Power grid access control system based on photovoltaic power generation
By using the photovoltaic power grid access control system, the randomness and harmonic problems of the photovoltaic power generation system are solved, the stable operation and efficient utilization of the power grid are realized, and the self-healing ability and power supply reliability of the power grid are improved.
Patent Information
- Application Number
- CN202411112327.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-14
- Publication Date
- 2024-11-22
AI Technical Summary
The output power of photovoltaic power generation systems is random, intermittent, and periodic. They do not have the ability to regulate power themselves. When connected to the grid through an inverter, they generate a large number of harmonics. The original grid connection standards have problems such as lack of grid planning, safe operation, and reliability, and cannot meet the needs of large-scale grid connection.
A power grid access control system based on photovoltaic power generation is adopted, including a monitoring module, a control module, and a protection module. It utilizes photovoltaic array monitoring units, grid parameter monitoring units, power control units, access point switching control units, overvoltage protection units, overcurrent protection units, and islanding protection units. Combined with artificial intelligence and big data technologies, improved artificial rabbit algorithms, Bi-LSTM networks, and blockchain technology, it can realize real-time monitoring, prediction, and control of photovoltaic power generation, optimize grid power flow distribution, diagnose grid faults, and perform real-time compensation.
It enables precise prediction and control of photovoltaic power generation, reduces grid voltage and frequency fluctuations, improves power supply quality and reliability, enables rapid diagnosis and repair of grid faults, optimizes grid power flow distribution, reduces operation and maintenance costs, and enhances the grid's self-healing and anti-interference capabilities.
Smart Images

Figure CN119010182A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of electric power technology, and in particular to an electric power grid access control system based on photovoltaic power generation. Background Art
[0002] With the growth of global energy consumption and the emphasis on environmental protection, the limitations of traditional fossil energy have become increasingly prominent. As a rich renewable energy, solar energy has the advantages of being safe, clean, widely distributed, and inexhaustible. Its development and utilization have received extensive attention. Among them, photovoltaic power generation technology has gradually become an important form of solar energy utilization.
[0003] However, photovoltaic power generation has some characteristics that bring challenges to the access to the power grid. For example, the output power of photovoltaic power stations is greatly affected by weather factors (such as light, temperature, humidity, etc.), and is random, intermittent and periodic, resulting in strong voltage fluctuations; it does not have the ability to regulate power itself, and needs to consider reactive power balance issues; and a large number of harmonics will be generated when connected to the power grid through an inverter.
[0004] In addition, in terms of technical standards, the original grid-connected standards lack important contents such as grid planning, safe operation and reliability, and some standards have exceeded their service life, which cannot meet the needs of large-scale grid connection of photovoltaic power stations. This has prompted relevant departments to formulate new technical regulations to standardize the technical requirements for photovoltaic power stations to access the grid, including power quality, operation control, grid safety, etc., to ensure the stable operation of the grid and the effective promotion and application of photovoltaic power generation. At the same time, it is also necessary to establish corresponding testing standards and management systems to facilitate photovoltaic system integrators to carry out standardized design and equipment selection, as well as the management of grid-connected photovoltaic power stations and the setting and transformation of relay protection systems by grid departments.
[0005] In order to solve the above problems, the present invention proposes a power grid access control system based on photovoltaic power generation. Summary of the invention
[0006] The purpose of the present invention is to propose a power grid access control system based on photovoltaic power generation to solve the problems raised in the background technology:
[0007] The output power of photovoltaic power stations is random, intermittent and periodic; they do not have the ability to regulate power themselves; a large number of harmonics are generated when connected to the grid through an inverter; the original grid-connected standards lack important content such as grid planning, safe operation and reliability, and some standards have exceeded their validity period, which cannot meet the needs of large-scale grid connection of photovoltaic power stations.
