Power distribution network control method

Abstract

The invention discloses a power distribution network control method in the field of power systems. Comprising the following steps: step 1), multi-source data fusion and state monitoring: establishing a comprehensive power distribution network data acquisition system, and widely collecting electrical parameters including real-time power generation power of a distributed power supply, power generation prediction data, real-time power of a load, a power change trend, current, voltage, impedance and the like of a power distribution network line, and meteorological data; and processing the multi-source data by using a data fusion method based on Kalman filtering so as to obtain the accurate real-time state of the power distribution network, and setting the collected generated power of the distributed power supply as PDG. Through accurate power generation prediction and cooperative control strategies, the impact of distributed energy access on the power distribution network is effectively reduced, the consumption proportion of distributed energy and classification and flexible control of loads are improved, load power utilization can be flexibly adjusted according to the operation state of the power distribution network, and the bearing capacity of the power distribution network for load changes is enhanced.

Description

Technical Field

[0001] The invention relates to the field of power systems, and in particular to a distribution network control method. Background Art

[0002] With the transformation of energy structure and the continuous growth of electricity demand, the effectiveness and advancement of control methods of distribution network, as a key link directly facing users in the power system, are of vital importance. At present, distribution network faces many challenges such as large-scale access of distributed energy, load diversity and increased uncertainty, and existing control methods have exposed a series of problems.

[0003] In the "A method, device and storage medium for collaborative stability control of low-voltage distribution networks" disclosed in Chinese patent CN119787433A, the method identifies the load by constructing a time-power characteristic model of the power source and the load, and designs a collaborative optimization access method for supercapacitors and lithium battery energy storage. However, when faced with complex and changeable distributed energy access scenarios, this method lacks in-depth analysis and precise control capabilities of different types of distributed energy generation characteristics, making it difficult to fully tap the potential of distributed energy, and lacks in the breadth and depth of multi-source information fusion, and cannot fully adapt to the dynamic changes of the distribution network.

[0004] The "distribution network power control method, system, intelligent device and readable storage medium" disclosed in Chinese patent CN119674958A mainly focuses on determining the power compensation strategy based on the distribution network attributes to achieve voltage zero-static error regulation and power coordination control. However, this method lacks an effective processing mechanism for the uncertainty of distributed power output prediction when dealing with the intermittent and volatile nature of distributed energy, which can easily lead to deviations in the power control strategy and affect the stable operation of the distribution network.

[0005] In the "A distribution network partition collaborative optimization control method and system" disclosed in Chinese patent CN119787515A, cluster analysis algorithm is used for partitioning and horizontal federated learning architecture is used to update model parameters. However, this method does not fully consider the dynamic changes of the distribution network topology structure during the partitioning process, and the security and efficiency of data transmission during the federated learning process need to be further improved. When faced with frequent changes in the distribution network structure caused by large-scale distributed energy access, it is difficult to quickly and accurately adjust the control strategy.

[0006] The Chinese patent CN119765371A discloses a method for optimizing voltage quality and network loss of a distribution network containing an energy storage system, which builds a two-layer optimization model for the energy storage system to improve the voltage quality and economic benefits of the distribution network. However, this method lacks flexibility in the dynamic matching of the energy storage system with distributed energy and loads, and it is difficult to achieve the optimal coordinated operation of the three under different working conditions.

[0007] Therefore, those skilled in the art have provided a distribution network control method to solve the problems raised in the above background art. Summary of the Invention

[0008] The object of the present invention is to provide a distribution network control method to solve the problems raised in the above background art.

[0009] To achieve the above object, the present invention provides the following technical solutions:

[0010] A distribution network control method includes the following steps:

[0011] Step 1), multi-source data fusion and state monitoring: Establish a comprehensive distribution network data acquisition system, widely collect electrical parameters such as real-time power generation power of distributed power sources, power generation prediction data, real-time power of loads, power change trends, current, voltage, and impedance of distribution network lines, as well as meteorological data. Use the data fusion method based on Kalman filtering to process the multi-source data to obtain the accurate real-time state of the distribution network. Let the collected power generation power of the distributed power source be P DG , the load power be PL, the line current be I, and the voltage be V. The fused estimated value can be obtained through the Kalman filtering algorithm , where X is the current measurement value vector [PDG, PL, I, V], is the estimated value at the previous moment, and K is the Kalman gain matrix, which is dynamically calculated according to system noise and measurement noise;

