Multi-source data electric energy quality energy-saving operation monitoring method and device based on AI Internet of Things

Abstract

The invention relates to a multi-source data electric energy quality energy-saving operation monitoring method and device based on AI Internet of Things, and the method comprises the steps: obtaining a transient electromagnetic signal of a permanent magnet operating mechanism circuit breaker, and extracting a time sequence characteristic parameter of the transient electromagnetic signal; judging whether a harmonic problem caused by an inverter exists or not and whether voltage suddenly rises or drops or not, and determining whether the time sequence characteristic parameters are associated with power quality disturbance or not; acquiring a voltage jump amplitude and a voltage swell rate to obtain a voltage fluctuation degree, and outputting a discharge strategy of the energy storage system through a pre-established discharge optimization model; respectively calculating to obtain a voltage distortion rate and a three-phase voltage unbalance degree, and selecting a charging strategy of the energy storage system through a pre-established charging optimization model; predicting a sudden rise rate and a sudden drop rate of the sample to obtain a prediction result, and forming an energy-saving optimization scheme of an energy storage charging and discharging strategy according to the prediction result; the charging and discharging power of the energy storage pool is automatically adjusted, an energy-saving optimization scheme is formed, and the stability of a power grid is effectively improved.

Description

Technical Field

[0001] The present disclosure relates to, specifically, a method and device for monitoring power quality and energy-saving operation based on AI Internet of Things multi-source data. Background Art

[0002] In smart city power plants, permanent magnet actuator circuit breakers are key equipment, and the timing characteristics of their transient electromagnetic signals directly reflect changes in the operating status of the power grid. When large-scale renewable energy is connected to the grid, the intermittent and volatile nature of renewable energy generation can easily cause power quality disturbances such as voltage surges and dips. These disturbances not only affect the stable operation of the power grid, but also cause potential damage to power equipment. Therefore, studying the correlation between the timing characteristics of transient electromagnetic signals of permanent magnet actuator circuit breakers and power quality disturbances is of great significance for improving the stability and safety of the power grid. In actual operation, when the voltage fluctuation reaches a certain type and degree, the AIoT monitoring system needs to quickly identify and analyze the characteristics of transient electromagnetic signals. Based on these characteristics, the AIoT monitoring system needs to optimize the charging and discharging strategy of the grid-side energy storage system to actively respond to power quality disturbances. However, there are technical contradictions in this process: on the one hand, the rapid response of the energy storage system requires high-precision identification of transient electromagnetic signals, but the signal itself is complex and uncertain, which increases the difficulty of identification; on the other hand, the charging and discharging strategy of the energy storage system needs to be highly matched with the real-time operating status of the power grid, but the changes in the power grid state are random and sudden, and difficult to accurately predict. In addition, the impact of this active power quality optimization strategy on line loss, the operating frequency of reactive compensation equipment and the overall energy saving effect also needs to be analyzed in depth. Since the charging and discharging behavior of the energy storage system will change the power distribution of the power grid, it may cause an increase or decrease in line loss, and also affect the operating frequency of reactive compensation equipment. How to find a balance among these mutually influencing factors is an important technical issue faced by smart city power plants when dealing with large-scale new energy grid connection. Summary of the invention

[0003] In order to solve the problems existing in the above-mentioned prior art, the purpose of the present invention is to provide a method and device for monitoring power quality and energy-saving operation based on AI Internet of Things multi-source data.

[0004] The present invention discloses a method for monitoring power quality and energy-saving operation based on AI IoT multi-source data, comprising the following steps: S1. Acquire a transient electromagnetic signal of a permanent magnetic operating mechanism circuit breaker and extract a time sequence characteristic parameter of the transient electromagnetic signal; S2, judging whether there is a harmonic problem caused by the inverter and whether a voltage surge or a voltage drop occurs, and determining whether the timing characteristic parameter is associated with the power quality disturbance; S3, obtaining a voltage mutation amplitude and a voltage surge rate, obtaining a voltage fluctuation degree according to the voltage mutation amplitude and the voltage surge rate, inputting the voltage fluctuation degree into a discharge optimization model, and outputting a discharge strategy for the energy storage system; S4. Calculate the voltage distortion rate and the three-phase voltage imbalance, and select the charging strategy of the energy storage system through the pre-established charging optimization model; S5. Evaluate the impact of charging strategy and discharging strategy on the harmonic current content of the transmission line and the energy loss corresponding to the power factor change, and determine the target energy-saving charging strategy and the target energy-saving discharging strategy; S6. Real-time monitoring of the instantaneous power curve and power spectrum density corresponding to the transient electromagnetic signals of the target energy-saving charging strategy and the target energy-saving discharging strategy, and performing time series analysis to obtain voltage value time series data; dividing the time series data into samples of a set length, predicting the sudden rise rate and sudden fall rate of the samples, obtaining prediction results, and forming an energy-saving optimization plan for the energy storage charging and discharging strategy according to the prediction results.

[0005] Preferably, the timing characteristic parameters include frequency components, amplitude changes, phase differences and harmonic content; The step S1 comprises: The electromagnetic torque stability parameter is calculated based on the original waveform data of the electromagnetic signal collected by the circuit breaker sensor and the magnetic ring strength of the permanent magnetic material to obtain a first transient electromagnetic signal sequence; Decomposing the first transient electromagnetic signal sequence in the frequency domain by fast Fourier transform, and generating a second transient electromagnetic signal sequence if the spectrum energy density is greater than the magnetic flux attenuation degree threshold; The third transient electromagnetic signal sequence is obtained by performing wavelet decomposition on the second transient electromagnetic signal sequence to obtain high-frequency coefficients and low-frequency coefficients, and extracting the electromagnetic displacement change and the switching state continuity characteristic curve from the high-frequency coefficients and the low-frequency coefficients. The recursive least squares method is used to fit the third transient electromagnetic signal sequence to obtain an amplitude change characteristic curve, and the fundamental component and harmonic component of the amplitude change characteristic curve are corrected by a phase compensation algorithm to obtain frequency component data and harmonic content data to construct a time series characteristic parameter matrix.

[0006] Preferably, the step S2 comprises: Acquire a transient electromagnetic signal from a power grid sampling device, perform Fourier series decomposition on the transient electromagnetic signal, calculate the cumulative energy of the frequency component within a preset frequency band, and obtain first frequency spectrum feature data based on the comparison of the frequency band energy increment with a preset threshold; Extracting the amplitude of the transient electromagnetic signal according to the first frequency spectrum characteristic data, calculating the amplitude change rate by discrete Fourier transform, and generating first voltage sudden change characteristic data if the amplitude change rate is greater than a preset amplitude change threshold; Extracting spectrum energy and transient response characteristics of the first voltage sudden change characteristic data by wavelet packet transformation to obtain second spectrum characteristic data; The second frequency spectrum feature data is analyzed by correlation calculation to extract time series correlation features. If the matching degree of the time series correlation features exceeds a preset threshold, it is determined that the transient electromagnetic signal is associated with the power quality disturbance caused by the new energy grid connection.

[0007] Preferably, the step S3 comprises: Acquire voltage mutation amplitude data and voltage surge rate data from the power grid sampling device, locate the mutation point of the voltage mutation amplitude data by wavelet edge detection, and obtain voltage fluctuation degree data; According to the voltage fluctuation degree data and the remaining capacity data of the energy storage device, a recursive least square method is used to perform data fitting to obtain a first power response parameter; The first power response parameter and the historical discharge data of the energy storage device are optimized and calculated by a pre-established discharge optimization model to obtain a discharge optimization curve and a second power response parameter; The discharge depth threshold of the energy storage device is calculated according to the second power response parameter, and the response strategy matrix is ​​constructed using the discharge depth threshold and the fluctuation response delay parameter to obtain the discharge strategy of the energy storage system; the discharge strategy includes determining the output power, discharge depth and speed of responding to voltage fluctuations of the energy storage system.

[0008] Preferably, the step S3 further comprises: The current energy storage value, load value and system operating parameters of the energy storage system are obtained, and the objective function is constructed in combination with the preset number of time periods and electricity price table; the target discharge capacity and target energy storage value are set as constraints, and the linear programming algorithm is used to solve the objective function under the constraints to obtain the target discharge strategy set; the discharge strategies whose discharge capacity exceeds the preset capacity threshold in the target strategy set are screened out, the discharge capacity is adjusted, and the target discharge strategy is obtained.

