High-voltage electricity-saving device and method

Abstract

The invention discloses a power-saving device and method for high-voltage power utilization, and relates to the technical field of power grid control, a self-adaptive dynamic control strategy and a load prediction function are introduced, and a reactive compensation module can adjust the compensation amount according to the real-time change of the load demand; when the load fluctuates violently, a response can be quickly made, unnecessary energy consumption is reduced, the main line and the load end are respectively provided with compensation equipment, and distributed compensation and main line compensation are dynamically matched through load fluctuation prediction and a self-adaptive control technology; when the load changes, the load end equipment preferentially and quickly responds, the coordination of compensation is effectively improved, redundant power loss is avoided, and the harmonic management module realizes accurate harmonic suppression through real-time harmonic data monitoring and parameter adaptive adjustment of the intelligent filter; the intelligent filter automatically adjusts the damping coefficient and the frequency gain according to the frequency and the power density of each harmonic component, and the harmonic component of the specific frequency is inhibited in time; and the harmonic wave fluctuation coping capability of the system is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid control, and in particular to a high-voltage electricity saving device and method. Background Art

[0002] Reactive power compensation and harmonic suppression in high-voltage power systems are key energy-saving measures aimed at reducing line losses and improving system operation stability. However, traditional centralized compensation methods lack sufficient dynamic response capabilities and mostly use fixed compensation adjustments, which are difficult to cope with rapidly changing load demands. When loads fluctuate, compensation equipment is difficult to respond in a timely manner, and over-compensation or under-compensation is prone to occur.

[0003] To this end, some traditional solutions set up centralized compensation equipment on high-voltage trunk lines, or install decentralized compensation devices at important load ends; however, in the case of large load fluctuations, this method is difficult to track load changes in a timely manner, the system response is delayed, and the compensation between the load end and the trunk line is difficult to achieve efficient coordination, resulting in additional energy consumption losses and affecting the improvement of energy-saving effects.

[0004] In terms of harmonic suppression, traditional methods mainly rely on fixed filters or passive filters to suppress harmonics of specific frequencies. Although such methods can reduce the impact of harmonics within a certain frequency band, the frequency and intensity of harmonic generation often fluctuate with load characteristics. Fixed filters are difficult to adapt to complex and variable harmonic distributions, resulting in harmonic problems beyond the frequency range that are difficult to effectively solve; in addition, some harmonic suppression equipment is easily affected by the strong current and high voltage in the high-voltage power system, and it is difficult to ensure the harmonic suppression effect when the load fluctuates greatly; therefore, a high-voltage power saving device and method are urgently needed to solve such problems. Summary of the invention

[0005] In view of the above existing problems, the present invention is proposed.

[0006] The present invention provides a high-voltage electricity saving device and method to solve the problems in existing high-voltage electricity saving solutions, such as lack of flexible dynamic response of reactive power compensation, single frequency coverage of harmonic suppression, and insufficient coordination between the two, which limits the improvement of overall energy efficiency of the system.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In one aspect, the present invention provides a high voltage electricity saving device, comprising:

[0009] Intelligent monitoring equipment, responsible for collecting and processing power parameters of each node in the high-voltage power system;

[0010] Intelligent monitoring equipment is installed on the main line, subnet and key load end, responsible for collecting current, voltage, power factor and harmonic data, using edge computing equipment to process the data in real time, extracting current and voltage fluctuations, power factor, reactive power demand, harmonic content and load change rate characteristics; and sending the extracted key feature data to the centralized control platform;

[0011] Centralized control platform to make short-term forecasts on the load demand of high-voltage power systems and identify load fluctuation trends and harmonic change trends in advance;

[0012] Reactive power compensation and harmonic suppression module, based on the output of the load prediction module, uses a dynamic control algorithm to adjust reactive power compensation in real time, and uses a compensation device to automatically adjust the compensation amount when the load changes;

[0013] The harmonic management module monitors and suppresses the harmonic components in the power-saving device to maintain stable power quality and avoid overheating and abnormal loss of equipment. The module continuously analyzes the harmonic content of each monitoring point and identifies possible harmonic sources and their fluctuation trends based on the harmonic spectrum characteristics.

