the stability and robustness of the detection and compensation performance in a complex power grid environment. This technology not only improves the intelligence level of the power optimization device, but also realizes the efficient management of harmonics, reactive power and three-phase loads, avoids the vicious feedback loop of the power grid caused by compensation errors, improves the overall power quality, and meets the needs of efficient energy saving and stable operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0019] Figure 1 The figure is a flow chart of the method of the present invention.

[0020] Figure 2 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] Example 1, please refer to Figure 1 As shown, the full-effect electric energy optimization rapid detection method described in this embodiment includes the following steps: S1: The electrical parameters of the power grid, including voltage, current, frequency, harmonic components and three-phase imbalance, are collected in real time through high-precision sensors installed on the power grid busbar; S2: Inject a small disturbance current into the power grid within the specified frequency range, measure the voltage response caused by the disturbance current, calculate the impedance value of the power grid at each frequency point, and calculate the ratio of the load current change rate to the voltage change rate by measuring the instantaneous change of the load current and voltage; S3: Draw the impedance frequency characteristic curve according to the calculated impedance value at each frequency point, analyze the abnormal degree of harmonic amplification of the power grid, and judge the dynamic response of the load under different working conditions according to the calculated ratio of the load current change rate to the voltage change rate; S4: According to the abnormal degree of harmonic amplification of the power grid and the dynamic response of the load under different working conditions, the accuracy of the detection instructions generated by the full-efficiency power optimization device is evaluated, and the detection instructions are divided into accurate detection instructions and inaccurate detection instructions according to the evaluation results; S5: Error correction is performed on the inaccurate detection instructions generated by the full-efficiency power optimization device, and the duty cycle is dynamically adjusted according to the error amplitude and phase deviation using PWM technology, so that the amplitude and phase of the output current match the target instruction; S6: By comparing the real-time values ​​of the detected harmonic current, reactive power and load current with the theoretical calculated values, the error correction effect is evaluated. If the error correction effect is poor, the detection model and compensation strategy are automatically adjusted.

[0023] In S1, high-precision sensors installed on the power grid busbar are used to collect the electrical parameters of the power grid in real time, including voltage, current, frequency, harmonic components and three-phase imbalance. Specifically: Monitor the real-time voltage value of the busbar and identify abnormal conditions such as voltage fluctuations, sudden rises or drops. Use a voltage transformer (VT) to convert the high voltage of the busbar into a low-voltage signal that matches the sensor; digitize the analog voltage signal through an analog-to-digital converter (ADC); sample the voltage waveform in real time and record its amplitude, frequency and phase information; use digital signal processing technology to extract the fundamental component and distortion components (such as harmonics) of the voltage.

[0024] Detect the current characteristics flowing through the bus, including instantaneous value, direction, harmonic components, etc., for analyzing the load status. Use a current transformer (CT) to convert the bus current into a low current signal; obtain the instantaneous value of the current signal through a sampling circuit and perform digital processing; perform Fourier transform (FFT) on the current signal to separate the fundamental and harmonic components; record the current amplitude, phase and its dynamic characteristics over time.

[0025] Determine the real-time frequency of the power grid, monitor frequency deviation, and warn of the impact of abnormal frequency on equipment operation. Perform periodic sampling on the collected voltage or current waveform; use zero-crossing detection or fast Fourier transform (FFT) to calculate the period of the waveform and infer the frequency: f=1 / T; where T is the period of the waveform; record the real-time value and change trend of the frequency, and evaluate the possible cause of the frequency deviation in combination with the bus load characteristics.

[0026] Analyze the harmonic components in the bus, including the amplitude and phase of each harmonic, and evaluate its impact on power quality. Perform fast Fourier transform (FFT) processing on the collected voltage and current signals to separate the harmonic components of different frequency components; Record the amplitude and phase of each harmonic and calculate the total harmonic distortion (THD): ;in, is the effective value of the nth harmonic, It is the effective value of the fundamental wave; compared with the standard value, it can determine whether the harmonic exceeds the standard and its impact range.

[0027] Detect the imbalance of the three-phase voltage and current of the busbar, and analyze whether the power grid has overload or uneven distribution problems. Collect the three-phase voltage and current signals separately, record the amplitude and phase of each phase; calculate the three-phase imbalance, which is defined as: Voltage imbalance: ; Current imbalance: ;in, and are the maximum, minimum and average values ​​of the three-phase voltages, respectively, and I corresponds to the current. The dynamic changes of the three-phase current distribution are tracked in real time to identify possible sources of load imbalance.