[0009] Preferably, the step S4 comprises: Acquire three-phase voltage waveform data from a power grid monitoring device, and obtain first voltage distortion data by performing Fourier transform analysis on the three-phase voltage waveform data; Decomposing the three-phase voltage waveform data into symmetrical components, calculating the phase deviation value and the voltage amplitude using the least square method, and obtaining first imbalance data; Perform compensation calculation according to the first voltage distortion data and the first imbalance data, determine the compensation voltage amplitude and phase by a symmetrical component method, and obtain first harmonic compensation data and first phase sequence compensation data; An optimization calculation of charging capacity is performed on the first harmonic compensation data and the first phase sequence compensation data, and constraint verification is performed on compensation parameters through a preset compensation model to screen out a charging strategy for the energy storage system.

[0010] Preferably, the step S4 further comprises: The harmonic content data is calculated from the harmonic content data in the transient electromagnetic signal, and the voltage distortion rate is calculated based on the harmonic content data; by measuring the instantaneous value of the three-phase voltage, the difference between the first target phase voltage and the second target phase voltage in the three-phase voltage, as well as the average value of the three-phase voltage are determined, and the three-phase voltage imbalance is calculated.

[0011] Preferably, the step S5 comprises: Obtaining harmonic current content and power factor benchmark data before the implementation of the charging and discharging strategy in the line monitoring device, and performing spectrum analysis on the benchmark data through Fourier transform to obtain first evaluation data; According to the first evaluation data, the line operation data after the implementation of the charging and discharging strategy is collected, and the harmonic current content change value is calculated by Fourier transform to obtain the second evaluation data; Calculating a power factor change value according to the second evaluation data, and obtaining third evaluation data through reactive power compensation calculation; A line loss calculator is used to comprehensively calculate the second evaluation data and the third evaluation data, and the energy loss value per unit length is integrated according to the line resistance parameter to obtain the total line energy loss data; If the harmonic current content in the total line energy loss data is lower than the reference value and the power factor is higher than the reference value, the current charging and discharging strategy is determined as the target energy-saving charging and discharging strategy.

[0012] Preferably, the step S6 comprises: Decomposing the electromagnetic signal by fast Fourier transform, and obtaining first power spectrum density data according to spectrum distribution calculation; Performing instantaneous power calculation on the first power spectrum density data, and performing segmented processing on the instantaneous power curve using a sliding window to obtain first time series data; Performing sample division according to the first time series data, extracting voltage values ​​from the divided sample segments, and obtaining second time series data; Performing long short-term memory network training on the second time series data, extracting voltage change characteristics from historical data, and obtaining a first prediction parameter; The voltage swell rate and the voltage sag rate are calculated by using the first prediction parameter, and a threshold comparator is used to make a judgment to obtain a second prediction parameter; If the surge rate or the sag rate exceeds the preset threshold, the current voltage change data is recorded, and the power adjustment amount is generated according to the change amplitude to obtain the energy storage optimization parameters.

[0013] The present disclosure also provides a power quality energy-saving operation monitoring device based on AI Internet of Things multi-source data, including: A transient electromagnetic signal acquisition module, which is used to acquire a transient electromagnetic signal of a permanent magnetic operating mechanism circuit breaker; A time series characteristic parameter extraction module, wherein the time series characteristic parameter extraction module is used to extract the time series characteristic parameters of the transient electromagnetic signal; A power quality disturbance association module, which is used to determine whether there is a harmonic problem caused by the inverter and whether a voltage surge or a voltage drop occurs, and to determine whether the timing characteristic parameter is associated with the power quality disturbance; A discharge strategy optimization module, wherein the discharge strategy optimization module is used to obtain a voltage mutation amplitude and a voltage surge rate, obtain a voltage fluctuation degree according to the voltage mutation amplitude and the voltage surge rate, input the voltage fluctuation degree into a discharge optimization model, and output a discharge strategy for the energy storage system; A charging strategy optimization module, which is used to calculate the voltage distortion rate and the three-phase voltage imbalance, and select a charging strategy for the energy storage system through a pre-established charging optimization model; An energy-saving strategy evaluation module, which is used to evaluate the impact of charging strategy and discharging strategy on the harmonic current content of the transmission line and the energy loss corresponding to the power factor change, and determine the target energy-saving charging strategy and the target energy-saving discharging strategy; An adjustment module, the adjustment module is used to monitor in real time the instantaneous power curve and power spectrum density corresponding to the transient electromagnetic signals of the target energy-saving charging strategy and the target energy-saving discharging strategy, and perform time series analysis to obtain voltage value time series data; divide the time series data into samples of a set length, predict the sudden rise rate and sudden fall rate of the samples, obtain the prediction results, and form an energy-saving optimization plan for the energy storage charging and discharging strategy according to the prediction results.

[0014] The advantages of the power quality energy-saving operation monitoring method and device based on AI Internet of Things multi-source data disclosed in the present invention are: By acquiring the transient electromagnetic signal of the permanent magnet operating mechanism circuit breaker and extracting the timing characteristic parameters, the power quality disturbance caused by the large-scale grid connection of new energy sources can be identified. According to the degree of voltage fluctuation, the present invention outputs the discharge strategy of the energy storage system, including output power, discharge depth and response speed. At the same time, by analyzing the voltage distortion rate and the three-phase voltage imbalance, a suitable charging strategy is selected to achieve harmonic suppression, imbalance compensation and power factor correction. The present invention also evaluates the impact of the charging and discharging strategy on the energy loss of the transmission line, and monitors the instantaneous power curve and power spectrum density in real time to predict the risk of voltage sudden change. When the grid is detected to be unstable, the present invention automatically adjusts the charging and discharging power of the energy storage cell to form an energy-saving optimization plan, effectively improving the stability of the grid and the efficiency of energy utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flow chart of a method and device for monitoring power quality and energy-saving operation based on AI Internet of Things multi-source data as described in the present invention. DETAILED DESCRIPTION

[0016] like Figure 1 As shown, the present disclosure discloses a method for monitoring power quality and energy-saving operation based on AI IoT multi-source data, comprising the following steps: S1. Acquire a transient electromagnetic signal of a permanent magnetic operating mechanism circuit breaker and extract a time sequence characteristic parameter of the transient electromagnetic signal; S2, judging whether there is a harmonic problem caused by the inverter and whether a voltage surge or a voltage drop occurs, and determining whether the timing characteristic parameter is associated with the power quality disturbance; S3, obtaining a voltage mutation amplitude and a voltage surge rate, obtaining a voltage fluctuation degree according to the voltage mutation amplitude and the voltage surge rate, inputting the voltage fluctuation degree into a discharge optimization model, and outputting a discharge strategy for the energy storage system; S4. Calculate the voltage distortion rate and the three-phase voltage imbalance, and select the charging strategy of the energy storage system through the pre-established charging optimization model; S5. Evaluate the impact of charging strategy and discharging strategy on the harmonic current content of the transmission line and the energy loss corresponding to the power factor change, and determine the target energy-saving charging strategy and the target energy-saving discharging strategy; S6. Real-time monitoring of the instantaneous power curve and power spectrum density corresponding to the transient electromagnetic signals of the target energy-saving charging strategy and the target energy-saving discharging strategy, and performing time series analysis to obtain voltage value time series data; dividing the time series data into samples of a set length, predicting the sudden rise rate and sudden fall rate of the samples, obtaining prediction results, and forming an energy-saving optimization plan for the energy storage charging and discharging strategy according to the prediction results.

[0017] Further, the timing characteristic parameters in step S1 may include frequency components, amplitude changes, phase differences, and harmonic content; More specifically, step S1 may include: The electromagnetic torque stability parameter is calculated based on the original waveform data of the electromagnetic signal collected by the circuit breaker sensor and the magnetic ring strength of the permanent magnetic material to obtain a first transient electromagnetic signal sequence; Decomposing the first transient electromagnetic signal sequence in the frequency domain by fast Fourier transform, and generating a second transient electromagnetic signal sequence if the spectrum energy density is greater than the magnetic flux attenuation degree threshold; The third transient electromagnetic signal sequence is obtained by performing wavelet decomposition on the second transient electromagnetic signal sequence to obtain high-frequency coefficients and low-frequency coefficients, and extracting the electromagnetic displacement change and the switching state continuity characteristic curve from the high-frequency coefficients and the low-frequency coefficients. The recursive least squares method is used to fit the third transient electromagnetic signal sequence to obtain an amplitude change characteristic curve, and the fundamental component and harmonic component of the amplitude change characteristic curve are corrected by a phase compensation algorithm to obtain frequency component data and harmonic content data to construct a time series characteristic parameter matrix.