[0012] Step 2), distributed energy prediction and coordinated control: Aiming at the intermittency and volatility of distributed energy, use deep learning algorithms such as long short-term memory networks to accurately predict the power generation power of distributed power sources. Train the LSTM model through a large amount of historical power generation data and corresponding meteorological data to obtain the predicted power P DG−forecast . Combining the prediction results, establish a coordinated control strategy for distributed energy and the distribution network. When it is predicted that the power generation power of the distributed power source will fluctuate greatly, adjust the operation mode of the distribution network in advance, such as maintaining voltage stability by adjusting the tap position of the on-load tap-changer transformer. Let the transformer ratio be k, according to the change amount of the predicted power of the distributed power source ΔP DG−forecast and the change amount of the load power ΔP L , use the formula to calculate the adjusted ratio, where k0 is the initial ratio, V0

[0013] is the initial voltage, V ref is the reference voltage, and α and β are coefficients determined according to the characteristics of the distribution network;

[0014] Step 3), Load Classification and Flexible Control: According to the electricity consumption characteristics of the load and its sensitivity to voltage and frequency changes, the load is divided into rigid loads, adjustable flexible loads, and interruptible loads. For adjustable flexible loads, such as some industrial production equipment and smart home appliances, a control model based on price incentives and demand response is established. When there is a power deficit or voltage anomaly in the distribution network, price signals or direct control commands are sent to users to guide them to adjust their electricity consumption behavior. Let the power adjustment amount of the adjustable flexible load be ΔP flex , according to the response coefficient γ of the user to the price incentive and the price change amount

[0015] Δp and the power adjustment ratio δ under the direct control command, ΔP can be obtained flex =γΔpP flex−base +δP flex−base , where

[0016] P flex−base is the reference power of the adjustable flexible load. For interruptible loads, such as some non-critical commercial electricity and agricultural irrigation electricity, a reasonable interruption strategy is formulated to quickly cut off part of the load in case of emergency to ensure the safe and stable operation of the distribution network;

[0017] Step 4), Optimal Scheduling Based on Model Predictive Control: Build a dynamic model of the distribution network, considering the dynamic characteristics of distributed power sources, loads, energy storage systems, and distribution network lines. Using the model predictive control algorithm, with the goal of minimizing the operation cost of the distribution network, reducing network losses, and improving voltage stability, the scheduling strategy of the distribution network is optimized in a rolling manner. Let the objective function , where is the power generation cost of the distributed power source is the electricity consumption cost of the load, is the charge and discharge cost of the energy storage system, is the voltage deviation,

[0018] is the network loss, and are the weight coefficients. In each control period, according to the real-time state and future prediction information of the distribution network, solve this objective function to obtain the optimal output of the distributed power source, load distribution, and charge and discharge strategy of the energy storage system;

[0019] Step 5), Fault Diagnosis and Self-Healing Control: Establish a fault diagnosis model for the distribution network based on fault feature recognition. Utilize information such as the sudden change characteristics of current and voltage during a fault and harmonic content, and adopt the support vector machine classification algorithm to quickly and accurately determine the fault type and location. When a fault is detected, start the self-healing control strategy. Through means such as automatic switch switching and islanding operation control of distributed power sources, quickly isolate the fault area and restore power supply to the non-fault area. For example, assume that the current mutation feature vector of the fault line is , and judge the fault type through the trained SVM model . If , it indicates the i-th type of fault. According to the fault location and the distribution network topology structure, formulate a self-healing control plan, such as controlling the opening and closing state of the automatic switch S j to achieve fault isolation and power supply restoration, where the state of S j is jointly determined by the fault diagnosis result and the self-healing control strategy.