[0028] S2: Inject a small disturbance current into the power grid within the specified frequency range, measure the voltage response caused by the disturbance current, calculate the impedance value of the power grid at each frequency point, and calculate the ratio of the load current change rate to the voltage change rate by measuring the instantaneous change of the load current and voltage.

[0029] A small AC disturbance current is generated by a controllable current source, and the amplitude is controlled within 5% of the rated current of the power grid to avoid interference with the operation of the power grid. The frequency range of the disturbance current can be selected from 50 Hz to 5 kHz to cover the fundamental and main harmonic frequency bands of the power grid. The disturbance current is applied to the target bus through an injection device (such as a power amplifier) ​​to ensure that the injection process minimizes interference with other electrical parameters.

[0030] The real-time voltage signal of the busbar is collected through the voltage transformer (VT), including the instantaneous changes caused by the disturbance current; the voltage signal is digitally sampled using a high-speed analog-to-digital converter (ADC). The disturbance current I(f) and the voltage response V(f) are measured synchronously; Calculate the grid impedance value at each frequency point using the formula: ; Record the impedance-frequency characteristic Z(f), draw the impedance-frequency curve, and identify the resonant frequency point and harmonic amplification area.

[0031] The real-time current and voltage signals at the load end are collected through the current transformer (CT) and the voltage transformer (VT), respectively; the waveforms of the current and voltage and their changes over time are recorded using high-speed sampling technology.

[0032] Calculate load current from continuously sampled data and voltage The amount of change: , ; Record instantaneous changes at multiple time points to ensure sampling accuracy and dynamic range coverage.

[0033] Using the changes in load current and voltage, calculate the ratio of load current change rate to voltage change rate LDC, the expression is: Analyze the size and change trend of LDC to reflect the sensitivity of load to voltage change. Classify LDC data under different working conditions to form a load dynamic characteristic model as the basis for subsequent compensation strategy optimization.

[0034] S3: Draw the impedance frequency characteristic curve according to the calculated impedance value at each frequency point, analyze the abnormal degree of power grid harmonic amplification, and judge the dynamic response of the load under different working conditions according to the calculated ratio of the load current change rate to the voltage change rate.

[0035] Collect the calculated impedance value Z(f) at each frequency point of the power grid, including amplitude and phase. Determine the frequency range (such as 50 Hz - 5 kHz) and frequency interval (such as 10 Hz or integer multiples of harmonics) for impedance calculation to ensure that the fundamental and main harmonic frequency bands are covered.

[0036] Smooth the collected data and remove the abnormal points caused by noise to ensure the accuracy of the curve. In the two-dimensional coordinate system, the frequency f is the horizontal axis and the impedance amplitude is With frequency f as the horizontal axis and impedance phase angle ∠Z(f) as the vertical axis, a curve of impedance amplitude changing with frequency is plotted. The amplitude and phase characteristic curves are combined to form a complete graph of the grid impedance frequency characteristics.

[0037] After analyzing the curve of impedance amplitude changing with frequency, the impedance amplitude abnormality index is generated to judge the abnormal degree of power grid harmonic amplification. The method for obtaining the impedance amplitude abnormality index is as follows: The original signal of the impedance amplitude is set to x(t), which contains useful signals and noise: ; In the formula, xsignal(t) is the useful impedance signal; xnoise(t) is the high frequency or random noise. Select the wavelet function ψ(t) and the decomposition level J, perform discrete wavelet transform on x(t), and the wavelet transform coefficient The calculation expression is: ;in, is the wavelet function after translation and scale transformation, a is the scale parameter, b is the translation parameter; the approximate coefficient and detail coefficient obtained by wavelet decomposition are expressed as: ;in, and are the scaling function and the wavelet function, respectively. is the approximate coefficient, is the detail coefficient; reconstruct the signal components at each level, analyze the frequency components of the reconstructed signal, extract the main frequency components and their amplitudes within the passband, ; In the formula, is the minimum frequency value of the frequency range, is the maximum frequency value of the frequency range; To reconstruct the signal, identify the frequency with the largest amplitude in the spectrum As the main frequency, it represents the frequency component with the largest amplitude in the signal spectrum, and records the corresponding amplitude Represents the amplitude of the main frequency, for each frequency component in the passband Amplitude Calculate the mean and standard deviation , Refers to the i-th frequency component in the signal spectrum, and calculates the impedance amplitude anomaly index. The expression is: ; HK is the impedance amplitude anomaly index.