[0018] More specifically, the original waveform data of the electromagnetic signal with a sampling frequency of at least 800 times per second is obtained from the circuit breaker sensor, and the electromagnetic torque stability parameter is calculated according to the magnetic ring strength of the permanent magnetic material and the voltage fluctuation amplitude to obtain the first transient electromagnetic signal sequence. For the first transient electromagnetic signal sequence, the signal is decomposed in the frequency domain by fast Fourier transform, and the transient response time of the circuit breaker is judged according to the spectrum energy density and the current peak value. If it is greater than the magnetic flux attenuation threshold of 0.3, a second transient electromagnetic signal sequence is generated. The second transient electromagnetic signal sequence is decomposed by five layers of wavelet to obtain high-frequency coefficients and low-frequency coefficients, and the electromagnetic displacement change and the switch state continuity characteristic curve are extracted from the coefficients to obtain the third transient electromagnetic signal sequence. The permanent magnetic force loop parameters are calculated according to the third transient electromagnetic signal sequence, and the characteristic curve is fitted by recursive least squares method to obtain the amplitude change characteristic curve. On the basis of the amplitude change characteristic curve, the electromagnetic signal is expanded by Fourier series according to the proportion of harmonic components to obtain the fundamental component and the harmonic component. The phase compensation algorithm is used to correct the phase of the fundamental component and the harmonic component to obtain the frequency component data, phase difference data and harmonic content data. The time series characteristic parameter matrix is ​​constructed based on the frequency component data, phase difference data and harmonic content data to complete the extraction of the time series characteristic parameters of the transient electromagnetic signal.

[0019] Exemplarily, the permanent magnetic operating mechanism circuit breaker uses a high-precision sensor array in the electromagnetic signal acquisition stage, and the sampling frequency is configured to be 800 times per second. At the sampling moment, the transient electromagnetic signal is calculated by the ratio of the electromagnetic torque stability parameter to the voltage fluctuation amplitude. In actual operation, the circuit breaker magnetic ring strength fluctuates within the range of 12 mm. If it is higher than this value, it will have an adverse effect on the permanent magnetic circuit. After obtaining the electromagnetic signal waveform, the waveform is converted to the frequency domain through Fourier transform. At this time, it will be found that the spectrum energy density is mainly concentrated in the range of 20 Hz to 60 Hz, and the current peak decays with the increase of the number of circuit breaker operations. When the attenuation exceeds the threshold of 0.3, it indicates that the performance of the permanent magnetic material has decreased. At this time, the original signal is resampled to improve the analysis accuracy. Five-layer wavelet decomposition of the sampled electromagnetic signal can obtain coefficients of different frequency bands, where the high-frequency coefficient reflects the transient characteristics of the switch, and the low-frequency coefficient contains steady-state operation information. By extracting the change in electromagnetic displacement in the coefficient, it is found that the displacement fluctuates between 0.8 mm and 1.2 mm when the switch state changes, and the duration is about 15 milliseconds to 25 milliseconds. The parameters of the permanent magnetic circuit include key indicators such as magnetic flux density and magnetic field strength. When fitting the characteristic curve using the recursive least squares method, the initial magnetic flux density is set to 1.2 Tesla and the magnetic field strength is set to 950 amperes per meter. The amplitude change characteristic curve is obtained through iterative calculation, and the slope of the curve reflects the performance change trend of the permanent magnetic material. When performing harmonic analysis, the fundamental frequency is 50 Hz. The content of the 2nd to 7th harmonics can be obtained through Fourier series expansion. During normal operation, the content of each harmonic does not exceed 10% of the fundamental wave, among which the 3rd and 5th harmonics account for a large proportion, about 8% and 6% respectively. During the phase correction process, it was found that there was a phase difference between the fundamental wave and each harmonic. The phase compensation algorithm was used to control the phase difference within 5 degrees. The compensated frequency component data more accurately reflects the operating status of the circuit breaker. At the same time, the phase difference data and harmonic content data also provide important basis for fault diagnosis. The constructed timing characteristic parameter matrix contains four dimensions: frequency component, amplitude change, phase difference and harmonic content. Each element in the matrix has a clear physical meaning. By analyzing the change law of the parameter matrix, the operating status of the permanent magnetic actuator circuit breaker can be effectively monitored.

[0020] Furthermore, step S2 may specifically include: Acquire a transient electromagnetic signal from a power grid sampling device, perform Fourier series decomposition on the transient electromagnetic signal, calculate the cumulative energy of the frequency component within a preset frequency band, and obtain first frequency spectrum feature data based on the comparison of the frequency band energy increment with a preset threshold; Extracting the amplitude of the transient electromagnetic signal according to the first frequency spectrum characteristic data, calculating the amplitude change rate by discrete Fourier transform, and generating first voltage sudden change characteristic data if the amplitude change rate is greater than a preset amplitude change threshold; Extracting spectrum energy and transient response characteristics of the first voltage sudden change characteristic data by wavelet packet transformation to obtain second spectrum characteristic data; The second frequency spectrum feature data is analyzed by correlation calculation to extract time series correlation features. If the matching degree of the time series correlation features exceeds a preset threshold, it is determined that the transient electromagnetic signal is associated with the power quality disturbance caused by the new energy grid connection.

[0021] More specifically, a first transient electromagnetic signal is obtained from a power grid sampling device, and the frequency component is cumulatively calculated in a preset frequency band by Fourier series decomposition, and the first spectrum characteristic data is obtained by comparing the frequency band energy increment with a preset threshold. The transient electromagnetic signal is amplitude extracted based on the first spectrum characteristic data, and the amplitude change rate is calculated by discrete Fourier transform, and the first voltage change characteristic data is obtained by comparing the amplitude change rate with a preset amplitude change threshold. According to the first spectrum characteristic data, if the frequency band energy increment is greater than the preset threshold, it is determined that there is an inverter harmonic, and the first harmonic characteristic data is generated. According to the first voltage change characteristic data, if the amplitude change rate is greater than the preset threshold, it is determined that a voltage sudden change occurs, and the first voltage sudden change characteristic data is generated. The spectrum energy and transient response characteristics of the first harmonic characteristic data and the first voltage sudden change characteristic data are extracted by wavelet packet transform to obtain the second spectrum characteristic data. The second spectrum characteristic data is analyzed by correlation calculation, and the time series correlation characteristics are extracted to obtain the power quality disturbance characteristic data. The power quality disturbance characteristic data is analyzed by matching degree calculation. If the matching degree exceeds the preset threshold, it is determined that the transient electromagnetic signal is associated with the power quality disturbance caused by the connection of new energy to the grid.

[0022] Exemplarily, when the grid sampling device obtains transient electromagnetic signals from the circuit breaker, the sampling frequency is set to 10 kHz, the sampling time is 200 milliseconds, and the complete transient process waveform is recorded. The amplitude range of the sampled original signal fluctuates between plus or minus 100 volts, and the signal contains fundamental waves and harmonic components. The signal is analyzed in the frequency domain by Fourier series decomposition, and the energy distribution is statistically calculated in the preset 50 Hz to 2000 Hz frequency band. During normal operation, the energy in this frequency band accounts for about 15%. When the frequency band energy accounts for more than 25%, it indicates that the harmonic content of the inverter output increases, and the first spectrum feature data shows abnormality. In the amplitude analysis, the amplitude is calculated every 1 millisecond for the 200 millisecond sampling window to obtain an amplitude sequence of 200 sampling points. The amplitude change rate is calculated by discrete Fourier transform. When the change rate exceeds 0.2, it is often accompanied by a sudden rise or fall in voltage. This voltage change characteristic reflects the transient disturbance in the power grid. Inverter harmonics are mainly manifested as switching frequency and its multiples. The common switching frequency is 2 kHz to 20 kHz. After frequency reduction modulation, harmonic components are generated in the low frequency band, of which the most significant are the sidebands near 100 Hz and 150 Hz. These characteristics are recorded in the first harmonic characteristic data. The voltage surge phenomenon usually lasts for 3 to 5 power frequency cycles, which is manifested as a rapid change in voltage amplitude, with the change amplitude exceeding 20% ​​of the rated value. The start time, duration and amplitude change of the surge are recorded in the first voltage surge characteristic data. Wavelet packet transform can provide time domain and frequency domain resolution at the same time. After 6 layers of signal decomposition, 64 frequency bands are obtained. The energy change of each frequency band reflects the time-varying characteristics of different frequency components. The second spectrum characteristic data is obtained by calculating the energy proportion of each frequency band. In the correlation analysis, a sliding time window is used to calculate the correlation coefficient between feature sequences. The window length is selected as 100 milliseconds and the sliding step is 10 milliseconds. When the correlation coefficient is greater than 0.8, it is considered that there is a significant correlation. This correlation is reflected in the power quality disturbance characteristic data. The degree of correlation between the transient electromagnetic signal and the disturbance caused by the grid connection of renewable energy is determined by the matching degree calculation. The matching degree calculation takes into account the three dimensions of spectrum characteristics, amplitude characteristics and timing characteristics. When the comprehensive matching degree exceeds the preset threshold of 0.75, it indicates that the transient process is indeed caused by the grid connection of renewable energy.