[0020] As a further solution of the present invention: It also includes communication and information security guarantee: Construct a reliable distribution network communication network, adopt a hybrid communication method combining wired communication and wireless communication, such as using optical fiber communication for backbone network transmission and wireless communication for the access of distributed power sources, loads, and terminal devices. To ensure information security during the communication process, use encryption algorithms such as AES to encrypt the transmitted data to prevent data from being stolen or tampered with. At the same time, establish a communication link status monitoring mechanism to real-time monitor indicators such as communication signal strength and bit error rate. When the communication link shows an abnormality, automatically switch to the backup communication link to ensure the reliable transmission of distribution network control information. Assume that the signal strength of the communication link is S and the bit error rate is BER. When S < S thresh or BER > BER

[0021] thresh , trigger the switching of the backup communication link, where S thresh and BER thresh are preset thresholds.

[0022] As a further solution of the present invention: The adaptive load regulation strategy in the step 2) is: Through real-time monitoring and analysis of load data, identify the electricity consumption patterns and change trends of different types of loads. For adjustable loads, when there is a power deficit or voltage abnormality in the distribution network, send personalized electricity consumption adjustment suggestions or directly control according to the preset rules and user preferences.

[0023] As a further solution of the present invention: When performing distributed energy clustering management in the step 3), when the light and wind speed conditions change, control the photovoltaic power station and the wind power station to adjust the power generation state to achieve complementary utilization and optimized scheduling of energy.

[0024] As a further solution of the present invention: The fault rapid diagnosis and self-healing reconstruction in step 4) further includes: quickly judging the fault type and location by analyzing the instantaneous current, voltage mutation characteristics, harmonic content and historical operation data of the equipment during the fault; automatically starting the self-healing control process, switching switches, adjusting the output of distributed power sources and the state of energy storage systems, isolating the fault area, restoring power supply to the non-fault area and intelligently reconstructing the distribution network.

[0025] As a further solution of the present invention: The dynamic security assessment and early warning mechanism in step 5) collects the operation data of the distribution network in real time, predicts the future operation state through an online simulation model, and issues an early warning signal and provides countermeasure suggestions when potential security risks are found.

[0026] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0027] 1. Through accurate power generation prediction and collaborative control strategies, the present invention effectively reduces the impact of distributed energy access on the distribution network, improves the consumption ratio of distributed energy, classifies and flexibly controls the load, can flexibly adjust the load power consumption according to the operation state of the distribution network, and enhances the bearing capacity of the distribution network to load changes.

[0028] 2. Based on the optimal scheduling of model predictive control, the present invention realizes the reduction of the operation cost of the distribution network and the reduction of network loss, efficient fault diagnosis and self-healing control, greatly shortens the fault power outage time, improves the power supply reliability, and reliable communication network and information security guarantee measures ensure the accurate transmission of distribution network control information, and enhance the stability and security of the distribution network operation. Specific embodiments

[0029] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0030] In the embodiments of the present invention, a distribution network control method includes the following steps:

[0031] Step 1), multi-source data fusion and state monitoring: Establish a comprehensive distribution network data acquisition system, widely collect electrical parameters such as real-time power generation power of distributed power sources, power generation prediction data, real-time power of loads, power change trends, current, voltage, and impedance of distribution network lines, as well as meteorological data, and use a data fusion method based on Kalman filtering to process the multi-source data to obtain the accurate real-time state of the distribution network. Let the collected power generation power of the distributed power source be P DG, the load power is \(P_{L}\), the line current is \(I\), and the voltage is \(V\). The fused estimated value can be obtained through the Kalman filtering algorithm. , where \(X\) is the current measurement value vector \([P_{DG}, P_{L}, I, V]\), is the estimated value at the previous moment, and \(K\) is the Kalman gain matrix, which is dynamically calculated based on system noise and measurement noise;

[0032] Step 2), Distributed energy prediction and coordinated control: Aiming at the intermittency and volatility of distributed energy, a deep learning algorithm such as the long short-term memory network (LSTM) is used to accurately predict the power generation of distributed power sources. The LSTM model is trained with a large amount of historical power generation data and corresponding meteorological data to obtain the predicted power \(P\). DG−forecast , combined with the prediction results, a coordinated control strategy for distributed energy and the distribution network is established. When it is predicted that the power generation of distributed power sources will fluctuate greatly, the operation mode of the distribution network is adjusted in advance, such as maintaining voltage stability by adjusting the tap position of the on-load tap-changer transformer. Let the transformer turns ratio be \(k\), according to the change in the predicted power of the distributed power source \(\Delta P\). DG−forecast and the change in load power \(\Delta P\). L , using the formula to calculate the adjusted turns ratio, where \(k_{0}\) is the initial turns ratio, \(V_{0}\).