[0038] When the impedance amplitude abnormality index increases significantly, it means that the power grid has a higher impedance at a specific frequency point, which easily leads to the amplification of the harmonic current corresponding to the frequency, forming a harmonic amplification effect. This phenomenon is usually closely related to the resonance of the power grid, especially when it is close to the resonant frequency point, the power grid impedance presents a peak value, and the amplified harmonic current will cause serious damage to the power grid and electrical equipment, such as equipment overheating, malfunction, and decreased power grid stability. Therefore, a larger impedance amplitude abnormality index means a higher degree of harmonic amplification abnormality and a worse power quality of the power grid.

[0039] On the contrary, when the impedance amplitude anomaly index decreases significantly, it means that the impedance of the power grid at this frequency point is close to zero or very small. At this time, the power grid presents a low impedance path to the harmonic current of this frequency, which may cause the harmonic energy of this frequency to be widely propagated in the power grid. This situation should not be ignored either, because the low impedance state will weaken the effectiveness of the harmonic filter and may aggravate the harmonic anomalies at other frequency points. In addition, too low impedance may also reduce the power grid's ability to regulate sudden load changes. Therefore, a too small impedance amplitude anomaly index may also mean that the potential risk of harmonic amplification needs special attention.

[0040] Record LDC data points under different working conditions. According to the actual load type or operating status, classify the working conditions (such as steady state, dynamic, impact load, etc.). According to the size of LDC, judge the sensitivity of the load to voltage changes: when >High threshold: The load response is sensitive, which may cause frequent fluctuations in the power grid; <Low threshold: The load responds slowly, which may affect the optimization compensation efficiency. According to the distribution of LDC data points, the load is divided into: High dynamic load: frequent and rapid current fluctuations correspond to large LDC values; Low dynamic load: stable operation corresponds to small LDC values; Impact load: has a large amplitude in a short period of time. and characteristics.

[0041] According to the LDC data, a load dynamic response model is constructed to match the optimization compensation strategy for different load types. The horizontal axis is the current change Draw a load dynamic response distribution diagram with the vertical axis as the vertical axis. Use colors or areas to distinguish different load conditions based on the size and distribution characteristics of the LDC.

[0042] The load fluctuation index is generated by analyzing the change of the ratio of the load current change rate to the voltage change rate to judge the dynamic response of the load under different working conditions. The method for obtaining the load fluctuation index is as follows: Obtain current change rate from real-time monitoring of the load and voltage change rate The time series data is in the form of an N×2 matrix, where N is the number of time points and the first column is , the second column is : ; Standardize each column of data in X to remove the dimension effect, so that the mean of the data is 0 and the standard deviation is 1: ;in, is the mean of each column of data, is the standard deviation of each column of data, calculate the standardized data matrix, ; C is a 2×2 matrix that describes and The variance and covariance between them, T is the matrix transpose, calculate the eigenvalues ​​λ1, λ2 of the covariance matrix C and the corresponding eigenvectors v1, v2, the expression is: Cv=λv; the eigenvalue represents the variance of the data in the direction of the corresponding eigenvector.

[0043] Sort the eigenvalues ​​so that λ1≥λ2, and the corresponding eigenvectors are v1 (principal component) and v2 (secondary principal component). Project to the principal component direction v1 to get the principal component score : It is the main change component of load fluctuation, containing the most fluctuation information. The load fluctuation index is defined as the absolute mean of the main component scores: ; Where LFI is the load fluctuation index.

[0044] When the load fluctuation index is large, it means that the load has a strong dynamic response characteristic under the current working condition, that is, the load current is very sensitive to voltage changes. In this case, the load may be a high dynamic load or an impact load, which is manifested as a large fluctuation in load current in a short period of time. This working condition usually occurs in high-power equipment or nonlinear loads that are frequently started and stopped, which can easily cause grid voltage fluctuations, increase harmonics, and even interfere with other equipment. Therefore, a large load fluctuation index reflects that the grid needs a fast and accurate compensation strategy to maintain stable power quality.