[0023] Furthermore, step S3 may specifically include: Acquire voltage mutation amplitude data and voltage surge rate data from the power grid sampling device, locate the mutation point of the voltage mutation amplitude data by wavelet edge detection, and obtain voltage fluctuation degree data; According to the voltage fluctuation degree data and the remaining capacity data of the energy storage device, a recursive least square method is used to perform data fitting to obtain a first power response parameter; The first power response parameter and the historical discharge data of the energy storage device are optimized and calculated by a pre-established discharge optimization model to obtain a discharge optimization curve and a second power response parameter; The discharge depth threshold of the energy storage device is calculated according to the second power response parameter, and the response strategy matrix is ​​constructed using the discharge depth threshold and the fluctuation response delay parameter to obtain the discharge strategy of the energy storage system; the discharge strategy includes determining the output power, discharge depth and speed of responding to voltage fluctuations of the energy storage system.

[0024] More specifically, the voltage mutation amplitude and voltage surge rate data are obtained from the grid sampling device, the mutation point of the voltage fluctuation sequence is located by wavelet edge detection, and the voltage fluctuation degree data is calculated according to the voltage change before and after the mutation point. The remaining capacity data of the energy storage device is obtained from the energy storage management unit, and the voltage fluctuation degree data is fitted by recursive least squares method, and the first power response parameter is calculated according to the fitting curve. The first power response parameter is calculated according to the pre-established discharge optimization model, and the discharge optimization curve is obtained by training the historical discharge data of the energy storage device. The power regulation accuracy is calculated segmentally using the discharge optimization curve to obtain the second power response parameter. The discharge depth threshold of the energy storage device is calculated according to the second power response parameter to obtain the discharge depth parameter. The response strategy matrix is ​​constructed according to the discharge depth parameter and the fluctuation response delay parameter to obtain the response speed parameter. The output power parameter, discharge depth parameter and response speed parameter of the energy storage device are calculated by the response strategy matrix to form the discharge strategy data.

[0025] Exemplarily, the grid sampling device collects voltage data every 20 milliseconds, identifies voltage mutation points through wavelet edge detection, and determines that when the voltage changes by more than 10% of the nominal value within 5 milliseconds, it is considered a mutation point, and the mutation amplitude and sudden rise rate are recorded. 100 sampling points before and after the mutation are analyzed to calculate the voltage fluctuation degree. The energy storage management unit monitors the remaining capacity of the battery pack in real time. Taking the lithium battery energy storage system as an example, the capacity is 10 megawatt-hours. When the current remaining capacity is 80%, the 100 recent sets of voltage fluctuation data are fitted by recursive least squares method to obtain the correlation curve between voltage fluctuation and energy storage response, which reflects the power response required under different fluctuation degrees. The discharge optimization model is constructed based on historical operation data and contains 1,000 sets of voltage fluctuation and energy storage response records. Each set of records contains the voltage fluctuation degree, power output value, discharge depth and response time. The discharge optimization curve is obtained through training with these data, which reflects the best response characteristics of the energy storage system. The power regulation accuracy is divided into three intervals: the low power interval is 0 to 2 MW, and the adjustment step is 0.1 MW; the medium power interval is 2 to 5 MW, and the adjustment step is 0.2 MW; the high power interval is 5 to 10 MW, and the adjustment step is 0.5 MW. The corresponding power interval is selected according to the degree of voltage fluctuation. The discharge depth parameter is closely related to the battery life. During normal operation, the discharge depth is controlled between 20% and 80%. For severe voltage fluctuations, the upper limit of the discharge depth can be increased to 90%. At the same time, considering the remaining life of the battery, when the number of cycles exceeds 3,000 times, the upper limit of the discharge depth is reduced to 70%. The response speed parameter reflects the energy storage system's ability to follow voltage fluctuations. The response delay time is calculated. For small power regulation, the response time is controlled within 50 milliseconds; medium power regulation is controlled within 100 milliseconds; and high power regulation is controlled within 200 milliseconds. The response strategy matrix contains three dimensions: output power, discharge depth, and response speed. Each element in the matrix corresponds to a specific set of response parameters. For example, when a voltage drop of 15% at a rate of 3% per millisecond is detected, a response strategy with an output power of 4 MW, a discharge depth of 60%, and a response time of 80 milliseconds is selected. This strategy not only ensures a rapid response to voltage fluctuations, but also avoids excessive discharge of the energy storage system.

[0026] Furthermore, step S3 may also include: The current energy storage value, load value and system operating parameters of the energy storage system are obtained, and the objective function is constructed in combination with the preset number of time periods and electricity price table; the target discharge capacity and target energy storage value are set as constraints, and the linear programming algorithm is used to solve the objective function under the constraints to obtain the target discharge strategy set; the discharge strategies whose discharge capacity exceeds the preset capacity threshold in the target strategy set are screened out, the discharge capacity is adjusted, and the target discharge strategy is obtained.

[0027] Specifically, the current energy storage value and the real-time load value are obtained from the energy storage management device, the operating parameter value of the energy storage device is read through the data collector, and the electricity price data value is segmented according to the preset time period division number to obtain the electricity price data of the first time period; the minimum cost objective function is constructed according to the electricity price data of the first time period, and the constraint condition calculation is performed using the target discharge amount and the target energy storage value to obtain the first optimization constraint condition; the minimum cost objective function is linearly solved using the first optimization constraint condition to obtain the first discharge strategy data; the discharge amount in the first discharge strategy data is compared with the preset capacity threshold, and if the discharge amount exceeds the threshold, the excess part is recorded to obtain the first discharge deviation data; the discharge adjustment coefficient is calculated according to the first discharge deviation data, and the discharge amount is linearly interpolated using the discharge adjustment coefficient to generate the target discharge strategy.

[0028] More specifically, the current energy storage value and the real-time load value are obtained from the energy storage management device, the operating parameter value of the energy storage device is read through the data collector, and the electricity price data value is segmented according to the preset time period division number to obtain the electricity price data for the first time period. The minimum cost objective function is constructed according to the electricity price data for the first time period, and the constraint conditions are obtained from the target discharge amount and the target energy storage value to obtain the first optimization constraint conditions. The minimum cost objective function is linearly programmed using the first optimization constraint conditions to obtain the first discharge strategy data. The discharge amount in the first discharge strategy data is compared with the preset capacity threshold. If the discharge amount exceeds the threshold, the excess part is recorded to obtain the first discharge deviation data. The discharge adjustment coefficient is calculated based on the first discharge deviation data, and the discharge amount is linearly interpolated to obtain the second discharge strategy data. Constraint verification is performed on the second discharge strategy data. If the target energy storage value constraint is not met, the discharge amount adjustment is repeated until the constraint conditions are met. The target discharge strategy data is generated according to the discharge strategy that meets the constraint conditions.