[0014] Furthermore, in the centralized control platform, the key characteristic data output by the intelligent monitoring equipment is input, and the long short-term memory network LSTM model is used to make short-term predictions on the load demand and harmonic trend to generate a prediction curve; the prediction results are directly passed to the reactive power compensation and harmonic suppression module.

[0015] Furthermore, in the harmonic management module, the intelligent filter automatically adjusts the filtering parameters according to the real-time harmonic data to suppress specific harmonic components.

[0016] In another aspect, the present invention provides a method for saving high voltage electricity, comprising:

[0017] Step S1, multi-point electricity consumption data collection,

[0018] Intelligent monitoring equipment is installed on the trunk line, subnet and load end to collect preliminary data in real time, including current, voltage, power factor and harmonic data; each monitoring device uses edge computing equipment to process preliminary data and send key feature data to the centralized control platform to reduce network load and data transmission delay;

[0019] Step S2, load fluctuation prediction,

[0020] Using the data received from the centralized control platform, the LSTM long short-term memory network model is used to perform short-term load fluctuation forecasts and generate forecast curves for load demand and harmonic trends. The forecast results are automatically output to the reactive power compensation module.

[0021] Step S3, adaptive reactive power compensation,

[0022] Based on the load prediction result of step S2, an adaptive dynamic control strategy is introduced into the reactive power compensation module to compensate for reactive power;

[0023] The reactive power compensation module automatically adjusts the compensation amount when the load changes;

[0024] Step S4, harmonic optimization and dynamic suppression,

[0025] Continuously monitor the harmonic content of each monitoring point, and use the harmonic analysis results to detect harmonic fluctuations at the load end; based on the harmonic data, the intelligent filter adjusts the filtering parameters in real time to suppress various harmonic components.

[0026] Furthermore, in step S1, the key characteristic data include current and voltage fluctuation characteristics, power factor, reactive power demand, harmonic content and spectrum characteristics, and load change rate.

[0027] Furthermore, in step S2, the method of generating the prediction curve of load demand and harmonic trend is as follows:

[0028] The centralized control platform receives data from intelligent monitoring devices, including current, voltage, power factor, reactive power demand, harmonic content, and load change rate. At each time step t, each feature is combined into a feature vector: in, represents the feature vector at time step t, I t Indicates the current value, V t Indicates voltage value, PF t Indicates power factor, Q t Represents reactive power demand, H t Indicates the harmonic content, R t Indicates the load change rate;

[0029] Arrange the data of the past n time steps into an input sequence As the input of the LSTM model, in, Represents the input sequence of time step t, including the feature vectors of the past n steps; in each LSTM unit, the input gate, forget gate and output gate are used to filter and transmit data, and the hidden state h is updated. t and cell state c t , to model the load change trend, the LSTM model architecture is:

[0030] Input Gate: Among them, i t represents the input gate value, σ represents the Sigmoid activation function, W i Represents the weight matrix input to the input gate, U i represents the weight matrix from the previous hidden state to the input gate, ht-1 represents the hidden state of the previous time step, b i represents the bias of the input gate;

[0031] Forget Gate: Among them, f t Represents the forget gate value, which controls the degree of retention of the previous state in the current time step, W f Represents the weight matrix input to the forget gate, U f represents the weight matrix from the previous hidden state to the forget gate, b f Represents the bias of the forget gate;

[0032] Candidate cell states: in, represents the candidate cell state, tanh represents the activation function, W c Represents the weight matrix input to the candidate cell state, U c Represents the weight matrix from the previous hidden state to the candidate cell state, b c represents the bias of the candidate cell state;

[0033] Update cell status: Among them, c t represents the cell state at the current time step, c t-1 represents the cell state at the previous time step, · represents element-wise multiplication;

[0034] Output Gate: Among them, t Represents the output gate value, W o Represents the weight matrix from input to output gate, U o represents the weight matrix from the previous hidden state to the output gate, b o represents the bias of the output gate;

[0035] Update hidden state: h t =o t ·tanh(c t ), where h t represents the hidden state of the current time step, tanh(c t ) represents the value of the current cell state after being adjusted by the activation function;

[0036] Finally, the hidden state h is passed through the output layer t Converted to target forecast values, i.e. load demand and harmonic trend at future time steps: in, represents the prediction of load demand and harmonic trend in the next time step, W y represents the weight matrix from the hidden state to the output layer, b y Represents the bias of the output layer.