[0045] On the contrary, when the load fluctuation index is small, it means that the dynamic response characteristics of the load are weak, that is, the load current is less sensitive to voltage changes. This situation may be a steady-state load or a low-dynamic load, which is manifested by small current and voltage fluctuations and relatively stable system operation. This type of working condition usually occurs in stable linear loads, with less disturbance to the power grid and lower compensation requirements for optimization devices. However, if the load fluctuation index is too small, it may mean that the system has problems with over-compensation or detection errors, and further analysis of the actual operating status is required.

[0046] S4: According to the abnormal degree of grid harmonic amplification and the dynamic response of the load under different working conditions, the accuracy of the detection instructions generated by the full-effect power optimization device is evaluated, and the detection instructions are divided into accuracy detection instructions and inaccuracy detection instructions according to the evaluation results.

[0047] The impedance amplitude anomaly index and the load fluctuation index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model uses each group of comprehensive feature vectors to predict the detection instruction accuracy value label generated by the full-effect electric energy optimization device as the prediction target, and takes minimizing the sum of prediction errors of the detection instruction accuracy value labels generated by all full-effect electric energy optimization devices as the training target. The machine learning model is trained until the sum of prediction errors converges and the model training is stopped. The accuracy value of the detection instruction generated by the full-effect electric energy optimization device is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0048] The method for obtaining the accuracy value of the detection instruction generated by the full-efficiency power optimization device is as follows: from the comprehensive feature vector training data of the trained machine learning model, the corresponding function expression is obtained: ; In the formula, is the output function of the model, HK is the impedance amplitude anomaly index, LFI is the load fluctuation index, The detection command accuracy value generated for the full power optimization device.

[0049] The acquired accuracy value of the detection instruction generated by the full-effect electric energy optimization device is compared with the accuracy reference threshold preset according to historical data. If the accuracy value of the detection instruction generated by the full-effect electric energy optimization device is greater than or equal to the preset accuracy reference threshold, it means that the accuracy of the detection instruction generated by the full-effect electric energy optimization device is high. At this time, no warning signal is generated, and the detection instruction is classified as an accuracy detection instruction; if the accuracy value of the detection instruction generated by the full-effect electric energy optimization device is less than the preset accuracy reference threshold, it means that the accuracy of the detection instruction generated by the full-effect electric energy optimization device is low. At this time, a warning signal is generated, and the detection instruction is classified as an inaccuracy detection instruction.

[0050] S5: Error correction is performed on the inaccurate detection instructions generated by the full-effect power optimization device. The PWM technology is used to dynamically adjust the duty cycle according to the error amplitude and phase deviation so that the amplitude and phase of the output current match the target instruction.

[0051] Detect the actual output current amplitude Ioutput and the target command amplitude Itarget to calculate the amplitude error : ; If eamp>0, the output current amplitude is insufficient and compensation needs to be increased; if eamp<0, the output current amplitude is too large and compensation needs to be reduced.

[0052] Detecting the phase of output current and target command phase Calculating Phase Deviation : ;like , the output current lags behind the target command, and the phase advance needs to be adjusted; if , the output current is ahead of the target instruction, and the phase delay needs to be adjusted.

[0053] PWM technology changes the amplitude and phase of the output current by adjusting the duty cycle D. The duty cycle is defined as the proportion of the high-level duration in one cycle: ; Where: Ton is the high level duration of the PWM signal, and T is the total period of the PWM signal.

[0054] According to the amplitude error eamp, adjust the PWM duty cycle Damp, the expression is: ; is the initial duty cycle; kamp is the amplitude adjustment coefficient, which is set according to the device response characteristics.

[0055] The adjusted duty cycle directly affects the output current amplitude: ; Indicates the amplitude of the corrected output current, which is obtained by adjusting the duty cycle of the compensation signal The output current amplitude achieved; Indicates the duty cycle of the PWM signal, which is used to control the proportional adjustment of the output current amplitude; represents a direct proportional relationship, that is Follow Increase and increase, with By increasing or decreasing the duty cycle, the output current amplitude can be adjusted in real time. According to the phase deviation Δφ, the switching delay time of PWM is adjusted , the expression is: ; By increasing or decreasing the switching delay time, the output current phase is synchronized with the target instruction. To achieve PWM signal control after phase correction: If (hysteresis), turn on the PWM switch in advance; if (Advance), delay the start of the PWM switch.