[0008] In order to achieve the above object, the present invention adopts the following technical solutions:
[0009] The power grid access control system based on photovoltaic power generation includes: a monitoring module, a control module and a protection module;
[0010] The monitoring module comprises:
[0011] Photovoltaic array monitoring unit: used to monitor the relevant parameters of the photovoltaic array in real time and obtain the power generation performance and status of the photovoltaic array;
[0012] Grid parameter monitoring unit: used to monitor grid-related parameters;
[0013] The control module comprises:
[0014] Power control unit: used to control the photovoltaic inverter according to the needs of the power grid and the output capacity of photovoltaic power generation;
[0015] Access point switching control unit: used to automatically optimize the access point of photovoltaic power generation based on the improved artificial rabbit algorithm in a power grid environment with multiple access points according to the operation status and needs of the power grid, and optimize the power grid flow distribution;
[0016] The protection module comprises:
[0017] Overvoltage protection unit: used to cut off the connection between photovoltaic power generation and the grid when the grid voltage exceeds the set upper limit;
[0018] Overcurrent protection unit: used to monitor the photovoltaic power generation output current and the grid input current, and take protective measures when the current exceeds the safety value;
[0019] Island protection unit: used to detect whether the power grid is in an island operation state, and immediately stop the operation of the photovoltaic power generation system when the power grid is detected to be in an island operation state.
[0020] Preferably, the power control unit establishes a photovoltaic power prediction model based on artificial intelligence and big data technology, and realizes short-term, medium-term and long-term prediction of photovoltaic power through in-depth analysis of historical meteorological data, photovoltaic panel performance data and grid load data;
[0021] The photovoltaic power generation prediction model is built based on the optimized Bi-LSTM network; the optimized Bi-LSTM network transforms the unidirectional optimized LSTM network into a bidirectional forward layer and reverse layer; the forward layer forwardly calculates once from the initial moment to the τ moment, obtains and maintains the output of the forward hidden layer at each moment; the reverse layer reversely calculates once from the initial moment to the τ moment, obtains and maintains the output of the reverse hidden layer at each moment, and finally combines the output results of the forward layer and the reverse layer at the corresponding moment to obtain the final output. Each neuron of the optimized LSTM network is a memory cell with three gates, namely, the forget gate f τ , input gate iτ and output gate o τ , as follows:
[0022] f τ =σ(w f ×[h τ-1 ,y τ ]+b f )
[0023] i τ =σ(w i ×[h τ-1 ,y τ ]+b i )
[0024] o τ =σ(w o ×[h τ-1 ,y τ ]+b o )
[0025] Among them, h τ-1 is the output of the previous moment; y τ is the input data; σ is the Sigmoid function; w f 、w i and w o is the model weight parameter; b f 、b i and b o is the model bias parameter;
[0026] Through the forget gate and the previous cell state c t-1 Multiplication determines the information that needs to be discarded in the memory cell at the previous moment, through the input gate i τ and the alternative content that needs to be updated C t Multiply to get the newly added information, and get the current cell state c by discarding the information and the root information t :
[0027] C τ =tanh(w C ×[h τ-1 ,y τ ]+b C )
[0028] c τ =f τ ×c τ-1 +i τ ×C τ
[0029] Among them, w C is the model weight parameter; b C is the model bias parameter;
[0030] Finally, the cell state value c is converted to t Perform linear mapping and output gate o τ Multiply to get the current output h t :
[0031] h τ =o τ ×tanh(c τ )
[0032] The model parameters are optimized based on the improved artificial rabbit algorithm; the improved artificial rabbit algorithm is as follows:
[0033] Search phase:
[0034] The search individual randomly selects another search individual in the population to update its position and add disturbance:
[0035] X i,t+1 =ω t ·x i,t +R·(x i,t -x j,t )+round(0.5·(0.05+r1))·z
[0036]
[0037] R=L·α
[0038]
[0039] g(l)=ranfperm(D)
[0040] Among them, X i,t+1 is the candidate position of the i-th rabbit at the t+1th iteration; ω t is the weight factor at the tth iteration; ω max and ω min are the maximum and minimum values of the weight factor respectively; x i,t is the current position of the i-th rabbit at the t-th iteration; x j,t is the current position of the jth rabbit at the tth iteration; R is the running operator; L is the running length; α(d) is the mapping function; g(l) is the random function; ranfperm(D) represents a random integer between 1 and D; z is a random number that obeys the standard normal distribution of N(0,1); r1, r2 and r3 are random numbers between [0,1]; e is a mathematical constant; T is the maximum number of iterations;