[0029] Exemplarily, the energy storage management device monitors the operating status of the energy storage unit in real time. The current energy storage value is 8 MWh, the real-time load value is 12 MW, and the operating parameters include key indicators such as charging and discharging efficiency of 95%, maximum discharge power of 10 MW, and minimum remaining power of 20%. According to the grid dispatching requirements, 24 hours are divided into 6 time periods, and the electricity price data for each time period are 0.8 yuan, 1.2 yuan, 0.6 yuan, 0.9 yuan, 1.5 yuan, and 0.7 yuan per kilowatt-hour respectively. The principle of maximizing energy storage benefits is considered in the process of constructing the objective function. Discharge is given priority during high electricity price periods, and discharge is avoided during low electricity price periods. At the same time, the constraints are set as the target discharge amount does not exceed 6 MWh, and the target energy storage value is not less than 1.6 MWh. The preliminary discharge strategy is calculated by the linear programming solver, discharging 4 MWh in the period when the electricity price is 1.5 yuan, and discharging 2 MWh in the period when the electricity price is 1.2 yuan. The discharge strategy is tested for capacity thresholds. The preset capacity threshold is 3 MWh of maximum discharge in a single period. It is found that the discharge in the period of 1.5 yuan electricity price exceeds the threshold by 1 MWh. The deviation data is recorded for subsequent adjustment. The discharge adjustment coefficient of 0.75 is calculated based on the deviation data, and the excess is proportionally distributed to adjacent periods. The adjusted discharge strategy is to discharge 3 MWh in the period of 1.5 yuan electricity price, 2.5 MWh in the period of 1.2 yuan electricity price, and 0.5 MWh in the period of 0.9 yuan electricity price. The adjusted strategy is constrained and verified. The verification results show that the target energy storage value is not less than 1.6 MWh. The remaining power of the energy storage device after executing the discharge strategy is 2 MWh. The optimal solution is found by linear programming method under various constraints. The generation of the discharge strategy takes into account both economic efficiency and safe operation of the energy storage device. Multi-period joint optimization avoids excessive discharge in a single period, and the refined adjustment of the discharge amount ensures the rational use of the energy storage capacity. The implementation effect of the discharge strategy is reflected in many aspects. The discharge amount of the energy storage device is positively correlated with the electricity price level, and the discharge amount is larger during the high electricity price period; the discharge amount in each period did not exceed the rated power and capacity threshold of the equipment; after the execution of the entire discharge strategy, the remaining power of the energy storage device can still meet the minimum remaining power requirement; by adjusting the discharge amount in adjacent time periods, a smooth transition of the discharge plan was achieved.

[0030] Furthermore, step S4 may specifically include: Acquire three-phase voltage waveform data from a power grid monitoring device, and obtain first voltage distortion data by performing Fourier transform analysis on the three-phase voltage waveform data; Decomposing the three-phase voltage waveform data into symmetrical components, calculating the phase deviation value and the voltage amplitude using the least square method, and obtaining first imbalance data; Perform compensation calculation according to the first voltage distortion data and the first imbalance data, determine the compensation voltage amplitude and phase by a symmetrical component method, and obtain first harmonic compensation data and first phase sequence compensation data; An optimization calculation of charging capacity is performed on the first harmonic compensation data and the first phase sequence compensation data, and constraint verification is performed on compensation parameters through a preset compensation model to screen out a charging strategy for the energy storage system.

[0031] More specifically, the three-phase voltage waveform data is obtained from the power grid monitoring device, and the waveform is subjected to harmonic analysis through Fourier transform, and the first voltage distortion data is obtained according to the fundamental voltage value and each harmonic voltage calculation. The three-phase voltage data is decomposed into symmetrical components, and the phase deviation value and the voltage amplitude are calculated by the least square method to obtain the first unbalance data. The harmonic compensation target value is calculated according to the first voltage distortion data, and the harmonic content of each order is optimized and sorted to obtain the first harmonic compensation data. The three-phase compensation target value is calculated based on the first unbalance data, and the compensation voltage amplitude and phase are determined by the symmetrical component method to obtain the first phase sequence compensation data. The power factor is calculated for the first harmonic compensation data and the first phase sequence compensation data, and the charging capacity is calculated using a preset compensation optimization model to obtain the first charging compensation data. Harmonic suppression instructions, unbalance compensation instructions, and power factor correction instructions are generated according to the first charging compensation data to obtain the second charging compensation data. Constraint verification is performed on the second charging compensation data, and the compensation parameters are adjusted until the charging limit requirements of the energy storage device are met to obtain the target charging strategy data.

[0032] For example, in the three-phase voltage waveform collected by the power grid monitoring device, the fundamental voltage of phase A is 220 volts, containing 3% of the third harmonic, 4% of the fifth harmonic and 2% of the seventh harmonic. The harmonic components are decomposed by Fourier transform, and the total voltage distortion rate is calculated to be 5.4%, which exceeds the 5% limit specified in the national standard. In the process of symmetrical component decomposition, it is found that the three-phase voltage is unbalanced. The voltage of phase A is 220 volts, the voltage of phase B is 208 volts, and the voltage of phase C is 215 volts. The phase deviation values ​​are 2 degrees, -3 degrees and 1 degree respectively. The unbalance calculated by the least squares method is 6.2%, which is higher than the 2% limit required by the power quality standard. In response to the harmonic problem, the compensation optimization first deals with the fifth harmonic with a higher content, and sets the compensation target to reduce it to below 2%, then the third harmonic to below 1.5%, and finally the seventh harmonic to below 1%. This hierarchical compensation method gradually improves the quality of the voltage waveform. The three-phase unbalanced compensation adopts the symmetrical component method, and it is calculated that the B-phase voltage needs to be increased by 12 volts to reach 220 volts, and the phase is adjusted to 120 degrees. The C-phase voltage is increased by 5 volts to 220 volts, and the phase is adjusted to 240 degrees. This compensation method makes the three-phase voltage reach a symmetrical state. When determining the charging capacity, considering that harmonic compensation requires a capacity of 2 megawatts, unbalanced compensation requires a capacity of 1.5 megawatts, and the power factor is increased from 0.92 to 0.98. 1 megawatt capacity is required. Comprehensive calculations show that the energy storage device needs to provide 4.5 megawatts of charging power. In the process of generating compensation instructions, the harmonic suppression instruction includes the compensation amplitude and phase of each harmonic, the unbalanced compensation instruction includes the compensation voltage and phase adjustment of each phase, and the power factor correction instruction includes the reactive power compensation value. The constraint verification phase focuses on the charging limits of the energy storage device, including the maximum charging power not exceeding 5 MW, the battery terminal voltage not exceeding 850 volts during charging, and the charging current not exceeding 600 amperes. At the same time, considering the limitation that the battery state of charge does not exceed 90%, the compensation parameters are appropriately adjusted. The charging strategy formed realizes the comprehensive management of power quality while ensuring the safe operation of the energy storage device. By coordinating and controlling the three aspects of harmonic suppression, unbalanced compensation and power factor correction, the grid operation indicators meet the relevant standard requirements.

[0033] Furthermore, step S4 may also include: The harmonic content data is calculated from the harmonic content data in the transient electromagnetic signal, and the voltage distortion rate is calculated based on the harmonic content data; by measuring the instantaneous value of the three-phase voltage, the difference between the first target phase voltage and the second target phase voltage in the three-phase voltage, as well as the average value of the three-phase voltage are determined, and the three-phase voltage imbalance is calculated.

[0034] Specifically, the harmonic content data is obtained from a transient electromagnetic signal acquisition device, the harmonic content data is decomposed by fast Fourier transform, and the harmonic component data is calculated according to the fundamental voltage value and the harmonic order value; the amplitude statistics of the harmonic component data are performed, and the ratio of each harmonic voltage to the fundamental voltage is calculated, and then the voltage distortion rate is obtained by cumulative summation; the instantaneous value of the three-phase voltage is obtained from a three-phase voltage sampling device, and after the effective value of the instantaneous value of the three-phase voltage is calculated, the maximum value is selected as the first target phase voltage, and the minimum value is selected as the second target phase voltage; the voltage difference is calculated according to the first target phase voltage and the second target phase voltage to obtain the first phase-to-phase voltage deviation data, and the three-phase voltage imbalance is calculated by dividing the first phase-to-phase voltage deviation data by the average value of the three-phase voltage.