[0056] Use the current sensor to collect the adjusted output current signal in real time to obtain the corrected amplitude Ioutput,new and phase ,new.

[0057] Recalculate the amplitude error eamp,new and phase deviation : ; ; Adjust the PWM duty cycle and delay time again according to the new error value to form a closed-loop control until the error converges to the set threshold (such as the amplitude error is less than 1%, the phase deviation is less than 1°).

[0058] S6: By comparing the real-time values ​​of the detected harmonic current, reactive power and load current with the theoretical calculated values, the error correction effect is evaluated. If the error correction effect is poor, the detection model and compensation strategy are automatically adjusted.

[0059] Using sensors and sampling equipment, the grid parameters are collected in real time: harmonic current Iharmonic, detected: the real-time amplitude of the harmonic component in the grid; reactive power Qdetected: the real-time reactive power of the system; load current Iload, detected: the real-time current value at the load end.

[0060] According to the system operating status and compensation strategy, the theoretical values ​​are calculated: the theoretical value of harmonic current Iharmonic, theoretical: obtained by the harmonic filtering model or FFT analysis; the theoretical value of reactive power Qtheoretical: calculated by the reactive power compensation model; the theoretical value of load current Iload, theoretical: predicted by the load dynamic model.

[0061] Compare the real-time value with the theoretical value and calculate the error: ; ; ; In the formula, is the harmonic current error, is the reactive power error, is the load current error; Define the allowable error range: harmonic error threshold, such as an error less than 5% of the theoretical value; reactive power error threshold, such as an error less than 2%; load current error threshold, such as an error less than 3%.

[0062] If all errors meet the threshold, the correction effect is good and there is no need to adjust the model and compensation strategy; if any error exceeds the threshold, the correction effect is poor and the detection model and compensation strategy need to be automatically adjusted.

[0063] Automatic adjustment of the detection model includes: checking the harmonic filter model parameters, such as filter cutoff frequency, Q factor, etc.; updating the harmonic decomposition algorithm (such as adjusting FFT resolution or sampling frequency). Checking the recognition accuracy of the load type in the reactive power compensation model; optimizing the calculation logic of the inductive or capacitive load compensation. Optimizing the update rules of the LDC parameters in the load dynamic response model; if the load response characteristics change, the model can be adjusted in real time through an adaptive algorithm (such as Kalman filtering).

[0064] Use historical data to retrain the model and adjust parameters, such as optimizing compensation parameters through the least squares method. Use machine learning algorithms (such as random forest or LSTM) combined with real-time data to update the model and enhance the model's adaptability to complex load conditions.

[0065] If the harmonic error is large, check whether the harmonic compensation instruction is accurate, optimize the compensation current amplitude and phase; dynamically adjust the compensation frequency band and harmonic order of the harmonic filter.

[0066] If the reactive power error is large, check whether the inductive or capacitive reactive power compensation is over-compensated or under-compensated; Adjust the real-time amount of reactive current injection to ensure the power factor is close to 1.

[0067] If the load current error is large, check the three-phase current balance, optimize the compensation current distribution ratio, and dynamically adjust the compensation response time to adapt to rapidly changing load conditions.

[0068] The optimized detection model is used to generate new compensation instructions, including the frequency and amplitude of the harmonic filtering current; the dynamic adjustment amount of reactive power compensation; and the target distribution ratio of the three-phase load current balance.

[0069] The adjusted compensation strategy is executed in real time, and the adjusted harmonic current, reactive power and load current are monitored. The error is calculated again based on the new detection data to determine whether further optimization is needed, thus forming a closed-loop adjustment mechanism.

[0070] In this embodiment, through the high-precision sensors installed on the power grid bus, electrical parameters such as voltage, current, frequency, harmonic components and three-phase imbalance are collected in real time, and a small disturbance current is injected within the specified frequency range to measure the power grid impedance. By calculating the ratio of the rate of change of load current to voltage, the frequency characteristics of the power grid impedance and the dynamic response of the load are analyzed. According to the degree of abnormal harmonic amplification and the load response characteristics, the accuracy of the detection instructions generated by the full-effect power optimization device is evaluated, and they are divided into accuracy detection instructions and inaccuracy detection instructions. For inaccuracy detection instructions, the duty cycle is dynamically adjusted using PWM technology, and the output current is matched with the target instruction through amplitude and phase error correction. Finally, by comparing the real-time value of the detection parameter with the theoretical value, the correction effect is evaluated. If the correction effect is not good, the detection model and compensation strategy are automatically adjusted to achieve efficient and accurate power optimization.