[0037] Furthermore, in step S3, the adaptive reactive power compensation method is:

[0038] Determine the compensation benchmark of the compensation module based on the reactive power prediction result of the next time step obtained from the LSTM model output in step S2 That is, the predicted reactive power demand at the next time step;

[0039] Calculate the reactive power error, which is the difference between the predicted reactive power demand and the current actual reactive power. Among them, E t represents the reactive power error of the current time step, α represents the weighting factor of the real-time error, represents the reactive power demand predicted in the next time step, Q t represents the actual reactive power at the current time step, β represents the dynamic factor of the historical error, and E t-1 Represents the reactive power error at the previous time step.

[0040] Furthermore, in step S3, the adaptive reactive power compensation method further includes:

[0041] Calculate the reactive power compensation amount based on the reactive power error. Among them, C t Indicates the reactive compensation amount of the current time step, C t-1 represents the reactive compensation amount in the previous time step, k p Represents the proportional gain coefficient, E t Represents the reactive power error of the current time step, k i represents the integral gain coefficient, Represents the cumulative value of reactive power error from the initial time step to the current time step;

[0042] The calculated reactive power compensation amount C t Transmitted to the compensation device, the device according to C t The compensation amount is adjusted dynamically according to the change of load to achieve real-time response to load fluctuations.

[0043] Furthermore, in step S4, the intelligent filter adjusts the filtering parameters in real time according to the harmonic analysis results, specifically:

[0044] The centralized control platform collects harmonic current components from each monitoring point and forms a harmonic vector at time step t Contains the amplitude of multiple harmonic frequency components, in, represents the harmonic vector at time step t, h t,N represents the amplitude of the nth-order harmonic component at time step t, and N represents the total number of harmonic orders monitored;

[0045] The harmonic component vector is obtained by using the Fast Fourier Transform (FFT). Perform spectrum analysis to obtain the power density spectrum of each harmonic component, P t,n =|h t,n | 2 , where P t,n represents the power density of the nth-order harmonic component at time step t, h t,n represents the amplitude of the nth order harmonic;

[0046] Set dynamic thresholds based on historical harmonic mean and standard deviation: T t,n =μ n +γσ n , where T t,n represents the nth-order harmonic dynamic threshold at time step t, μ n represents the historical mean of the nth order harmonic, σ n represents the historical standard deviation of the nth order harmonic, γ represents the safety factor, which adjusts the sensitivity of the threshold; when P t,n >T t,n When the nth-order harmonic fluctuation exceeds the preset range, the filter is adjusted.

[0047] Further, in step S4, for the detected abnormal harmonic components, the intelligent filter is used according to the frequency f of each order harmonic component. n and power density P t,n Automatically adjust the filter damping factor ζ n and frequency gain G n , suppress specific harmonics, specifically:

[0048] Adjust the damping coefficient: n =ζ0+η·(P t,n -T t,n ), where ζ n represents the damping coefficient of the nth order harmonic, ζ0 represents the basic damping coefficient, η represents the dynamic gain factor, and adjusts the response intensity when the power density exceeds the threshold. t,n -T t,n Indicates the magnitude by which the power density exceeds the threshold;

[0049] To adjust the frequency gain: G n represents the frequency gain of the nth order harmonic, G0 represents the fundamental frequency gain, T t,n / P t,n Indicates dynamic adjustment ratio;

[0050] Record the harmonic suppression effect of the filter after adjustment and calculate the actual reduction value of the harmonic component after adjustment:

[0051] ΔH t,n =ht,n -h t+1,n , where ΔH t,n Indicates the amplitude reduction of the nth-order harmonic component, h t+1,n Indicates the amplitude of the nth harmonic in the next time step after adjustment; if ΔH t,n If the expected suppression level is not reached, feedback is given to the filter to further adjust ζ n and G n The value of .

[0052] The beneficial effects of the present invention are:

[0053] The present invention introduces an adaptive dynamic control strategy and a load prediction function, and the reactive power compensation module can adjust the compensation amount according to the real-time change of the load demand; it can respond quickly when the load fluctuates violently, thereby reducing unnecessary energy consumption.