[0035] More specifically, the harmonic content data is obtained from the transient electromagnetic signal acquisition device, the signal is decomposed by fast Fourier transform, and the first harmonic component data is calculated according to the fundamental voltage value and the harmonic order value. The amplitude statistics of the first harmonic component data are performed, and the first voltage distortion data is calculated according to the ratio of each harmonic voltage to the fundamental voltage. The first voltage distortion data is accumulated and summed, and the voltage distortion rate is calculated according to the ratio of the total harmonic voltage to the fundamental voltage. The instantaneous value of the three-phase voltage is obtained from the three-phase voltage sampling device, and the three-phase voltage effective value sequence is calculated. The three-phase voltage effective value sequence is sorted according to the amplitude, and the maximum value is selected as the first target phase voltage, and the minimum value is selected as the second target phase voltage. The voltage difference is calculated according to the first target phase voltage and the second target phase voltage to obtain the first phase-to-phase voltage deviation data. The first phase-to-phase voltage deviation data is divided by the average value of the three-phase voltage to calculate the three-phase voltage imbalance.

[0036] For example, the transient electromagnetic signal contains rich harmonic information. The sampling device collects signals at a frequency of 10 kHz and records a data window of 200 milliseconds. Through fast Fourier transform, each harmonic component can be accurately extracted. The measured data of a distribution station shows that the fundamental voltage is 220 volts, the third harmonic voltage is 6.6 volts, the fifth harmonic voltage is 11 volts, and the seventh harmonic voltage is 4.4 volts. When performing harmonic component statistics, each harmonic voltage is divided by the fundamental voltage to obtain the harmonic content percentage. After the above data is converted, the third harmonic content is 3%, the fifth harmonic content is 5%, and the seventh harmonic content is 2%. These values ​​reflect the relative strength of each harmonic in the power grid, among which the fifth harmonic content is the highest. This phenomenon is more common in industrial areas where there are a large number of frequency conversion equipment. The calculation of the voltage distortion rate comprehensively considers the influence of all harmonics. The square root of the sum of the squares of each harmonic voltage is summed and then divided by the fundamental voltage to obtain a total harmonic distortion rate of 6.2%. This value exceeds the 5% limit specified by the national standard, indicating that the harmonic pollution of the power grid is relatively serious. In the three-phase voltage sampling, 1,000 sampling points are collected for each phase voltage to calculate the effective value. The effective value of phase A voltage is 225 volts, the effective value of phase B voltage is 210 volts, and the effective value of phase C voltage is 218 volts. Through sorting and comparison, it is found that the maximum value appears in phase A and the minimum value appears in phase B. These two phases are selected as the target phases for calculating the imbalance. The voltage difference between the two target phases is 15 volts, and the average value of the three-phase voltage is 217.7 volts. Dividing the voltage difference by the average value, the three-phase voltage imbalance is 6.9%, which greatly exceeds the 2% limit specified in the power quality standard. This imbalance phenomenon is often related to the unreasonable distribution of single-phase high-power loads. This imbalance is particularly prominent in industrial areas. If there are single-phase loads such as large arc furnaces and electrified railways, it often leads to more serious three-phase imbalance. In the power supply system of a steel plant, when the arc furnace starts working, the three-phase voltage imbalance will quickly rise to more than 8%, accompanied by obvious harmonic distortion, and the total harmonic distortion rate can reach 7.5%. The comprehensive evaluation of voltage distortion and three-phase imbalance has important guiding significance for improving power quality. By accurately calculating these two indicators, it can not only reflect the operating status of the power supply system, but also provide a basis for subsequent compensation measures. Actual operation data shows that through reasonable compensation means, the voltage distortion rate can be controlled within 4% and the three-phase imbalance can be reduced to less than 1.5%.

[0037] Furthermore, step S5 may specifically include: Obtaining harmonic current content and power factor benchmark data before the implementation of the charging and discharging strategy in the line monitoring device, and performing spectrum analysis on the benchmark data through Fourier transform to obtain first evaluation data; According to the first evaluation data, the line operation data after the implementation of the charging and discharging strategy is collected, and the harmonic current content change value is calculated by Fourier transform to obtain the second evaluation data; Calculating a power factor change value according to the second evaluation data, and obtaining third evaluation data through reactive power compensation calculation; A line loss calculator is used to comprehensively calculate the second evaluation data and the third evaluation data, and the energy loss value per unit length is integrated according to the line resistance parameter to obtain the total line energy loss data; If the harmonic current content in the total line energy loss data is lower than the reference value and the power factor is higher than the reference value, the current charging and discharging strategy is determined as the target energy-saving charging and discharging strategy.

[0038] More specifically, the harmonic current content and power factor benchmark data before the implementation of the charging and discharging strategy are obtained from the line monitoring device, and the benchmark data is subjected to spectrum analysis through Fourier transform to obtain the first evaluation data. The line operation data after the implementation of the charging and discharging strategy is collected, and the harmonic current content change value is calculated by Fourier transform to obtain the second evaluation data. The power factor change value is calculated according to the second evaluation data, and the third evaluation data is obtained by reactive power compensation calculation. The line loss calculator is used to perform a comprehensive calculation on the second evaluation data and the third evaluation data, and the energy loss value per unit length is obtained according to the line resistance parameter. The energy loss value per unit length is integrated to obtain the total energy loss data of the line. The total energy loss data of the line is judged according to the pre-set energy-saving benchmark. If the harmonic current content is lower than the benchmark value and the power factor is higher than the benchmark value, the current charging and discharging strategy parameters are recorded. The charging and discharging strategy parameters that meet the energy-saving benchmark are marked as target energy-saving charging and discharging strategy data.

[0039] For example, the line monitoring device collected the operation data of the 10 kV transmission line. Before the implementation of the charging and discharging strategy, there was obvious harmonic pollution in the line, among which the 3rd harmonic current content was 4.5%, the 5th harmonic current content was 6.2%, the 7th harmonic current content was 3.8%, and the power factor was 0.85. These data constituted the benchmark value for the evaluation. The data after the implementation of the strategy were analyzed by Fourier transform, and it was found that the 3rd harmonic current content was reduced to 2.8%, the 5th harmonic current content was reduced to 3.5%, the 7th harmonic current content was reduced to 2.2%, and the total harmonic current content was reduced from 8.5% to 5.0%, indicating that the charging and discharging strategy has a significant effect on harmonic suppression. The improvement of power factor is reflected in the compensation of reactive power. Through the charging and discharging process of the energy storage device, the line power factor is increased to 0.95, and the reactive power is compensated from the original 750 kN to 320 kN. This improvement directly affects the energy loss of the line. When calculating line losses, considering that the line resistance is 0.25 ohms per kilometer, for a 5-kilometer line section, the energy loss per unit length is reduced from the original 2.8 kilowatts per kilometer to 1.5 kilowatts per kilometer through the improvement of harmonic current and power factor. The calculation results of the total energy loss of the line show that the total loss before the implementation of the strategy was 14 kilowatts, and after implementation it was reduced to 7.5 kilowatts, a loss reduction of 46%. The pre-set energy-saving benchmark requires a 30% reduction in harmonic current content and an increase in power factor of more than 0.08. The implementation effect of the current strategy has exceeded this benchmark. The specific parameters of the energy-saving charging and discharging strategy include the operating instructions of the energy storage device, increasing the charging power to 2 megawatts when the harmonic content is high, and adjusting the charging current waveform through an intelligent control algorithm to make it opposite to the harmonic current phase, thereby achieving harmonic suppression. In terms of power factor correction, the energy storage device dynamically adjusts the charging and discharging power according to the load characteristics. The discharge power is controlled at 1.5 MW during peak hours and the charging power is maintained at 1 MW during off-peak hours, which not only ensures the improvement of power quality but also avoids excessive charging and discharging of the energy storage device. The determined target energy-saving charging and discharging strategy not only meets the preset energy-saving benchmark, but also shows good adaptability in actual operation. Even under conditions with large load fluctuations, it can still maintain stable harmonic suppression and power factor improvement effects, and the line energy loss is always maintained at a low level.

[0040] Furthermore, step S6 may specifically include: Decomposing the electromagnetic signal by fast Fourier transform, and obtaining first power spectrum density data according to spectrum distribution calculation; Performing instantaneous power calculation on the first power spectrum density data, and performing segmented processing on the instantaneous power curve using a sliding window to obtain first time series data; Performing sample division according to the first time series data, extracting voltage values ​​from the divided sample segments, and obtaining second time series data; Performing long short-term memory network training on the second time series data, extracting voltage change characteristics from historical data, and obtaining a first prediction parameter; The voltage swell rate and the voltage sag rate are calculated by using the first prediction parameter, and a threshold comparator is used to make a judgment to obtain a second prediction parameter; If the surge rate or the sag rate exceeds the preset threshold, the current voltage change data is recorded, and the power adjustment amount is generated according to the change amplitude to obtain the energy storage optimization parameters.

