Power distribution network electric energy quality optimization control method and control system

Abstract

The invention discloses a power distribution network electric energy quality optimization control method and control system. The method comprises the steps that standard multi-source real-time data in the operation process of a power distribution network are acquired; performing feature extraction on the standard multi-source real-time data to obtain multi-source real-time features; the multi-source real-time features are input into the trained power distribution network power quality optimization control model for solving, and an optimization control scheme is obtained; and according to the optimization control scheme, controlling each device in the power distribution network to realize power quality optimization control of the power distribution network. Wherein the power distribution network power quality optimization control model takes a voltage qualified rate, harmonic suppression and three-phase balance as optimization objectives to construct a multi-objective optimization function, and takes a voltage amplitude, a total harmonic distortion rate and a current unbalance degree as constraint conditions. The industrial problems that a traditional method is poor in adaptability and a data driving model is insufficient in safety are effectively solved, and the accuracy, the response speed, the safety and the reliability of power quality optimization control of the power distribution network are remarkably improved.

Description

Technical Field

[0001] This invention relates to a power quality optimization control method and control system for power distribution networks, belonging to the field of power quality optimization control technology. Background Technology

[0002] Currently, power quality control in distribution networks mainly relies on two methods: traditional pure mechanistic modeling or pure data-driven approaches, both of which have significant limitations. Pure mechanistic modeling builds models based on physical formulas such as power flow calculations and harmonic propagation laws in distribution networks. While it possesses strong physical interpretability, it depends on precise system parameters and fixed scenario assumptions. When faced with dynamic conditions such as fluctuations in renewable energy output and sudden load changes, the model has poor adaptability and struggles to adjust control strategies in real time. Pure data-driven models rely on large amounts of data to train algorithms such as neural networks. Although they can handle complex operating conditions, they lack physical mechanistic constraints and are prone to decision-making biases in data-scarce or extreme scenarios, making it difficult to ensure the safe operation of the distribution network. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art and provide a power quality optimization control method and control system for distribution networks. Based on the power quality optimization control model of distribution networks, the optimization control scheme effectively solves the industry problems of poor adaptability of traditional methods and insufficient security of data-driven models, and significantly improves the accuracy, response speed and safety reliability of power quality optimization control of distribution networks.

[0004] To achieve the above objectives, the present invention is implemented using the following technical solution:

[0005] On one hand, this invention discloses a method for optimizing and controlling power quality in a power distribution network, comprising the following steps:

[0006] Acquire standard multi-source real-time data during the operation of the power distribution network;

[0007] Feature extraction is performed on the standard multi-source real-time data to obtain multi-source real-time features;

[0008] The multi-source real-time features are input into the trained power quality optimization control model of the distribution network for solution to obtain the optimized control scheme.

[0009] According to the optimized control scheme, each device in the distribution network is controlled to achieve optimized power quality control of the distribution network;

[0010] The power quality optimization control model for the distribution network constructs a multi-objective optimization function with voltage qualification rate, harmonic suppression and three-phase balance as optimization objectives, and uses voltage amplitude, total harmonic distortion rate and current imbalance as constraints.

[0011] Furthermore, acquiring standard multi-source real-time data during the operation of the distribution network includes the following steps:

[0012] Acquire raw, multi-source, real-time data during the operation of the power distribution network;

[0013] After normalizing the original multi-source real-time data, standard multi-source real-time data is obtained.

[0014] The standard multi-source real-time data includes real-time power grid measurement data, real-time equipment status data, and real-time environmental load data.

[0015] The real-time data of the power grid measurement includes real-time data of node voltage, real-time data of branch current, and real-time data of harmonic components.

[0016] The real-time equipment status data includes real-time data on the rated capacity of the SVG and real-time data on the filter impedance.

[0017] The real-time environmental load data includes real-time three-phase load rate data and real-time flexible load response delay data.