[0054] The present invention sets compensation equipment on the trunk line and the load end respectively, and dynamically coordinates distributed compensation with trunk compensation through load fluctuation prediction and adaptive control technology; when the load changes, the load end equipment responds quickly first, while the trunk compensation equipment makes fine adjustments according to the overall load conditions, effectively improving the coordination of compensation and avoiding unnecessary power loss.

[0055] The present invention, the harmonic management module realizes accurate harmonic suppression through real-time harmonic data monitoring and adaptive adjustment of parameters of the intelligent filter; the intelligent filter automatically adjusts the damping coefficient and frequency gain according to the frequency and power density of each harmonic component, and timely suppresses the harmonic components of specific frequencies; and improves the system's ability to cope with harmonic fluctuations.

[0056] The present invention uses the LSTM model to make short-term predictions on load demand and harmonic trends, so that reactive power compensation and harmonic suppression can respond to expected fluctuations in advance and reduce the system's dependence on sudden changes. By identifying the trends of loads and harmonics in advance and adjusting compensation and filtering parameters in advance, the overall efficiency and energy-saving effect are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0058] Figure 1 It is a schematic structural diagram of the high-voltage electricity saving device of the present invention;

[0059] Figure 2 It is a schematic diagram of the flow chart of the high voltage electricity saving method of the present invention. DETAILED DESCRIPTION

[0060] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0061] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0062] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0063] Example 1, reference Figure 1 , this embodiment provides a high voltage electricity saving device, comprising:

[0064] Intelligent monitoring equipment, responsible for collecting and processing power parameters of each node in the high-voltage power system;

[0065] Intelligent monitoring equipment is installed on the trunk line, subnet and key load end, responsible for collecting current, voltage, power factor and harmonic data, using edge computing equipment to process the data in real time, extracting current and voltage fluctuations, power factor, reactive power demand, harmonic content and load change rate characteristics; and sending the extracted key feature data to the centralized control platform;

[0066] Centralized control platform to make short-term forecasts on the load demand of high-voltage power systems and identify load fluctuation trends and harmonic change trends in advance;

[0067] In the centralized control platform, key characteristic data output by intelligent monitoring equipment is input, and the long short-term memory network LSTM model is used to make short-term predictions on load demand and harmonic trends to generate prediction curves; the prediction results are directly transmitted to the reactive power compensation and harmonic suppression module;

[0068] Reactive power compensation and harmonic suppression module, based on the output of the load prediction module, uses a dynamic control algorithm to adjust reactive power compensation in real time, and uses a compensation device to automatically adjust the compensation amount when the load changes;

[0069] The harmonic management module monitors and suppresses the harmonic components in the power-saving device to maintain stable power quality and avoid overheating and abnormal loss of equipment. The module continuously analyzes the harmonic content of each monitoring point and identifies possible harmonic sources and their fluctuation trends based on the harmonic spectrum characteristics.

[0070] In the harmonic management module, the intelligent filter automatically adjusts the filtering parameters according to the real-time harmonic data to suppress specific harmonic components.

[0071] Example 2, reference Figure 2 This embodiment provides a high voltage electricity saving method, comprising the following steps:

[0072] Step S1, multi-point electricity consumption data collection,

[0073] Intelligent monitoring equipment is installed on the trunk line, subnet and load end to collect preliminary data in real time, including current, voltage, power factor and harmonic data; each monitoring device uses edge computing equipment to process preliminary data and send key feature data to the centralized control platform to reduce network load and data transmission delay;

[0074] In step S1, the key characteristic data include current and voltage fluctuation characteristics, power factor, reactive power demand, harmonic content and spectrum characteristics, and load change rate;

[0075] Specifically, edge computing is combined with centralized control to reduce the computing pressure of the centralized control platform. Key feature data is collected in layers, so that the centralized control platform can accurately identify the power parameters and load changes in each area, thereby improving the efficiency of system resource utilization. In addition, multi-level, distributed intelligent monitoring and compensation make the system more sensitive to local load regulation, realize the coordinated optimization of trunk line and load-end compensation, and reduce the operating loss of the overall system.

[0076] Step S2, load fluctuation prediction,

[0077] Using the data received from the centralized control platform, the LSTM long short-term memory network model is used to perform short-term load fluctuation forecasts and generate forecast curves for load demand and harmonic trends. The forecast results are automatically output to the reactive power compensation module.