[0041] More specifically, transient electromagnetic signal data is obtained from a power grid monitoring device, the signal is decomposed by fast Fourier transform, and the first power spectrum density data is calculated according to the spectrum distribution. The instantaneous power is calculated for the first power spectrum density data, and the instantaneous power curve is segmented by a sliding window to obtain the first time series data. Samples are divided according to the first time series data, and the voltage value is extracted for each sample segment to obtain the second time series data. Long short-term memory network training is performed on the second time series data, and voltage change characteristics are extracted from historical data to obtain the first prediction parameter. The voltage swell rate and sag rate are calculated using the first prediction parameter, and the voltage is judged by a threshold comparator to obtain the second prediction parameter. The grid stability state is judged according to the second prediction parameter. If the swell rate or sag rate exceeds the preset threshold, the current voltage change data is recorded. The charge and discharge power of the recorded voltage change data is calculated, and the power adjustment amount is generated according to the change amplitude to obtain the energy storage optimization parameter. The charge and discharge power is adjusted using the energy storage optimization parameter to generate a new energy storage charge and discharge strategy parameter.

[0042] Exemplarily, the transient electromagnetic signal collected by the power grid monitoring device is recorded at a sampling frequency of 10 kHz, and the spectrum distribution is obtained after fast Fourier transform. The power spectrum density has an obvious main peak at the base frequency of 50 Hz, with an amplitude of 100 kW per Hz, and there are secondary peaks at 100 Hz and 150 Hz, with amplitudes of 15 kW per Hz and 8 kW per Hz, respectively. The calculation of the instantaneous power curve uses a 200 millisecond sliding window with a window overlap rate of 50%. 2000 sampling points are recorded in each window. This segmentation method not only ensures the continuity of the data, but also captures rapid power changes. Analysis of a certain period shows that the instantaneous power fluctuates between 5 MW and 7 MW, with obvious periodic changes. The time series samples are divided in a fixed length manner. Each sample contains 10 continuous sliding window data, which is equivalent to an observation time of 2 seconds. This division method allows each sample to contain a complete power fluctuation cycle. The extracted voltage values ​​show that the voltage fluctuates between 95% and 105% of the rated value of 220 volts during normal operation. The long short-term memory network adopts a three-layer structure with 128 hidden neurons. Through training, the voltage change law is mastered and the voltage change trend in the next 100 milliseconds is predicted. The root mean square error of the prediction result is controlled within 2%. The change characteristics extracted from the prediction results include the voltage change rate and duration. The calculation results of the voltage swell rate show that the voltage change rate under normal conditions does not exceed 1% of the nominal value per millisecond, and can reach 5% per millisecond when a disturbance occurs. The preset warning threshold is 3% per millisecond. Once the predicted value exceeds this threshold, it means that the grid may be unstable. The energy storage response strategy is calculated based on the voltage change data. When a voltage swell is predicted, the energy storage device increases the charging power from the original 1 MW to 2 MW, and the charging duration is dynamically adjusted according to the voltage change duration. On the contrary, when a voltage sag is predicted, the discharge power is increased to 2.5 MW. The optimized charging and discharging strategy showed good adaptability in actual operation. For voltage fluctuations lasting less than 50 milliseconds, the energy storage device can complete power adjustment within 20 milliseconds, effectively suppressing voltage fluctuations. For larger disturbances, the adjustment of charging and discharging power is divided into multiple steps, and the adjustment amplitude of each step is controlled within 0.5 megawatts, avoiding excessive response of the energy storage device. Actual measured data shows that the optimized control strategy reduces the voltage fluctuation amplitude by 40%, and the power fluctuation is effectively suppressed.

[0043] The present disclosure also provides a power quality energy-saving operation monitoring device based on AI Internet of Things multi-source data, including: A transient electromagnetic signal acquisition module, used to acquire a transient electromagnetic signal of a permanent magnetic operating mechanism circuit breaker; A time series characteristic parameter extraction module is used to extract the time series characteristic parameters of transient electromagnetic signals; A power quality disturbance association module is used to determine whether there is a harmonic problem caused by the inverter and whether a voltage surge or a voltage drop occurs, and to determine whether the timing characteristic parameter is associated with the power quality disturbance; A discharge strategy optimization module is used to obtain a voltage mutation amplitude and a voltage surge rate, obtain a voltage fluctuation degree according to the voltage mutation amplitude and the voltage surge rate, input the voltage fluctuation degree into a discharge optimization model, and output a discharge strategy for the energy storage system; The charging strategy optimization module is used to calculate the voltage distortion rate and the three-phase voltage imbalance, and select the charging strategy of the energy storage system through the pre-established charging optimization model; Energy-saving strategy evaluation module, used to evaluate the impact of charging strategy and discharging strategy on the harmonic current content of the transmission line and the energy loss corresponding to the power factor change, and determine the target energy-saving charging strategy and target energy-saving discharging strategy; The adjustment module is used to monitor in real time the instantaneous power curve and power spectrum density corresponding to the transient electromagnetic signals of the target energy-saving charging strategy and the target energy-saving discharging strategy, and perform time series analysis to obtain voltage value time series data; divide the time series data into samples of a set length, predict the sudden rise rate and sudden fall rate of the samples, obtain the prediction results, and form an energy-saving optimization plan for the energy storage charging and discharging strategy according to the prediction results.

[0044] In the description of the present disclosure, it is necessary to understand that the directions or positional relationships indicated by directional words such as "front, back, up, down, left, right", "lateral, vertical, horizontal" and "top, bottom" are usually based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present disclosure and simplifying the description. Unless otherwise specified, these directional words do not indicate or imply that the device or element referred to must have a specific orientation or be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the scope of protection of the present disclosure.

[0045] For those skilled in the art, various other corresponding changes and deformations can be made according to the technical solutions and concepts described above, and all of these changes and deformations should fall within the protection scope of the claims of the present disclosure.

Claims

1. A method for monitoring power quality and energy-saving operation based on AI IoT multi-source data, characterized in that: The following steps are involved: S1. Acquire a transient electromagnetic signal of a permanent magnetic operating mechanism circuit breaker and extract a time sequence characteristic parameter of the transient electromagnetic signal; S2, judging whether there is a harmonic problem caused by the inverter and whether a voltage surge or a voltage drop occurs, and determining whether the timing characteristic parameter is associated with the power quality disturbance; S3, obtaining a voltage mutation amplitude and a voltage surge rate, obtaining a voltage fluctuation degree according to the voltage mutation amplitude and the voltage surge rate, inputting the voltage fluctuation degree into a discharge optimization model, and outputting a discharge strategy for the energy storage system; S4. Calculate the voltage distortion rate and the three-phase voltage imbalance, and select the charging strategy of the energy storage system through the pre-established charging optimization model; S5. Evaluate the impact of charging strategy and discharging strategy on the harmonic current content of the transmission line and the energy loss corresponding to the power factor change, and determine the target energy-saving charging strategy and the target energy-saving discharging strategy; S6. Real-time monitoring of the instantaneous power curve and power spectrum density corresponding to the transient electromagnetic signals of the target energy-saving charging strategy and the target energy-saving discharging strategy, and performing time series analysis to obtain voltage value time series data; dividing the time series data into samples of a set length, predicting the sudden rise rate and sudden fall rate of the samples, obtaining prediction results, and forming an energy-saving optimization plan for the energy storage charging and discharging strategy according to the prediction results.

2. According to claim 1, the method for monitoring power quality and energy-saving operation based on AI Internet of Things multi-source data is characterized in that: The timing characteristic parameters include frequency components, amplitude changes, phase differences and harmonic content; The step S1 comprises: The electromagnetic torque stability parameter is calculated based on the original waveform data of the electromagnetic signal collected by the circuit breaker sensor and the magnetic ring strength of the permanent magnetic material to obtain a first transient electromagnetic signal sequence; Decomposing the first transient electromagnetic signal sequence in the frequency domain by fast Fourier transform, and generating a second transient electromagnetic signal sequence if the spectrum energy density is greater than the magnetic flux attenuation degree threshold; The third transient electromagnetic signal sequence is obtained by performing wavelet decomposition on the second transient electromagnetic signal sequence to obtain high-frequency coefficients and low-frequency coefficients, and extracting the electromagnetic displacement change and the switching state continuity characteristic curve from the high-frequency coefficients and the low-frequency coefficients. The recursive least squares method is used to fit the third transient electromagnetic signal sequence to obtain an amplitude change characteristic curve, and the fundamental component and harmonic component of the amplitude change characteristic curve are corrected by a phase compensation algorithm to obtain frequency component data and harmonic content data to construct a time series characteristic parameter matrix.