[0018] Furthermore, obtaining multi-source real-time features includes the following steps:

[0019] For the real-time power grid measurement data, time-frequency analysis is performed using wavelet transform to extract the real-time features of the power grid measurement.

[0020] For the real-time equipment status data and real-time environmental load data, weights are allocated using an attention mechanism to extract real-time equipment status features and real-time environmental load features.

[0021] Based on the real-time characteristics of power grid measurements, equipment status, and environmental load, electrical correlation features between nodes are extracted using graph convolution operations to obtain multi-source real-time features.

[0022] Furthermore, the training steps for the power quality optimization control model of the distribution network are as follows:

[0023] Acquire standard multi-source historical data and corresponding equipment control command sequences during the operation of the power distribution network;

[0024] Based on the aforementioned standard multi-source historical data, feature extraction is performed to obtain multi-source historical features;

[0025] Based on the multi-source historical features and the corresponding device control command sequence, a training sample set is constructed, wherein each training sample includes a sample state, a sample action, and an actual reward value. The sample state is composed of multi-source historical features, the sample action is composed of a device control command sequence, and the actual reward value is calculated based on a multi-objective optimization function.

[0026] The training sample set is input into the pre-constructed power quality optimization control model of the distribution network for iterative training until the preset training termination condition is met, and the trained power quality optimization control model of the distribution network is output.

[0027] Furthermore, each iteration of training includes the following steps:

[0028] For any training sample, the sample state and the corresponding sample action are input into the power quality optimization control model of the distribution network with the current parameters, and the predicted reward value and state value are obtained through the forward calculation of the model.

[0029] Based on the predicted reward value and the actual reward value, the state value is corrected according to the time series difference error to obtain the corrected state value;

[0030] The model parameters of the power quality optimization control model for the distribution network are updated based on the corrected state value.

[0031] Furthermore, the expression for the state value is as follows:

[0032]

[0033] In the formula, Represents the state at time t The value of the state under the following conditions; Expressing expectations; To express summation; This represents the discount factor at time k; Represents the state at time k Next action The reward value afterward; This represents the state at time t; This represents the state at time k; This represents the action at time k;

[0034] The expression for the corrected state value is as follows:

[0035]

[0036] In the formula, Represents the state at time t The corrected state value; Indicates the correction factor; Represents the state at time t The actual reward value below; Represents the state at time t The predicted reward value.

[0037] Furthermore, the expression for the multi-objective optimization function is as follows:

[0038]

[0039] In the formula, Represents a multi-objective optimization function;

[0040] Indicates the first weighting coefficient; This indicates the annual control cost of the equipment;

[0041] This represents the second weighting coefficient; Indicates the voltage qualification rate;

[0042] Indicates the third weighting coefficient; This represents the total harmonic distortion rate of the voltage.

[0043] This represents the fourth weighting coefficient; This indicates the degree of current imbalance.

[0044] Furthermore, the expression for the constraint condition of the voltage amplitude is as follows:

[0045]

[0046] In the formula, Indicates the lower limit of voltage amplitude; Indicates the upper limit of voltage amplitude;

[0047] This represents the voltage phasor of the i-th node; This represents the voltage amplitude at the i-th node;

[0048] The expression for the constraint condition of the total harmonic distortion rate is as follows:

[0049]

[0050] In the formula, This represents the total harmonic distortion rate of the voltage. Indicates the amplitude of the first harmonic voltage; This represents the amplitude of the h-th harmonic voltage; This indicates the upper limit of the total harmonic distortion rate of the voltage.

[0051] The expression for the constraint condition of the current unbalance is as follows:

[0052]

[0053] In the formula, Indicates the degree of current imbalance; This represents the maximum effective value of the three-phase current; This represents the minimum effective value of the three-phase current. This represents the average value of the effective values ​​of the three-phase current. This represents the upper limit threshold of the current imbalance.

[0054] Furthermore, the optimized control scheme includes reactive power output adjustment commands for the static reactive power generator and switching gear commands for the filter.