[0033] is the initial voltage, \(V\). ref is the reference voltage, and \(\alpha\) and \(\beta\) are coefficients determined according to the characteristics of the distribution network;

[0034] Step 3), Load classification and flexible control: According to the electricity consumption characteristics of the load and its sensitivity to voltage and frequency changes, the load is divided into rigid loads, adjustable flexible loads, and interruptible loads. For adjustable flexible loads, such as some industrial production equipment and smart home appliances, a control model based on price incentives and demand response is established. When there is a power deficit or voltage anomaly in the distribution network, price signals or direct control commands are sent to users to guide them to adjust their electricity consumption behavior. Let the power adjustment amount of the adjustable flexible load be \(\Delta P\). flex , according to the response coefficient \(\gamma\) of the user to price incentives and the price change amount.

[0035] \(\Delta p\) and the power adjustment ratio \(\delta\) under the direct control command, \(\Delta P\) can be obtained. flex = \(\gamma\Delta pP\). flex−base + \(\delta P\). flex−base , where

[0036] \(P\). flex−base is the reference power of the adjustable flexible load. For interruptible loads, such as some non-critical commercial electricity and agricultural irrigation electricity, a reasonable interruption strategy is formulated to quickly cut off some loads in case of emergency to ensure the safe and stable operation of the distribution network;

[0037] Step 4), Optimization scheduling based on model predictive control: Build a dynamic model of the distribution network, consider the dynamic characteristics of distributed power sources, loads, energy storage systems, and distribution network lines, and use the model predictive control algorithm to roll-optimize the scheduling strategy of the distribution network with the goal of minimizing the operating cost of the distribution network, reducing network losses, and improving voltage stability. Let the objective function , where is the generation cost of the distributed power source is the electricity consumption cost of the load, is the charge and discharge cost of the energy storage system, is the voltage deviation,

[0038] is the network loss, and are the weight coefficients. In each control period, solve this objective function according to the real-time state and future prediction information of the distribution network to obtain the optimal output of the distributed power source, load distribution, and charge and discharge strategy of the energy storage system;

[0039] Step 5), Fault diagnosis and self-healing control: Establish a fault diagnosis model of the distribution network based on fault feature recognition, use information such as the mutation characteristics of current and voltage and harmonic content during a fault, and adopt the support vector machine classification algorithm to quickly and accurately judge the fault type and fault location. When a fault is detected, start the self-healing control strategy, and quickly isolate the fault area and restore power supply to the non-fault area through means such as automatic switch switching and island operation control of distributed power sources. For example, let the current mutation feature vector of the fault line be , and judge the fault type through the trained SVM model . If , it means the i-th type of fault. According to the fault location and the distribution network topology structure, formulate a self-healing control plan, such as controlling the opening and closing state of the automatic switch S j to achieve fault isolation and power supply restoration, where the state of S j is jointly determined by the fault diagnosis result and the self-healing control strategy.

[0040] Among them, it also includes communication and information security guarantee: Build a reliable distribution network communication network, adopt a hybrid communication method combining wired communication and wireless communication, such as using fiber optic communication for backbone network transmission and wireless communication for the access of distributed power sources, loads, and terminal devices. To ensure information security during communication, use encryption algorithms such as AES to encrypt the transmitted data to prevent data from being stolen or tampered with. At the same time, establish a communication link status monitoring mechanism to real-time monitor indicators such as communication signal strength and bit error rate. When the communication link is abnormal, automatically switch to the standby communication link to ensure the reliable transmission of distribution network control information. Let the signal strength of the communication link be S and the bit error rate be BER. When S < Sthresh or BER > BER

[0041] thresh when it triggers the switching of the standby communication link, where S thresh and BER thresh are preset thresholds.