[0071] Example 2, please refer to Figure 2 As shown, the full-efficiency electric energy optimization rapid detection system described in this embodiment includes a real-time data acquisition module, a power grid characteristic analysis module, a harmonic and load analysis module, a detection instruction accuracy evaluation module, an error correction module and an optimization module: Real-time data acquisition module: through the high-precision sensors installed on the power grid bus, the electrical parameters of the power grid are collected in real time, including voltage, current, frequency, harmonic components and three-phase imbalance; Grid characteristic analysis module: Injects a small disturbance current into the grid within a specified frequency range, measures the voltage response caused by the disturbance current, and calculates the impedance value of the grid at each frequency point, and calculates the ratio of the load current change rate to the voltage change rate by measuring the instantaneous change of the load current and voltage; Harmonic and load analysis module: Draw the impedance frequency characteristic curve according to the calculated impedance value at each frequency point, analyze the abnormal degree of harmonic amplification of the power grid, and judge the dynamic response of the load under different working conditions according to the calculated ratio of the load current change rate to the voltage change rate; Detection instruction accuracy evaluation module: evaluates the accuracy of the detection instructions generated by the full-efficiency power optimization device according to the abnormal degree of grid harmonic amplification and the dynamic response of the load under different working conditions, and divides the detection instructions into accurate detection instructions and inaccurate detection instructions according to the evaluation results; Error correction module: performs error correction on the inaccurate detection instructions generated by the full-effect power optimization device, and uses PWM technology to dynamically adjust the duty cycle according to the error amplitude and phase deviation, so that the amplitude and phase of the output current match the target instruction; Optimization module: By comparing the real-time values ​​of the detected harmonic current, reactive power and load current with the theoretical calculated values, the error correction effect is evaluated. If the error correction effect is poor, the detection model and compensation strategy are automatically adjusted.

[0072] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0073] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0074] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0075] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A fast detection method for full-effect power optimization, characterized by: The following steps are involved: S1: The electrical parameters of the power grid, including voltage, current, frequency, harmonic components and three-phase imbalance, are collected in real time through high-precision sensors installed on the power grid busbar; S2: Inject a small disturbance current into the power grid within the specified frequency range, measure the voltage response caused by the disturbance current, calculate the impedance value of the power grid at each frequency point, and calculate the ratio of the load current change rate to the voltage change rate by measuring the instantaneous change of the load current and voltage; S3: Draw the impedance frequency characteristic curve according to the calculated impedance value at each frequency point, analyze the abnormal degree of harmonic amplification of the power grid, and judge the dynamic response of the load under different working conditions according to the calculated ratio of the load current change rate to the voltage change rate; S4: According to the abnormal degree of harmonic amplification of the power grid and the dynamic response of the load under different working conditions, the accuracy of the detection instructions generated by the full-efficiency power optimization device is evaluated, and the detection instructions are divided into accurate detection instructions and inaccurate detection instructions according to the evaluation results; S5: Error correction is performed on the inaccurate detection instructions generated by the full-efficiency power optimization device, and the duty cycle is dynamically adjusted according to the error amplitude and phase deviation using PWM technology, so that the amplitude and phase of the output current match the target instruction; S6: By comparing the real-time values ​​of the detected harmonic current, reactive power and load current with the theoretical calculated values, the error correction effect is evaluated. If the error correction effect is poor, the detection model and compensation strategy are automatically adjusted.