[0055] According to the optimized control scheme, controlling each device in the power distribution network includes the following steps:

[0056] The reactive power output adjustment command of the static reactive power generator and the switching gear command of the filter are converted into control signals of the static reactive power generator and operation signals of the switching equipment, and then sent to the equipment actuator through power line carrier communication or wireless communication.

[0057] On the other hand, this invention discloses a power quality optimization control system for distribution networks, applicable to the aforementioned power quality optimization control method for distribution networks, characterized by comprising:

[0058] The data acquisition module is used to acquire standard multi-source real-time data during the operation of the power distribution network;

[0059] The feature extraction module is used to extract features from the standard multi-source real-time data to obtain multi-source real-time features;

[0060] An optimization control module is used to input the multi-source real-time features into a trained power quality optimization control model for distribution networks to solve for an optimized control scheme.

[0061] The equipment control module is used to control each device in the power distribution network according to the optimized control scheme to achieve optimized power quality control of the power distribution network.

[0062] The power quality optimization control model for the distribution network constructs a multi-objective optimization function with voltage qualification rate, harmonic suppression and three-phase balance as optimization objectives, and uses voltage amplitude, total harmonic distortion rate and current imbalance as constraints.

[0063] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0064] The present invention relates to a distribution network power quality optimization control method and control system. First, by acquiring standard multi-source real-time data and extracting features, it achieves accurate perception and feature enhancement of the distribution network's operating status, providing a reliable data foundation for optimization decisions. Second, it solves the problem based on a trained distribution network power quality optimization control model. This model employs multi-objective collaborative optimization of voltage qualification rate, harmonic suppression, and three-phase balance, overcoming the limitations of traditional methods that often focus on a single indicator. Multiple constraints ensure that all control decisions are executed within the safe operating boundary, effectively guaranteeing the operational safety of the distribution network. This method can respond promptly to dynamic conditions such as fluctuations in renewable energy output and sudden load changes. Compared to traditional fixed-strategy control methods, it can adapt to changes in the distribution network's operating status, significantly improving the intelligence level and operational efficiency of distribution network power quality control, and providing effective technical support for the high-quality and reliable operation of distribution networks under the new power system context. Attached Figure Description

[0065] Figure 1 This is a flowchart of the power quality optimization control method for power distribution networks provided in Embodiment 1 of the present invention;

[0066] Figure 2 These are reactive power-voltage boosting effect curves under different methods provided in Embodiment 1 of the present invention. Detailed Implementation

[0067] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0068] Example 1

[0069] This embodiment 1 provides a method for optimizing and controlling the power quality of a distribution network, such as... Figure 1 As shown, it includes the following steps:

[0070] Acquire standard multi-source real-time data during the operation of the power distribution network;

[0071] Feature extraction is performed on standard multi-source real-time data to obtain multi-source real-time features;

[0072] The multi-source real-time features are input into the trained power quality optimization control model of the distribution network for solution, and the optimized control scheme is obtained.

[0073] Based on the optimized control scheme, each device in the distribution network is controlled to achieve optimized power quality control of the distribution network;

[0074] Among them, the power quality optimization control model of the distribution network constructs a multi-objective optimization function with voltage qualification rate, harmonic suppression and three-phase balance as optimization objectives, and uses voltage amplitude, total harmonic distortion rate and current imbalance as constraints.

[0075] The technical concept of this invention is as follows: First, by acquiring standard multi-source real-time data and extracting features, accurate perception and feature enhancement of the distribution network's operating status are achieved, providing a reliable data foundation for optimization decisions. Second, the solution is based on a trained distribution network power quality optimization control model. This model uses multi-objective collaborative optimization of voltage qualification rate, harmonic suppression, and three-phase balance to overcome the limitations of traditional methods that often focus on a single indicator. Multiple constraints ensure that all control decisions are executed within the safe operating boundary, effectively guaranteeing the operational safety of the distribution network. This method can respond promptly to dynamic operating conditions such as fluctuations in new energy output and sudden load changes. Compared with traditional fixed-strategy control methods, it can adapt to changes in the distribution network's operating status, significantly improving the intelligence level and operational efficiency of distribution network power quality control, and providing effective technical support for the high-quality and reliable operation of the distribution network under the new power system background.