[0078] In step S2, the method of generating the prediction curve of load demand and harmonic trend is:

[0079] The centralized control platform receives data from intelligent monitoring devices, including current, voltage, power factor, reactive power demand, harmonic content, and load change rate. At each time step t, each feature is combined into a feature vector: in, represents the feature vector at time step t, I t Indicates the current value, V t Indicates voltage value, PF t Indicates power factor, Q t Represents reactive power demand, H t Indicates the harmonic content, Rt Indicates the load change rate;

[0080] Arrange the data of the past n time steps into an input sequence As the input of the LSTM model, in, Represents the input sequence of time step t, including the feature vectors of the past n steps; in each LSTM unit, the input gate, forget gate and output gate are used to filter and transmit data, and the hidden state h is updated. t and cell state c t , to model the load change trend, the LSTM model architecture is:

[0081] Input Gate: Among them, i t represents the input gate value, σ represents the Sigmoid activation function, W i Represents the weight matrix input to the input gate, U i represents the weight matrix from the previous hidden state to the input gate, h t-1 represents the hidden state of the previous time step, b i represents the bias of the input gate;

[0082] Forget Gate: Among them, f t Represents the forget gate value, which controls the degree of retention of the previous state in the current time step, W f Represents the weight matrix input to the forget gate, U f represents the weight matrix from the previous hidden state to the forget gate, b f Represents the bias of the forget gate;

[0083] Candidate cell states: in, represents the candidate cell state, tanh represents the activation function, W c Represents the weight matrix input to the candidate cell state, U c Represents the weight matrix from the previous hidden state to the candidate cell state, b c represents the bias of the candidate cell state;

[0084] Update cell status: Among them, c t represents the cell state at the current time step, c t-1 represents the cell state at the previous time step, · represents element-wise multiplication;

[0085] Output Gate: Among them, t Represents the output gate value, W o Represents the weight matrix from input to output gate, U orepresents the weight matrix from the previous hidden state to the output gate, b o represents the bias of the output gate;

[0086] Update hidden state: h t =o t ·tanh(c t ), where h t represents the hidden state of the current time step, tanh(c t ) represents the value of the current cell state after being adjusted by the activation function;

[0087] Finally, the hidden state h is passed through the output layer t Converted to target forecast values, i.e. load demand and harmonic trend at future time steps: in, represents the prediction of load demand and harmonic trend in the next time step, W y represents the weight matrix from the hidden state to the output layer, b y represents the bias of the output layer;

[0088] Specifically, the intelligent reactive power compensation module responds to changes in reactive power demand in real time to avoid over-compensation or under-compensation. The adaptive dynamic control strategy adjusts the compensation amount in real time based on the load fluctuation prediction results, greatly reducing energy loss when load fluctuations are frequent. The harmonic suppression module effectively reduces the reactive power loss caused by harmonics.

[0089] Step S3, adaptive reactive power compensation,

[0090] Based on the load prediction result of step S2, an adaptive dynamic control strategy is introduced into the reactive power compensation module to compensate for reactive power;

[0091] The reactive power compensation module automatically adjusts the compensation amount when the load changes;

[0092] In step S3, the adaptive reactive power compensation method is:

[0093] Determine the compensation benchmark of the compensation module based on the reactive power prediction result of the next time step obtained from the LSTM model output in step S2 That is, the predicted reactive power demand at the next time step;

[0094] Calculate the reactive power error, which is the difference between the predicted reactive power demand and the current actual reactive power. Among them, E t represents the reactive power error of the current time step, α represents the weighting factor of the real-time error, represents the reactive power demand predicted in the next time step, Q trepresents the actual reactive power at the current time step, β represents the dynamic factor of the historical error, and E t-1 represents the reactive power error of the previous time step;

[0095] In step S3, the adaptive reactive power compensation method further includes:

[0096] Calculate the reactive power compensation amount based on the reactive power error. Among them, C t Indicates the reactive compensation amount of the current time step, C t-1 represents the reactive compensation amount in the previous time step, k p Represents the proportional gain coefficient, E t Represents the reactive power error of the current time step, k i represents the integral gain coefficient, Represents the cumulative value of reactive power error from the initial time step to the current time step;

[0097] The calculated reactive power compensation amount C t Transmitted to the compensation device, the device according to C t Dynamically adjust the compensation amount according to the change of load to achieve real-time response to load fluctuations;

[0098] Specifically, the reactive power compensation and harmonic suppression modules perform adaptive control and dynamically adjust the compensation amount and filtering parameters based on edge computing devices, reducing the energy consumption caused by voltage fluctuations and significantly improving the stability of the system.