[0042] Among them, the adaptive load regulation strategy in step 2) is: by real-time monitoring and analyzing load data, identifying the electricity consumption patterns and change trends of different types of loads, for adjustable loads, when there is a power deficit or voltage anomaly in the distribution network, according to the preset rules and user preferences, sending personalized electricity consumption adjustment suggestions or directly controlling.

[0043] Among them, when performing the cluster management of distributed energy in step 3), when the light and wind speed conditions change, controlling the photovoltaic power station and the wind power station to adjust the power generation status to achieve complementary utilization and optimal scheduling of energy.

[0044] Among them, the fault rapid diagnosis and self-healing reconstruction in step 4) further includes: quickly judging the fault type and location by analyzing the instantaneous current, voltage mutation characteristics, harmonic content and historical operation data of the equipment during the fault; automatically starting the self-healing control process, switching switches, adjusting the output of distributed power sources and the state of the energy storage system, isolating the fault area, restoring power supply to the non-fault area and performing intelligent reconstruction of the distribution network.

[0045] Among them, the dynamic security assessment and early warning mechanism in step 5) collects the operation data of the distribution network in real time, predicts the future operation status through an online simulation model, and issues an early warning signal and provides countermeasures when potential security risks are found.

[0046] In this embodiment, through accurate power generation prediction and collaborative control strategies, the impact of distributed energy access on the distribution network is effectively reduced, the consumption ratio of distributed energy is improved, the classification and flexible control of loads can flexibly adjust the electricity consumption of loads according to the operation status of the distribution network, and the bearing capacity of the distribution network for load changes is enhanced.

[0047] In this embodiment, the optimal scheduling based on model predictive control realizes the reduction of the operation cost of the distribution network and the reduction of network losses. The efficient fault diagnosis and self-healing control greatly shorten the fault power outage time, improve the power supply reliability. The reliable communication network and information security guarantee measures ensure the accurate transmission of the control information of the distribution network and improve the stability and security of the operation of the distribution network.

[0048] The above are only the preferred specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention should cover within the protection scope of the present invention any equivalent substitution or change made according to the technical solution and inventive concept of the present invention.

Claims

1. A distribution network control method, characterized in that: It includes the following steps: Step 1), Multi-source data fusion and status monitoring: Establish a comprehensive power distribution network data acquisition system, widely collect real-time power generation power of distributed power sources, power generation prediction data, real-time power of loads, power change trends, current, voltage, and impedance electrical parameters of power distribution network lines, as well as meteorological data. Use the data fusion method based on Kalman filtering to process multi-source data to obtain the accurate real-time status of the power distribution network. Let the power generation power of the collected distributed power source be P DG , the load power be PL, the line current be I, and the voltage be V. The estimated value after fusion can be obtained through the Kalman filtering algorithm , where X is the current measurement value vector [PDG, PL, I, V], is the estimated value at the previous moment, and K is the Kalman gain matrix, which is dynamically calculated according to system noise and measurement noise; Step 2), Distributed Energy Forecasting and Cooperative Control: Aiming at the intermittency and volatility of distributed energy, deep learning algorithms such as long short-term memory network are used to accurately predict the power generation of distributed power sources. The LSTM model is trained through a large amount of historical power generation data and corresponding meteorological data to obtain the predicted power P DG−forecast , combined with the prediction results, a cooperative control strategy for distributed energy and the distribution network is established. When it is predicted that the power generation of distributed power sources will fluctuate greatly, the operation mode of the distribution network is adjusted in advance, such as maintaining voltage stability by adjusting the tap position of the on-load tap-changer transformer. Let the transformer ratio be k, and according to the change in the predicted power of the distributed power source ΔP DG−forecast and the change in load power ΔP L , the adjusted ratio is calculated using the formula , where k0 is the initial ratio, V0 is the initial voltage, V ref is the reference voltage, and α and β are coefficients determined according to the characteristics of the distribution network; Step 3), Load Classification and Flexible Control: According to the electricity consumption characteristics of the load and its sensitivity to voltage and frequency changes, the load is classified into rigid load, adjustable flexible load, and interruptible load. For adjustable flexible loads, such as some industrial production equipment and smart home appliances, a control model based on price incentives and demand response is established. When there is a power deficit or voltage anomaly in the distribution network, by sending price signals or direct control instructions to users, the users are guided to adjust their electricity consumption behaviors. Let the power adjustment amount of the adjustable flexible load be ΔP flex , according to the response coefficient γ of the user to the price incentive and the price change Δp and the power regulation ratio δ under the direct control instruction, ΔP can be obtained flex =γΔpP flex−base +δP flex−base , where P flex−base is the reference power of the adjustable flexible load. For interruptible loads, such as some non-critical commercial electricity and agricultural irrigation electricity, reasonable interruption strategies are formulated to quickly cut off part of the load in case of emergency to ensure the safe and stable operation of the distribution network; Step 4), Optimal Scheduling Based on Model Predictive Control: Construct a dynamic model of the distribution network, considering the dynamic characteristics of distributed power sources, loads, energy storage systems, and distribution network lines. Use the model predictive control algorithm to roll-optimize the scheduling strategy of the distribution network with the goal of minimizing the operating cost of the distribution network, reducing network losses, and improving voltage stability. Set the objective function , where is the generation cost of distributed power sources is the electricity consumption cost of the load, is the charge and discharge cost of the energy storage system, is the voltage deviation, is the network loss, and are the weight coefficients. In each control period, according to the real-time state and future prediction information of the distribution network, the objective function is solved to obtain the optimal distributed power generation output, load distribution, and charge and discharge strategies of the energy storage system; Step 5), Fault diagnosis and self-healing control: Establish a fault diagnosis model for the distribution network based on fault feature recognition. Utilize the mutation characteristics of current and voltage and harmonic content information during a fault, and adopt the support vector machine classification algorithm to quickly and accurately determine the fault type and location. When a fault is detected, start the self-healing control strategy, and quickly isolate the fault area and restore the power supply to the non-fault area through means such as automatic switch switching and island operation control of distributed power sources.