[0076] The specific steps are as follows:

[0077] Step 1: Obtain standard multi-source real-time data during the operation of the power distribution network.

[0078] Acquiring standard multi-source real-time data during the operation of the power distribution network includes the following steps:

[0079] Acquire raw, multi-source, real-time data during the operation of the power distribution network;

[0080] After normalizing the raw multi-source real-time data, standard multi-source real-time data is obtained.

[0081] The standard multi-source real-time data includes real-time power grid measurement data, real-time equipment status data, and real-time environmental load data.

[0082] Real-time power grid measurement data includes real-time node voltage data, real-time branch current data, and real-time harmonic component data.

[0083] Real-time equipment status data includes real-time rated capacity data of SVG (Static Var Generator) and real-time filter impedance data;

[0084] Real-time environmental load data includes real-time three-phase load rate data and real-time flexible load response delay data.

[0085] Considering that the differences in the dimensions of parameters such as voltage, current, and temperature can interfere with model calculations, the collected data is standardized and all parameters are uniformly mapped to the [0,1] interval. After normalization, standard multi-source real-time data is obtained. The standard multi-source real-time data is stored in the distribution network cloud database to form a structured dataset with timestamps, providing basic data for subsequent steps.

[0086] The normalization expression is as follows:

[0087]

[0088] In the formula, Represents the i-th feature parameter Real-time data, Represents real-time data The normalized value; Representing characteristic parameters The minimum value; Representing characteristic parameters The maximum value.

[0089] Step 2: Extract features from standard multi-source real-time data to obtain multi-source real-time features.

[0090] In this embodiment, different data types need to be processed in a targeted manner during feature extraction: for measurement data such as voltage and current, the focus is on decomposing signal components related to harmonic exceedance and voltage deviation to ensure that abnormal features can be accurately identified; for meteorological-load correlation data, the focus is on extracting change features related to new energy output and load fluctuations to provide support for dynamic operating condition adaptation.

[0091] Specifically, obtaining multi-source real-time features includes the following steps:

[0092] For real-time power grid measurement data, time-frequency analysis is performed using wavelet transform to extract real-time power grid measurement features.

[0093] For real-time equipment status data and real-time environmental load data, an attention mechanism is used to assign weights and extract real-time features of equipment status and real-time environmental load.

[0094] Based on the real-time characteristics of power grid measurements, equipment status, and environmental load, the electrical correlation features between nodes are extracted using graph convolution operations to obtain multi-source real-time features.

[0095] The method for extracting electrical correlation features between nodes based on graph convolution operations includes the following steps: taking the nodes of the pilot distribution network as graph nodes and the line impedance as edges, inputting standardized voltage, current and load data, learning the electrical correlation features between nodes through graph convolution layers, and outputting multi-source real-time features as the state input of the subsequent model.

[0096] Step 3: Input the multi-source real-time features into the trained power quality optimization control model of the distribution network for solution to obtain the optimized control scheme.

[0097] 3.1 The power quality optimization control model of the distribution network constructs a multi-objective optimization function with voltage qualification rate, harmonic suppression and three-phase balance as optimization objectives, and uses voltage amplitude, total harmonic distortion rate and current imbalance as constraints.

[0098] 3.1.1 Multi-objective optimization function.

[0099] This embodiment completes the configuration of multi-objective optimization function parameters based on the operation and maintenance requirements of the pilot distribution network.