[0041] Development phase:
[0042] In each iteration, individuals in the population generate b holes around themselves along the dimension of the search space to reduce the probability of being preyed upon. The location of the kth hiding hole of the i-th rabbit is as follows:
[0043]
[0044] Among them, hole i,k,t represents the hiding cave location randomly selected from K caves; r4 is a random number between [0,1]; p(b) is the cave mapping function;
[0045] The random hidden positions are updated as follows:
[0046] X i,t+1 =x j,t +R·(r4·hole i,v,t -x j,t )
[0047]
[0048] Among them, hole i,v,t is a cave randomly selected from K caves; p r (b) is the cave random mapping function; r5 is a random number between [0,1];
[0049] After implementing both strategies, the rabbit's position x i,t+1 Updated as follows:
[0050]
[0051] Moving from the exploration phase to the development phase:
[0052] Through the energy factor β t Simulate the transition from exploration to exploitation;
[0053]
[0054] Among them, r6 is a random number between [0,1]; when β t >1, the rabbit takes a detour to find food and enters the global exploration stage; when β t When ≤1, the rabbit randomly chooses a cave to hide in, and the local development stage is carried out;
[0055] The population density around the optimal individual is also monitored based on the population density detection mechanism. When the population density reaches the threshold, the iteration is stopped; the details are as follows:
[0056] n=find(||x b -x i || <r ρ )
[0057]
[0058] Among them, n is the optimal individual x b Number of artificial rabbits around; r ρ is the density radius; find means to find the artificial rabbit x within the density radius i number; ρ is the population density around the optimal individual; N is the total population.
[0059] Preferably, the power control unit also builds a power grid fault diagnosis model based on the optimized Bi-LSTM network combined with a decision tree, monitors and analyzes the physical quantities in the power grid in real time, diagnoses the type, location and cause of the power grid fault, and sends out corresponding control signals and alarm signals.
[0060] Preferably, when the power control unit detects a fault, the control module automatically isolates the fault area, sends a control instruction to the protection module, controls the output of the system, and restores the power grid to normal operation.
[0061] Preferably, the access point switching control unit selects the best access point and access capacity based on the system operation model and the improved artificial rabbit algorithm to achieve optimal distribution of power grid flow and minimize losses.
[0062] Preferably, the protection module also installs a converter, a compensation device and a filter on the system circuit to adjust the output voltage or current, and compensate for the grid voltage fluctuation caused by the photovoltaic power fluctuation in real time; and also detects and filters the harmonic current generated by the system in real time.
[0063] Preferably, the system also realizes information sharing and collaborative control between the system and other distributed energy sources based on blockchain technology.
[0064] Compared with the prior art, the present invention provides a power grid access control system based on photovoltaic power generation, which has the following beneficial effects:
[0065] The present invention uses a photovoltaic power generation power prediction model to monitor and accurately predict the output power of photovoltaic power generation in real time and control it. At the same time, it combines with devices such as converters to effectively reduce the impact of photovoltaic power fluctuations on grid voltage and frequency, so that the grid voltage and frequency are kept within a specified range, and the power supply quality and reliability of the grid are improved. The grid fault diagnosis model is also used to monitor and analyze the operating status of the photovoltaic power generation system and the grid in real time, timely discover and diagnose faults, achieve rapid fault location and repair, reduce power outage time and scope, and reduce the operation and maintenance cost of the grid. When the grid fails or is disturbed by external interference, the system can quickly isolate the photovoltaic power generation system to prevent the spread of the fault. At the same time, after the grid returns to normal, the photovoltaic power generation system is quickly reconnected to the grid to improve the self-healing ability and anti-interference ability of the grid. The access point switching control unit is also used to intelligently adjust the access point and access mode of photovoltaic power generation to achieve the optimized distribution of grid currents, reduce line losses, and improve the utilization rate of photovoltaic power generation. BRIEF DESCRIPTION OF THE DRAWINGS
[0066] Figure 1 This is a system block diagram mentioned in Embodiment 1 of the present invention;
[0067] Figure 2 This is a schematic diagram of the weight change mentioned in Example 1 of the present invention.