2. The full-effect electric energy optimization rapid detection method according to claim 1 is characterized in that: In S3, the impedance amplitude abnormality index is generated after analyzing the curve of impedance amplitude changing with frequency. The impedance amplitude abnormality index is obtained as follows: Set the original signal of impedance amplitude to x(t), select wavelet function ψ(t) and decomposition level J, perform discrete wavelet transform on x(t), and the wavelet transform coefficients The calculation expression is: ;in, is the wavelet function after translation and scale transformation, a is the scale parameter, b is the translation parameter; the approximate coefficient and detail coefficient obtained by wavelet decomposition are expressed as: ;in, and are the scaling function and the wavelet function, respectively. is the approximate coefficient, is the detail coefficient; reconstruct the signal components at each level, analyze the frequency components of the reconstructed signal, extract the main frequency components and their amplitudes within the passband, ; In the formula, is the minimum frequency value of the frequency range, is the maximum frequency value of the frequency range; To reconstruct the signal, identify the frequency with the largest amplitude in the spectrum As the main frequency, it represents the frequency component with the largest amplitude in the signal spectrum, and records the corresponding amplitude Represents the amplitude of the main frequency, for each frequency component in the passband Amplitude Calculate the mean and standard deviation , Refers to the i-th frequency component in the signal spectrum, and calculates the impedance amplitude anomaly index. The expression is: ; HK is the impedance amplitude anomaly index.

3. The full-effect electric energy optimization rapid detection method according to claim 2 is characterized in that: In S3, the load fluctuation index is generated after analyzing the change of the ratio of the load current change rate to the voltage change rate. The method for obtaining the load fluctuation index is: Obtain current change rate from real-time monitoring of the load and voltage change rate The time series data is in the form of an N×2 matrix, where N is the number of time points and the first column is , the second column is : ; Standardize each column of data in X: ;in, is the mean of each column of data, is the standard deviation of each column of data, calculate the standardized data matrix, ; C is a 2×2 matrix that describes and The variance and covariance between them, T is the matrix transpose, calculate the eigenvalues ​​λ1, λ2 of the covariance matrix C and the corresponding eigenvectors v1, v2, the expression is: Cv=λv; the eigenvalue represents the variance of the data in the direction of the corresponding eigenvector; Sort the eigenvalues ​​so that λ1≥λ2, the corresponding eigenvectors are v1 and v2, and the standardized data Project to the principal component direction v1 to get the principal component score : is the change component of load fluctuation, and the load fluctuation index is defined as the absolute mean of the principal component scores: ; Where LFI is the load fluctuation index.

4. The full-effect electric energy optimization rapid detection method according to claim 1 is characterized in that: In S4, the accuracy of the detection instructions generated by the full-effect electric energy optimization device is evaluated according to the abnormal degree of harmonic amplification in the power grid and the dynamic response of the load under different working conditions. Specifically, the impedance amplitude abnormality index and the load fluctuation index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model uses each group of comprehensive feature vectors to predict the accuracy value label of the detection instructions generated by the full-effect electric energy optimization device as the prediction target, and takes minimizing the sum of prediction errors of the accuracy value labels of the detection instructions generated by all full-effect electric energy optimization devices as the training target. The machine learning model is trained until the sum of prediction errors converges and the model training is stopped. The accuracy value of the detection instructions generated by the full-effect electric energy optimization device is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

5. The full-effect electric energy optimization rapid detection method according to claim 4 is characterized in that: The acquired accuracy value of the detection instruction generated by the full-effect electric energy optimization device is compared with the accuracy reference threshold preset according to historical data. If the accuracy value of the detection instruction generated by the full-effect electric energy optimization device is greater than or equal to the preset accuracy reference threshold, it means that the accuracy of the detection instruction generated by the full-effect electric energy optimization device is high. At this time, no warning signal is generated, and the detection instruction is classified as an accuracy detection instruction; if the accuracy value of the detection instruction generated by the full-effect electric energy optimization device is less than the preset accuracy reference threshold, it means that the accuracy of the detection instruction generated by the full-effect electric energy optimization device is low. At this time, a warning signal is generated, and the detection instruction is classified as an inaccuracy detection instruction.

6. The full-effect electric energy optimization rapid detection method according to claim 5 is characterized in that: In S5, error correction is performed on the inaccuracy detection instruction generated by the full-efficiency power optimization device. The PWM technology is used to dynamically adjust the duty cycle according to the error amplitude and phase deviation so that the amplitude and phase of the output current match the target instruction. Specifically: Detect the actual output current amplitude Ioutput and the target command amplitude Itarget to calculate the amplitude error : ; If eamp>0, the output current amplitude is insufficient and the compensation needs to be increased; if eamp<0, the output current amplitude is too large and the compensation needs to be reduced; Detecting the phase of output current and target command phase Calculating Phase Deviation : ;like , the output current lags behind the target command, and the phase advance needs to be adjusted; if , the output current is ahead of the target instruction, and the phase delay needs to be adjusted.