[0100] First, the equipment cost parameters are calculated, and the annual operation and maintenance costs of SVG equipment and the annual energy consumption costs of filters are statistically analyzed to determine the total annual control cost for both types of equipment. Second, the weighting coefficients are dynamically adjusted. During extreme weather periods, to ensure stable power supply, the weight of the voltage qualification target is increased, while the weights of other targets are decreased. During off-peak periods, due to low electricity demand, the weight of cost control is increased, and the weights of other targets are adjusted to balance the needs of different scenarios. Simultaneously, the monthly statistics of voltage qualification duration and total operating time are compiled to calculate the voltage qualification rate, which serves as the core input parameter for the power quality indicators in the optimization function, ensuring that the function reflects the actual operation and maintenance priorities.

[0101] To achieve accurate fault location, a characteristic voltage signal is injected using voltage-frequency control of SOP (Soft Open Point, intelligent soft switch). By superimposing a characteristic component of a specific frequency onto the mains frequency voltage, an identifiable fault feature is formed, providing a unique identifier for subsequent signal analysis.

[0102] Specifically, the expression for the multi-objective optimization function is as follows:

[0103]

[0104] In the formula, Represents a multi-objective optimization function;

[0105] Indicates the first weighting coefficient; This indicates the annual control cost of the equipment;

[0106] This represents the second weighting coefficient; Indicates the voltage qualification rate; , Indicates the duration of voltage compliance; Indicates the annual runtime;

[0107] Indicates the third weighting coefficient; This represents the total harmonic distortion rate of the voltage.

[0108] This represents the fourth weighting coefficient; This indicates the degree of current imbalance.

[0109] 3.1.2 Constraints.

[0110] In this embodiment, a mechanism constraint system that can be embedded in the control platform is built by combining the actual parameters of the pilot distribution network with industry operation standards.

[0111] Regarding voltage safety constraints, based on the 10kV distribution network operation requirements, voltage safety ranges are set for each node. When the system detects that the voltage of a node exceeds the range, a constraint warning is automatically triggered and the node is marked as pending adjustment.

[0112] For harmonic control, referencing the harmonic mitigation standards for public power grids, an upper limit for the total harmonic distortion rate of voltage is set. By monitoring the amplitude of the fundamental voltage and each harmonic voltage in real time, the distortion rate is automatically calculated. If the upper limit is exceeded, the harmonic mitigation process is initiated.

[0113] For three-phase balance constraints, a threshold for three-phase current imbalance is set, and the maximum, minimum and average values ​​of the three-phase currents are statistically analyzed in real time and the imbalance is calculated. When the value exceeds the threshold, it is marked as a key optimization target and priority is given to allocating regulation resources.

[0114] A mechanism constraint system for distribution network operation is established around the three core objectives of voltage compliance, harmonic suppression, and three-phase balance. Based on power supply standards, the voltage amplitude of the 10kV distribution network must be limited to a safe range; according to harmonic control requirements, the upper limit of the total harmonic distortion rate is controlled; and with reference to three-phase balance specifications, a current imbalance threshold is set. These constraints define physical boundaries, ensuring that optimization decisions conform to the operational rules of the distribution network and avoiding safety risks such as voltage exceeding limits and harmonic exceedances.

[0115] Specifically, the expression for the constraint condition of voltage amplitude is as follows:

[0116]

[0117] In the formula, Indicates the lower limit of voltage amplitude; Indicates the upper limit of voltage amplitude;

[0118] This represents the voltage phasor of the i-th node; This represents the voltage amplitude at the i-th node;

[0119] The expression for the constraint condition of the total harmonic distortion rate is as follows:

[0120]

[0121] In the formula, This represents the total harmonic distortion rate of the voltage. Indicates the amplitude of the first harmonic voltage; This represents the amplitude of the h-th harmonic voltage; This represents the upper limit of the total harmonic distortion rate of the voltage, which is set to 5% in this embodiment;

[0122] The expression for the constraint condition of current unbalance is as follows:

[0123]

[0124] In the formula, Indicates the degree of current imbalance; This represents the maximum effective value of the three-phase current; This represents the minimum effective value of the three-phase current. This represents the average value of the effective values ​​of the three-phase current. This represents the upper limit threshold of the current imbalance.