2. The distribution network control method according to claim 1, characterized in that: It also includes: Communication and Information Security Assurance: Build a reliable communication network for the distribution network, adopting a hybrid communication method that combines wired communication and wireless communication. For example, fiber optic communication is used for backbone network transmission, and wireless communication is used for the access of distributed power sources, loads, and terminal devices. To ensure information security during communication, encryption algorithms such as AES are used to encrypt the transmitted data to prevent data from being stolen or tampered with. At the same time, a communication link status monitoring mechanism is established to monitor the communication signal strength and bit error rate indicators in real time. When an abnormality occurs in the communication link, it automatically switches to the backup communication link to ensure the reliable transmission of distribution network control information. Let the signal strength of the communication link be S and the bit error rate be BER. When S < S thresh or BER > BER thresh it triggers the switching of the backup communication link, where S thresh and BER thresh are preset thresholds.

3. A distribution network control method according to claim 1, characterized in that: The adaptive load regulation strategy in step 2) is as follows: By real-time monitoring and analyzing load data, identify the electricity consumption patterns and change trends of different types of loads. For adjustable loads, when there is a power deficit or voltage anomaly in the distribution network, send personalized electricity consumption adjustment suggestions or directly control according to pre-set rules and user preferences.

4. A distribution network control method according to claim 1, characterized in that: When performing cluster management of distributed energy in step 3), when the light and wind speed conditions change, control the photovoltaic power station and the wind power station to adjust the power generation state to achieve complementary utilization and optimized scheduling of energy.

5. A distribution network control method according to claim 1, characterized in that: The fault rapid diagnosis and self-healing reconstruction in step 4) further includes: Quickly determine the fault type and location by analyzing the mutation characteristics of current and voltage, harmonic content, and equipment historical operation data at the moment of fault; Automatically start the self-healing control process, switch switches, adjust the output of distributed power sources and the state of the energy storage system, isolate the fault area, restore the power supply to the non-fault area, and perform intelligent reconstruction of the distribution network.

6. A distribution network control method according to claim 1, characterized in that: The dynamic security assessment and early warning mechanism in step 5) collects the operation data of the distribution network in real time, predicts the future operation state through an online simulation model, and issues an early warning signal and provides countermeasure suggestions when potential security risks are found.

Images