[0099] Step S4, harmonic optimization and dynamic suppression,

[0100] Continuously monitor the harmonic content of each monitoring point, and use the harmonic analysis results to detect the harmonic fluctuations at the load end; based on the harmonic data, the intelligent filter adjusts the filtering parameters in real time to suppress various harmonic components;

[0101] In step S4, the intelligent filter adjusts the filtering parameters in real time according to the harmonic analysis results. Specifically:

[0102] The centralized control platform collects harmonic current components from each monitoring point and forms a harmonic vector at time step t Contains the amplitude of multiple harmonic frequency components, in, represents the harmonic vector at time step t, h t,N represents the amplitude of the nth-order harmonic component at time step t, and N represents the total number of harmonic orders monitored;

[0103] The harmonic component vector is obtained by using the Fast Fourier Transform (FFT). Perform spectrum analysis to obtain the power density spectrum of each harmonic component, P t,n =|h t,n |2 , where P t,n represents the power density of the nth-order harmonic component at time step t, h t,n represents the amplitude of the nth order harmonic;

[0104] Set dynamic thresholds based on historical harmonic mean and standard deviation: T t,n =μ n +γσ n , where T t,n represents the nth-order harmonic dynamic threshold at time step t, μ n represents the historical mean of the nth order harmonic, σ n represents the historical standard deviation of the nth order harmonic, γ represents the safety factor, which adjusts the sensitivity of the threshold; when P t,n >T t,n When the nth-order harmonic fluctuation exceeds the preset range, the filter is adjusted;

[0105] In step S4, for the detected abnormal harmonic components, the intelligent filter is used according to the frequency f of each order harmonic component. n and power density P t,n Automatically adjust the filter damping factor ζ n and frequency gain G n , suppress specific harmonics, specifically:

[0106] Adjust the damping coefficient: n =ζ0+η·(P t,n -T t,n ), where ζ n represents the damping coefficient of the nth order harmonic, ζ0 represents the basic damping coefficient, and η represents the dynamic gain factor, which adjusts the response intensity when the power density exceeds the threshold, P t,n -T t,n Indicates the magnitude by which the power density exceeds the threshold;

[0107] To adjust the frequency gain: G n represents the frequency gain of the nth order harmonic, G0 represents the fundamental frequency gain, T t,n / P t,n Indicates dynamic adjustment ratio;

[0108] Record the harmonic suppression effect of the filter after adjustment and calculate the actual reduction value of the harmonic component after adjustment:

[0109] ΔH t,n =h t,n -h t+1,n , where ΔH t,n Indicates the amplitude reduction of the nth-order harmonic component, h t+1,n Indicates the amplitude of the nth harmonic in the next time step after adjustment; if ΔHt,n If the expected suppression level is not reached, feedback is given to the filter to further adjust ζ n and G n The value of

[0110] Specifically, the long short-term memory network LSTM and fast Fourier transform FFT are used to predict future load demand and harmonic change trends based on historical load fluctuation data, and identify possible power parameter fluctuations in advance; the intelligent filter continuously and adaptively adjusts during harmonic monitoring to ensure the harmonic suppression effect while improving the operating stability of the equipment in a high-voltage environment; for the detected abnormal harmonic components, the filtering parameters and damping coefficients are intelligently adjusted within the threshold.

[0111] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A high voltage electricity saving device, characterized in that: include, Intelligent monitoring equipment, responsible for collecting and processing power parameters of each node in the high-voltage power system; Intelligent monitoring equipment is installed on the main line, subnet and key load end, responsible for collecting current, voltage, power factor and harmonic data, using edge computing equipment to process the data in real time, extracting current and voltage fluctuations, power factor, reactive power demand, harmonic content and load change rate characteristics; and sending the extracted key feature data to the centralized control platform; Centralized control platform to make short-term forecasts on the load demand of high-voltage power systems and identify load fluctuation trends and harmonic change trends in advance; Reactive power compensation and harmonic suppression module, based on the output of the load prediction module, uses a dynamic control algorithm to adjust reactive power compensation in real time, and uses a compensation device to automatically adjust the compensation amount when the load changes; The harmonic management module monitors and suppresses the harmonic components in the power-saving device. The module continuously analyzes the harmonic content of each monitoring point and identifies possible harmonic sources and their fluctuation trends based on the harmonic spectrum characteristics.