[0125] 3.2 Training of the power quality optimization control model for distribution networks.

[0126] The specific training steps are as follows:

[0127] 3.2.1 Obtain standard multi-source historical data and corresponding equipment control command sequences during the operation of the power distribution network.

[0128] This embodiment selects a provincial 10kV distribution network pilot area, which includes 33 distribution nodes, 5 SVG compensation devices, and 20MW of distributed photovoltaic power. To cover all operating conditions, including wet and dry seasons and winter and summer seasons, data collection was conducted continuously for one year. Specifically, the distribution network SCADA system collected the voltage and branch current of each node at a frequency of 1 minute / time, and the 3rd, 5th, and 7th harmonic components at a frequency of 5 minutes / time. The SVG local monitoring terminal was used to obtain operating parameters such as the rated capacity and real-time reactive power output of the equipment. At the same time, the meteorological monitoring station and load management platform were linked to collect data on irradiance, ambient temperature, and three-phase load factor.

[0129] 3.2.2. Based on the standard multi-source historical data, feature extraction is performed to obtain multi-source historical features.

[0130] This step is based on the same principle as step 2, and will not be repeated here.

[0131] 3.2.3 Based on multi-source historical features and corresponding equipment control command sequences, a training sample set is constructed. Each training sample includes a sample state, a sample action, and an actual reward value. The sample state is composed of multi-source historical features, the sample action is composed of equipment control command sequences, and the actual reward value is calculated based on a multi-objective optimization function.

[0132] 3.2.4 Input the training sample set into the pre-constructed power quality optimization control model of the distribution network for iterative training until the preset training termination condition is met, and output the trained power quality optimization control model of the distribution network.

[0133] Each iteration of training includes the following steps:

[0134] For any training sample, the sample state and corresponding sample action are input into the power quality optimization control model of the distribution network with current parameters, and the predicted reward value and state value are obtained through the forward calculation of the model.

[0135] Based on the predicted reward value and the actual reward value, the state value is corrected according to the time series difference error to obtain the corrected state value;

[0136] The model parameters of the power quality optimization control model for the distribution network are updated based on the corrected state values.

[0137] The expression for the state value is as follows:

[0138]

[0139] In the formula, Represents the state at time t The value of the state under the following conditions; Expressing expectations; To express summation; This represents the discount factor at time k; Represents the state at time k Next action The reward value afterward; This represents the state at time t; This represents the state at time k; This represents the action at time k;

[0140] The revised expression for the state value is as follows:

[0141]

[0142] In the formula, Represents the state at time t The corrected state value; Indicates the correction factor; Represents the state at time t The actual reward value below; Represents the state at time t The predicted reward value.

[0143] Step 4: Based on the optimized control scheme, control each device in the distribution network to achieve optimized power quality control of the distribution network.

[0144] The optimized control scheme includes reactive power output adjustment commands for the static reactive power generator and switching gear commands for the filter.

[0145] According to the optimized control scheme, controlling each device in the power distribution network includes the following steps:

[0146] The reactive power output adjustment command of the static var generator and the switching gear command of the filter are converted into control signals of the static var generator and operation signals of the switching equipment, and then sent to the equipment actuator through power line carrier communication or wireless communication.

[0147] When the power quality optimization control model of the distribution network identifies a node voltage deviation from the safe range, it outputs a control command within the mechanistic constraints and sends it to the equipment actuator. After the command is executed, the node voltage change is monitored in real time, the reward value corresponding to the actual optimization effect is calculated, compared with the model's predicted reward value, and the model parameters are adjusted according to the preset correction coefficient to optimize the subsequent decision-making logic to adapt to dynamic operating conditions.