[0068] Meaning of the symbols in the figure:
[0069] 1. Monitoring module; 11. Photovoltaic array monitoring unit; 12. Grid parameter monitoring unit; 2. Control module; 21. Power control unit; 22. Access point switching control unit; 3. Protection module; 31. Overvoltage protection unit; 32. Overcurrent protection unit; 33. Island protection unit. DETAILED DESCRIPTION
[0070] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.
[0071] The present invention uses a photovoltaic power prediction model to monitor and accurately predict the output power of photovoltaic power generation in real time, and combines devices such as converters to effectively reduce the impact of photovoltaic power fluctuations on grid voltage and frequency, so that the grid voltage and frequency are kept within a specified range, and the power supply quality and reliability of the grid are improved. The power grid fault diagnosis model is also used to monitor and analyze the operating status of the photovoltaic power generation system and the grid in real time, timely discover and diagnose faults, achieve rapid fault location and repair, reduce power outage time and scope, and reduce the operation and maintenance cost of the grid. When the grid fails or is disturbed by external interference, the system can quickly isolate the photovoltaic power generation system to prevent the spread of the fault. At the same time, after the grid returns to normal, the photovoltaic power generation system is quickly reconnected to the grid to improve the self-healing ability and anti-interference ability of the grid. The access point switching control unit is also used to intelligently adjust the access point and access mode of photovoltaic power generation to achieve the optimized distribution of grid currents, reduce line losses, and improve the utilization rate of photovoltaic power generation. Specifically include the following contents.
[0072] Embodiment 1:
[0073] See also Figure 1-2 , the power grid access control system based on photovoltaic power generation of the present invention comprises: a monitoring module 1, a control module 2 and a protection module 3;
[0074] Monitoring module 1 includes:
[0075] Photovoltaic array monitoring unit 11: used to monitor the output voltage, current, power and other related parameters of the photovoltaic array in real time, and obtain the power generation performance and status of the photovoltaic array; for example, by installing sensors at key locations such as the junction box and inverter of the photovoltaic array, the data is transmitted to the control system to achieve real-time monitoring of the photovoltaic array. For example, in a 5MW photovoltaic power station, current sensors and voltage sensors are installed at the junction box of each row of photovoltaic panels, and data is collected every 10 minutes to promptly detect possible faults or performance degradation problems in the array.
[0076] Grid parameter monitoring unit 12: used to monitor the voltage, frequency, phase and other parameters of the grid to ensure that the grid is in normal operation and to prepare for the access of photovoltaic power generation; for example, high-precision voltage transformers and current transformers, as well as frequency monitoring devices are installed at the incoming line end of the substation to obtain the operating parameters of the grid in real time. When the grid voltage fluctuates by more than ±5% or the frequency fluctuates by more than ±0.2Hz, the system sends out an early warning signal.