2. A high voltage electricity saving device according to claim 1, characterized in that: In the centralized control platform, the key characteristic data output by the intelligent monitoring equipment is input, and the long short-term memory network LSTM model is used to make short-term predictions on load demand and harmonic trends to generate prediction curves; the prediction results are directly passed to the reactive power compensation and harmonic suppression module.

3. A high voltage electricity saving device according to claim 2, characterized in that: In the harmonic management module, the intelligent filter automatically adjusts the filtering parameters according to the real-time harmonic data to suppress specific harmonic components.

4. A method for saving high voltage electricity, characterized in that: A high voltage electricity saving device according to any one of claims 1 to 3 comprises the following steps: Step S1, multi-point electricity consumption data collection, Intelligent monitoring equipment is installed on the trunk line, subnet and load end to collect preliminary data in real time, including current, voltage, power factor and harmonic data; each monitoring device uses edge computing equipment to process preliminary data and send key feature data to the centralized control platform; Step S2, load fluctuation prediction, Using the data received from the centralized control platform, the LSTM long short-term memory network model is used to perform short-term load fluctuation forecasts and generate forecast curves for load demand and harmonic trends. The forecast results are automatically output to the reactive power compensation module. Step S3, adaptive reactive power compensation, Based on the load prediction result of step S2, an adaptive dynamic control strategy is introduced into the reactive power compensation module to compensate for reactive power; The reactive power compensation module automatically adjusts the compensation amount when the load changes; Step S4, harmonic optimization and dynamic suppression, Continuously monitor the harmonic content of each monitoring point, and use the harmonic analysis results to detect harmonic fluctuations at the load end; based on the harmonic data, the intelligent filter adjusts the filtering parameters in real time to suppress various harmonic components.

5. A high voltage electricity saving method according to claim 4, characterized in that: In step S1, key characteristic data include current and voltage fluctuation characteristics, power factor, reactive power demand, harmonic content and spectrum characteristics, and load change rate.

6. A high voltage electricity saving method according to claim 5, characterized in that: In step S2, the method of generating the prediction curve of load demand and harmonic trend is: The centralized control platform receives data from intelligent monitoring devices, including current, voltage, power factor, reactive power demand, harmonic content, and load change rate. At each time step t, each feature is combined into a feature vector: in, represents the feature vector at time step t, I t Indicates the current value, V t Indicates voltage value, PF t Indicates power factor, Q t Represents reactive power demand, H t Indicates the harmonic content, R t Indicates the load change rate; Arrange the data of the past n time steps into an input sequence As the input of the LSTM model, in, Represents the input sequence of time step t, including the feature vectors of the past n steps; in each LSTM unit, the input gate, forget gate and output gate are used to filter and transmit data, and the hidden state h is updated. t and cell state c t , to model the load change trend, the LSTM model architecture is: Input Gate: Among them, i t represents the input gate value, σ represents the Sigmoid activation function, W i Represents the weight matrix input to the input gate, U i represents the weight matrix from the previous hidden state to the input gate, h t-1 represents the hidden state of the previous time step, b i represents the bias of the input gate; Forget Gate: Among them, f t Represents the forget gate value, which controls the degree of retention of the previous state in the current time step, W f Represents the weight matrix input to the forget gate, U f represents the weight matrix from the previous hidden state to the forget gate, b f represents the bias of the forget gate; Candidate cell states: in, represents the candidate cell state, tanh represents the activation function, W c Represents the weight matrix input to the candidate cell state, U c Represents the weight matrix from the previous hidden state to the candidate cell state, b c represents the bias of the candidate cell state; Update cell status: Among them, c t represents the cell state at the current time step, c t-1 represents the cell state at the previous time step, · represents element-wise multiplication; Output Gate: Among them, t Represents the output gate value, W o Represents the weight matrix from input to output gate, U o represents the weight matrix from the previous hidden state to the output gate, b o represents the bias of the output gate; Update hidden state: h t =o t ·tanh(c t ), where h t represents the hidden state of the current time step, tanh(c t ) represents the value of the current cell state after being adjusted by the activation function; Finally, the hidden state h is passed through the output layer t Converted to target forecast values, i.e. load demand and harmonic trend at future time steps: in, represents the prediction of load demand and harmonic trend in the next time step, W y represents the weight matrix from the hidden state to the output layer, b y Represents the bias of the output layer.