7. The full-effect electric energy optimization rapid detection method according to claim 6 is characterized in that: PWM technology changes the amplitude and phase of the output current by adjusting the duty cycle D. The duty cycle is defined as the proportion of the high-level duration in one cycle: ; Where: Ton is the high level duration of the PWM signal, T is the total period of the PWM signal; According to the amplitude error eamp, adjust the PWM duty cycle Damp, the expression is: ; is the initial duty cycle; kamp is the amplitude adjustment coefficient, which is set according to the device response characteristics; the adjusted duty cycle directly affects the output current amplitude: ; According to the phase deviation , adjust the PWM switching delay time , the expression is: :like , turn on the PWM switch in advance; if , delay turning on the PWM switch; Use the current sensor to collect the adjusted output current signal in real time to obtain the corrected amplitude Ioutput,new and phase ,new; recalculate the amplitude error eamp,new and phase deviation : ; ; Adjust the PWM duty cycle and delay time again according to the new error value to form a closed-loop control until the error converges to the set threshold.

8. The full-effect electric energy optimization rapid detection method according to claim 7 is characterized in that: In S6, sensors and sampling devices are used to collect grid parameters in real time: harmonic current Iharmonic,detected: the real-time amplitude of harmonic components in the grid; reactive power Qdetected: the real-time reactive power of the system; Load current Iload, detected: real-time current value at the load end; According to the system operating status and compensation strategy, calculate the theoretical value of harmonic current Iharmonic,theoretical, the theoretical value of reactive power Qtheoretical and the theoretical value of load current Iload,theoretical; Compare the real-time value with the theoretical value, calculate the error, and define the allowable error range; if all errors meet the threshold, the correction effect is good and there is no need to adjust the model and compensation strategy; if any error exceeds the threshold, the correction effect is poor and the detection model and compensation strategy need to be automatically adjusted.

9. The full-effect electric energy optimization rapid detection method according to claim 8, characterized in that: If the harmonic error is large, check whether the harmonic compensation instruction is accurate, optimize the compensation current amplitude and phase; dynamically adjust the compensation frequency band of the harmonic filter and the harmonic order; If the reactive power error is large, check whether the inductive or capacitive reactive power compensation is over-compensated or under-compensated; Adjust the real-time amount of reactive current injection; If the load current error is large, check the three-phase current balance, optimize the compensation current distribution ratio, and dynamically adjust the compensation response time.

10. A full-effect electric energy optimization rapid detection system, used to implement the full-effect electric energy optimization rapid detection method according to any one of claims 1 to 9, characterized in that: It includes real-time data acquisition module, power grid characteristics analysis module, harmonic and load analysis module, detection instruction accuracy evaluation module, error correction module and optimization module: Real-time data acquisition module: through the high-precision sensors installed on the power grid bus, the electrical parameters of the power grid are collected in real time, including voltage, current, frequency, harmonic components and three-phase imbalance; Grid characteristic analysis module: Injects a small disturbance current into the grid within a specified frequency range, measures the voltage response caused by the disturbance current, and calculates the impedance value of the grid at each frequency point, and calculates the ratio of the load current change rate to the voltage change rate by measuring the instantaneous change of the load current and voltage; Harmonic and load analysis module: Draw the impedance frequency characteristic curve according to the calculated impedance value at each frequency point, analyze the abnormal degree of harmonic amplification of the power grid, and judge the dynamic response of the load under different working conditions according to the calculated ratio of the load current change rate to the voltage change rate; Detection instruction accuracy evaluation module: evaluates the accuracy of the detection instructions generated by the full-efficiency power optimization device according to the abnormal degree of grid harmonic amplification and the dynamic response of the load under different working conditions, and divides the detection instructions into accurate detection instructions and inaccurate detection instructions according to the evaluation results; Error correction module: performs error correction on the inaccurate detection instructions generated by the full-effect power optimization device, and uses PWM technology to dynamically adjust the duty cycle according to the error amplitude and phase deviation, so that the amplitude and phase of the output current match the target instruction; Optimization module: By comparing the real-time values ​​of the detected harmonic current, reactive power and load current with the theoretical calculated values, the error correction effect is evaluated. If the error correction effect is poor, the detection model and compensation strategy are automatically adjusted.

Images