3. According to claim 1, the method for monitoring power quality and energy-saving operation based on AI Internet of Things multi-source data is characterized in that: The step S2 comprises: Acquire a transient electromagnetic signal from a power grid sampling device, perform Fourier series decomposition on the transient electromagnetic signal, calculate the cumulative energy of the frequency component within a preset frequency band, and obtain first frequency spectrum feature data based on the comparison of the frequency band energy increment with a preset threshold; Extracting the amplitude of the transient electromagnetic signal according to the first frequency spectrum characteristic data, calculating the amplitude change rate by discrete Fourier transform, and generating first voltage sudden change characteristic data if the amplitude change rate is greater than a preset amplitude change threshold; Extracting spectrum energy and transient response characteristics of the first voltage sudden change characteristic data by wavelet packet transformation to obtain second spectrum characteristic data; The second frequency spectrum feature data is analyzed by correlation calculation to extract time series correlation features. If the matching degree of the time series correlation features exceeds a preset threshold, it is determined that the transient electromagnetic signal is associated with the power quality disturbance caused by the new energy grid connection.

4. According to claim 1, the method for monitoring power quality and energy-saving operation based on AI Internet of Things multi-source data is characterized in that: The step S3 comprises: Acquire voltage mutation amplitude data and voltage surge rate data from the power grid sampling device, locate the mutation point of the voltage mutation amplitude data by wavelet edge detection, and obtain voltage fluctuation degree data; According to the voltage fluctuation degree data and the remaining capacity data of the energy storage device, a recursive least square method is used to perform data fitting to obtain a first power response parameter; The first power response parameter and the historical discharge data of the energy storage device are optimized and calculated by a pre-established discharge optimization model to obtain a discharge optimization curve and a second power response parameter; The discharge depth threshold of the energy storage device is calculated according to the second power response parameter, and the response strategy matrix is ​​constructed using the discharge depth threshold and the fluctuation response delay parameter to obtain the discharge strategy of the energy storage system; the discharge strategy includes determining the output power, discharge depth and speed of responding to voltage fluctuations of the energy storage system.

5. According to claim 4, the method for monitoring power quality and energy-saving operation based on AI Internet of Things multi-source data is characterized in that: The step S3 further comprises: The current energy storage value, load value and system operating parameters of the energy storage system are obtained, and the objective function is constructed in combination with the preset number of time periods and electricity price table; the target discharge capacity and target energy storage value are set as constraints, and a linear programming algorithm is used to solve the objective function under the constraints to obtain the target discharge strategy set; the discharge strategies whose discharge capacity exceeds the preset capacity threshold in the target strategy set are screened out, the discharge capacity is adjusted, and the target discharge strategy is obtained.

6. According to claim 1, the method for monitoring power quality and energy-saving operation based on AI Internet of Things multi-source data is characterized in that: The step S4 comprises: Acquire three-phase voltage waveform data from a power grid monitoring device, and obtain first voltage distortion data by performing Fourier transform analysis on the three-phase voltage waveform data; Decomposing the three-phase voltage waveform data into symmetrical components, calculating the phase deviation value and the voltage amplitude using the least square method, and obtaining first imbalance data; Perform compensation calculation according to the first voltage distortion data and the first imbalance data, determine the compensation voltage amplitude and phase by a symmetrical component method, and obtain first harmonic compensation data and first phase sequence compensation data; An optimization calculation of charging capacity is performed on the first harmonic compensation data and the first phase sequence compensation data, and constraint verification is performed on compensation parameters through a preset compensation model to screen out a charging strategy for the energy storage system.

7. According to claim 6, the method for monitoring power quality and energy-saving operation based on AI Internet of Things multi-source data is characterized in that: The step S4 further comprises: The harmonic content data is calculated from the harmonic content data in the transient electromagnetic signal, and the voltage distortion rate is calculated based on the harmonic content data; by measuring the instantaneous value of the three-phase voltage, the difference between the first target phase voltage and the second target phase voltage in the three-phase voltage, as well as the average value of the three-phase voltage are determined, and the three-phase voltage imbalance is calculated.

8. According to claim 1, the method for monitoring power quality and energy-saving operation based on AI Internet of Things multi-source data is characterized in that: The step S5 comprises: Obtaining harmonic current content and power factor benchmark data before the implementation of the charging and discharging strategy in the line monitoring device, and performing spectrum analysis on the benchmark data through Fourier transform to obtain first evaluation data; According to the first evaluation data, the line operation data after the implementation of the charging and discharging strategy is collected, and the harmonic current content change value is calculated by Fourier transform to obtain the second evaluation data; Calculating a power factor change value according to the second evaluation data, and obtaining third evaluation data through reactive power compensation calculation; A line loss calculator is used to comprehensively calculate the second evaluation data and the third evaluation data, and the energy loss value per unit length is integrated according to the line resistance parameter to obtain the total line energy loss data; If the harmonic current content in the total line energy loss data is lower than the reference value and the power factor is higher than the reference value, the current charging and discharging strategy is determined as the target energy-saving charging and discharging strategy.

9. According to claim 1, the method for monitoring power quality and energy-saving operation based on AI Internet of Things multi-source data is characterized in that: The step S6 comprises: Decomposing the electromagnetic signal by fast Fourier transform, and obtaining first power spectrum density data according to spectrum distribution calculation; Performing instantaneous power calculation on the first power spectrum density data, and performing segmented processing on the instantaneous power curve using a sliding window to obtain first time series data; Performing sample division according to the first time series data, extracting voltage values ​​from the divided sample segments, and obtaining second time series data; Performing long short-term memory network training on the second time series data, extracting voltage change characteristics from historical data, and obtaining a first prediction parameter; The voltage swell rate and the voltage sag rate are calculated by using the first prediction parameter, and a threshold comparator is used to make a judgment to obtain a second prediction parameter; If the surge rate or the sag rate exceeds the preset threshold, the current voltage change data is recorded, and the power adjustment amount is generated according to the change amplitude to obtain the energy storage optimization parameters.

10. A power quality energy-saving operation monitoring device based on AI Internet of Things multi-source data, characterized in that: include: A transient electromagnetic signal acquisition module, which is used to acquire a transient electromagnetic signal of a permanent magnetic operating mechanism circuit breaker; A time series characteristic parameter extraction module, wherein the time series characteristic parameter extraction module is used to extract the time series characteristic parameters of the transient electromagnetic signal; A power quality disturbance association module, which is used to determine whether there is a harmonic problem caused by the inverter and whether a voltage surge or a voltage drop occurs, and to determine whether the timing characteristic parameter is associated with the power quality disturbance; A discharge strategy optimization module, wherein the discharge strategy optimization module is used to obtain a voltage mutation amplitude and a voltage surge rate, obtain a voltage fluctuation degree according to the voltage mutation amplitude and the voltage surge rate, input the voltage fluctuation degree into a discharge optimization model, and output a discharge strategy for the energy storage system; A charging strategy optimization module, which is used to calculate the voltage distortion rate and the three-phase voltage imbalance, and select a charging strategy for the energy storage system through a pre-established charging optimization model; An energy-saving strategy evaluation module, which is used to evaluate the impact of charging strategy and discharging strategy on the harmonic current content of the transmission line and the energy loss corresponding to the power factor change, and determine the target energy-saving charging strategy and the target energy-saving discharging strategy; An adjustment module is used to monitor in real time the instantaneous power curve and power spectrum density corresponding to the transient electromagnetic signals of the target energy-saving charging strategy and the target energy-saving discharging strategy, and perform time series analysis to obtain voltage value time series data; divide the time series data into samples of a set length, predict the sudden rise rate and sudden fall rate of the samples, obtain the prediction results, and form an energy-saving optimization plan for the energy storage charging and discharging strategy according to the prediction results.

Images