7. A high voltage electricity saving method according to claim 6, characterized in that: In step S3, the adaptive reactive power compensation method is: Determine the compensation benchmark of the compensation module based on the reactive power prediction result of the next time step obtained from the LSTM model output in step S2 That is, the predicted reactive power demand at the next time step; Calculate the reactive power error, which is the difference between the predicted reactive power demand and the current actual reactive power. Among them, E t represents the reactive power error of the current time step, α represents the weighting factor of the real-time error, represents the reactive power demand predicted in the next time step, Q t represents the actual reactive power at the current time step, β represents the dynamic factor of the historical error, and E t-1 Represents the reactive power error at the previous time step.

8. A high voltage electricity saving method according to claim 7, characterized in that: In step S3, the adaptive reactive power compensation method further includes: Calculate the reactive power compensation amount based on the reactive power error. Among them, C t Indicates the reactive compensation amount of the current time step, C t-1 represents the reactive compensation amount in the previous time step, k p Represents the proportional gain coefficient, E t represents the reactive power error of the current time step, k i represents the integral gain coefficient, Represents the cumulative value of reactive power error from the initial time step to the current time step; The calculated reactive power compensation amount C t Transmitted to the compensation device, the device according to C t The compensation amount is adjusted dynamically according to the changes in 9. A high voltage electricity saving method according to claim 8, characterized in that: In step S4, the intelligent filter adjusts the filtering parameters in real time according to the harmonic analysis results. Specifically: The centralized control platform collects harmonic current components from each monitoring point and forms a harmonic vector at time step t Contains the amplitude of multiple harmonic frequency components, in, represents the harmonic vector at time step t, h t,N represents the amplitude of the nth-order harmonic component at time step t, and N represents the total number of harmonic orders monitored; The harmonic component vector is obtained by using the Fast Fourier Transform (FFT). Perform spectrum analysis to obtain the power density spectrum of each harmonic component, P t,n =|h t,n | 2 , where P t,n represents the power density of the nth-order harmonic component at time step t, h t,n represents the amplitude of the nth order harmonic; Set dynamic thresholds based on historical harmonic mean and standard deviation: T t,n =μ n +γσ n , where T t,n represents the nth-order harmonic dynamic threshold at time step t, μ n represents the historical mean of the nth order harmonic, σ n represents the historical standard deviation of the nth order harmonic, γ represents the safety factor, and adjusts the sensitivity of the threshold; when P t,n >T t,n When the nth-order harmonic fluctuation exceeds the preset range, the filter is adjusted.

10. A high voltage electricity saving method according to claim 9, characterized in that: In step S4, for the detected abnormal harmonic components, the intelligent filter is used according to the frequency f of each order harmonic component. n and power density P t,n Automatically adjust the filter damping factor ζ n and frequency gain G n , suppress specific harmonics, specifically: Adjust the damping coefficient: n =ζ0+η·(P t,n -T t,n ), where ζ n represents the damping coefficient of the nth order harmonic, ζ0 represents the basic damping coefficient, η represents the dynamic gain factor, and adjusts the response intensity when the power density exceeds the threshold. t,n -T t,n Indicates the magnitude by which the power density exceeds the threshold; To adjust the frequency gain: G n represents the frequency gain of the nth order harmonic, G0 represents the fundamental frequency gain, T t,n / P t,n Indicates dynamic adjustment ratio; Record the harmonic suppression effect of the filter after adjustment and calculate the actual reduction value of the harmonic component after adjustment: ΔH t,n =h t,n -h t+1,n , where ΔH t,n Indicates the amplitude reduction of the nth-order harmonic component, h t+1,n Indicates the amplitude of the nth harmonic in the next time step after adjustment; if ΔH t,n If the expected suppression level is not reached, feedback is given to the filter to further adjust ζ n and G n The value of .

Images