[0148] The following table compares the performance of this invention with traditional PID (Proportional-Integral-Derivative) control and purely data-driven power quality optimization control methods for distribution networks:

[0149]

[0150] Combination Figure 2 As shown, the power quality optimization control method for distribution networks provided by this invention is significantly superior to traditional PID control and pure data-driven methods in terms of control accuracy, response speed, operational stability and overall performance, fully demonstrating the effectiveness and superiority of the "data-mechanism fusion" technical approach in solving complex power quality problems in distribution networks.

[0151] Example 2

[0152] This embodiment 2 provides a power quality optimization and control system for a distribution network, applicable to the power quality optimization and control method for a distribution network described in embodiment 1 above, including:

[0153] The data acquisition module is used to acquire standard multi-source real-time data during the operation of the power distribution network;

[0154] The feature extraction module is used to extract features from standard multi-source real-time data to obtain multi-source real-time features.

[0155] The optimization control module is used to input multi-source real-time features into the trained power quality optimization control model of the distribution network for solution, and obtain the optimization control scheme.

[0156] The equipment control module is used to control various devices in the distribution network according to the optimized control scheme in order to achieve optimized power quality control of the distribution network.

[0157] Among them, the power quality optimization control model of the distribution network constructs a multi-objective optimization function with voltage qualification rate, harmonic suppression and three-phase balance as optimization objectives, and uses voltage amplitude, total harmonic distortion rate and current imbalance as constraints.

[0158] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0159] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0160] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0161] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0162] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for optimizing and controlling power quality in a distribution network, characterized in that, Includes the following steps: Acquire standard multi-source real-time data during the operation of the power distribution network; Feature extraction is performed on the standard multi-source real-time data to obtain multi-source real-time features; The multi-source real-time features are input into the trained power quality optimization control model of the distribution network for solution to obtain the optimized control scheme. According to the optimized control scheme, each device in the distribution network is controlled to achieve optimized power quality control of the distribution network; The power quality optimization control model for the distribution network constructs a multi-objective optimization function with voltage qualification rate, harmonic suppression and three-phase balance as optimization objectives, and uses voltage amplitude, total harmonic distortion rate and current imbalance as constraints.

2. The power quality optimization and control method for distribution networks according to claim 1, characterized in that, The acquisition of standard multi-source real-time data during the operation of the power distribution network includes the following steps: Acquire raw, multi-source, real-time data during the operation of the power distribution network; After normalizing the original multi-source real-time data, standard multi-source real-time data is obtained. The standard multi-source real-time data includes real-time power grid measurement data, real-time equipment status data, and real-time environmental load data. The real-time data of the power grid measurement includes real-time data of node voltage, real-time data of branch current, and real-time data of harmonic components. The real-time equipment status data includes real-time data on the rated capacity of the SVG and real-time data on the filter impedance. The real-time environmental load data includes real-time three-phase load rate data and real-time flexible load response delay data.

3. The power quality optimization and control method for distribution networks according to claim 2, characterized in that, Obtaining multi-source real-time features includes the following steps: For the real-time power grid measurement data, time-frequency analysis is performed using wavelet transform to extract the real-time features of the power grid measurement. For the real-time equipment status data and real-time environmental load data, weights are allocated using an attention mechanism to extract real-time equipment status features and real-time environmental load features. Based on the real-time characteristics of power grid measurements, equipment status, and environmental load, electrical correlation features between nodes are extracted using graph convolution operations to obtain multi-source real-time features.

4. The power quality optimization control method for distribution networks according to claim 1, characterized in that, The training steps for the power quality optimization control model of the distribution network are as follows: Acquire standard multi-source historical data and corresponding equipment control command sequences during the operation of the power distribution network; Based on the aforementioned standard multi-source historical data, feature extraction is performed to obtain multi-source historical features; Based on the multi-source historical features and the corresponding device control command sequence, a training sample set is constructed, wherein each training sample includes a sample state, a sample action, and an actual reward value. The sample state is composed of multi-source historical features, the sample action is composed of a device control command sequence, and the actual reward value is calculated based on a multi-objective optimization function. The training sample set is input into the pre-constructed power quality optimization control model of the distribution network for iterative training until the preset training termination condition is met, and the trained power quality optimization control model of the distribution network is output.

5. The power quality optimization and control method for distribution networks according to claim 4, characterized in that, Each iteration of training includes the following steps: For any training sample, the sample state and the corresponding sample action are input into the power quality optimization control model of the distribution network with the current parameters, and the predicted reward value and state value are obtained through the forward calculation of the model. Based on the predicted reward value and the actual reward value, the state value is corrected according to the time series difference error to obtain the corrected state value; The model parameters of the power quality optimization control model for the distribution network are updated based on the corrected state value.

6. The power quality optimization control method for distribution networks according to claim 5, characterized in that, The expression for the state value is as follows: ； In the formula, Represents the state at time t The value of the state under the following conditions; Expressing expectations; To express summation; This represents the discount factor at time k; Represents the state at time k Next action The reward value afterward; This represents the state at time t; This represents the state at time k; This represents the action at time k; The expression for the corrected state value is as follows: ； In the formula, Represents the state at time t The corrected state value; Indicates the correction factor; Represents the state at time t The actual reward value below; Represents the state at time t The predicted reward value.

7. The power quality optimization control method for distribution networks according to claim 1, characterized in that, The expression for the multi-objective optimization function is as follows: ； In the formula, Represents a multi-objective optimization function; Indicates the first weighting coefficient; This indicates the annual control cost of the equipment; This represents the second weighting coefficient; Indicates the voltage qualification rate; Indicates the third weighting coefficient; This represents the total harmonic distortion rate of the voltage. This represents the fourth weighting coefficient; This indicates the degree of current imbalance.

8. The power quality optimization control method for distribution networks according to claim 1, characterized in that, The expression for the constraint condition of the voltage amplitude is as follows: ； In the formula, Indicates the lower limit of voltage amplitude; Indicates the upper limit of voltage amplitude; This represents the voltage phasor of the i-th node; This represents the voltage amplitude at the i-th node; The expression for the constraint condition of the total harmonic distortion rate is as follows: ； In the formula, This represents the total harmonic distortion rate of the voltage. Indicates the amplitude of the first harmonic voltage; This represents the amplitude of the h-th harmonic voltage. This indicates the upper limit of the total harmonic distortion rate of the voltage. The expression for the constraint condition of the current unbalance is as follows: ； In the formula, Indicates the degree of current imbalance; This represents the maximum effective value of the three-phase current; This represents the minimum effective value of the three-phase current. This represents the average value of the effective values ​​of the three-phase current. This represents the upper limit threshold of the current imbalance.

9. The power quality optimization control method for distribution networks according to claim 1, characterized in that, The optimized control scheme includes reactive power output adjustment commands for the static reactive power generator and switching gear commands for the filter. According to the optimized control scheme, controlling each device in the power distribution network includes the following steps: The reactive power output adjustment command of the static reactive power generator and the switching gear command of the filter are converted into control signals of the static reactive power generator and operation signals of the switching equipment, and then sent to the equipment actuator through power line carrier communication or wireless communication.

10. A power quality optimization and control system for a distribution network, applicable to the power quality optimization and control method for a distribution network as described in any one of claims 1-9, characterized in that, include: The data acquisition module is used to acquire standard multi-source real-time data during the operation of the power distribution network; The feature extraction module is used to extract features from the standard multi-source real-time data to obtain multi-source real-time features; An optimization control module is used to input the multi-source real-time features into a trained power quality optimization control model for distribution networks to solve for an optimized control scheme. The equipment control module is used to control each device in the power distribution network according to the optimized control scheme to achieve optimized power quality control of the power distribution network. The power quality optimization control model for the distribution network constructs a multi-objective optimization function with voltage qualification rate, harmonic suppression and three-phase balance as optimization objectives, and uses voltage amplitude, total harmonic distortion rate and current imbalance as constraints.

Images