[0077] The control module 2 includes:
[0078] Power control unit 21: used to control the photovoltaic inverter according to the demand of the power grid and the output capacity of photovoltaic power generation to achieve the regulation of active power and reactive power; based on artificial intelligence and big data technology, a photovoltaic power generation power prediction model is established, and through in-depth analysis of historical meteorological data, photovoltaic panel performance data and power grid load data, short-term (such as hourly level), medium-term (such as daily level) and long-term (such as weekly and monthly level) prediction of photovoltaic power generation is achieved;
[0079] The photovoltaic power generation prediction model is built based on the optimized Bi-LSTM network; the optimized Bi-LSTM network transforms the unidirectional optimized LSTM network into a bidirectional forward layer and reverse layer; the forward layer is forward calculated from the initial moment to the moment τ, and the output of the forward hidden layer at each moment is obtained and maintained; the reverse layer is reversely calculated from the initial moment to the moment τ, and the output of the reverse hidden layer at each moment is obtained and maintained. Finally, the output results of the forward layer and the reverse layer at the corresponding moment are combined to obtain the final output. Each neuron in the optimized LSTM network is a memory cell with three gates, namely the forget gate f τ , input gate i τ and output gate o τ , as follows:
[0080] f τ =σ(w f ×[h τ-1 ,y τ ]+b f )
[0081] i τ =σ(w i ×[h τ-1 ,y τ ]+b i )
[0082] o τ =σ(w o ×[h τ-1 ,y τ ]+b o )
[0083] Among them, h τ-1 is the output of the previous moment; y τ is the input data; σ is the Sigmoid function; w f 、w i and w o is the model weight parameter; b f 、b i and b o is the model bias parameter;
[0084] Through the forget gate and the previous cell state c t-1Multiplication determines the information that needs to be discarded in the memory cell at the previous moment, through the input gate i τ and the alternative content that needs to be updated C t Multiply to get the newly added information, and get the current cell state c by discarding the information and the root information t :
[0085] C τ =tanh(w C ×[h τ-1 ,y τ ]+b C )
[0086] c τ =f τ ×c τ-1 +i τ ×C τ
[0087] Among them, w C is the model weight parameter; b C is the model bias parameter;
[0088] Finally, the cell state value c is converted to t Perform linear mapping and output gate o τ Multiply to get the current output h t :
[0089] h τ =o τ ×tanh(c τ )
[0090] The model parameters are optimized based on the improved artificial rabbit algorithm; the improved artificial rabbit algorithm is as follows:
[0091] Search phase:
[0092] The search individual randomly selects another search individual in the population to update its position and add disturbance:
[0093] X i,t+1 =ω t ·x i,t +R·(x i,t -x j,t )+round(0.5·(0.05+r1))·z
[0094]
[0095] R=L·α
[0096]
[0097] g(l)=ranfperm(D)
[0098] Among them, X i,t+1 is the candidate position of the i-th rabbit at the t+1th iteration; ω t is the weight factor at the tth iteration. The specific change process can be referred to Figure 2 , in the early stage of iteration, the inertia weight decreases slowly, which makes it have good global search ability and reach a certain convergence accuracy faster. In the later stage of iteration, its solution is easy to fall into the local optimum. At this time, the smaller inertia weight can have better local search ability to reach the optimal solution; ω max and ω min are the maximum and minimum values of the weight factor respectively; x i,t is the current position of the i-th rabbit at the t-th iteration; x j,t is the current position of the jth rabbit at the tth iteration; R is the running operator; L is the running length; α(d) is the mapping function; g(l) is the random function; ranfperm(D) represents a random integer between 1 and D; z is a random number that obeys the standard normal distribution of N(0,1); r1, r2 and r3 are random numbers between [0,1]; e is a mathematical constant; T is the maximum number of iterations;
[0099] Development phase:
[0100] In each iteration, individuals in the population generate b holes around themselves along the dimension of the search space to reduce the probability of being preyed upon. The location of the kth hiding hole of the i-th rabbit is as follows:
[0101]
[0102] Among them, hole i,k,t represents the hiding cave location randomly selected from K caves; r4 is a random number between [0,1]; p(b) is the cave mapping function;
[0103] The random hidden positions are updated as follows:
[0104] X i,t+1 =x j,t +R·(r4·hole i,v,t -x j,t )
[0105]
[0106] Among them, is a cave randomly selected from K caves; p r (b) is the cave random mapping function; r5 is a random number between [0,1];
[0107] After implementing the two strategies, the rabbit's position is updated x i,t+1as follows:
[0108]
[0109] Moving from the exploration phase to the development phase:
[0110] By energy factor β t Simulate the transition from exploration to exploitation;
[0111]
[0112] Among them, r6 is a random number between [0,1]; when β t >1, the rabbit takes a detour to find food and enters the global exploration stage; when β t When ≤1, the rabbit randomly chooses a cave to hide in, and the local development stage is carried out;
[0113] The population density around the optimal individual is also monitored based on the population density detection mechanism. When the population density reaches the threshold, the iteration is stopped; the details are as follows:
[0114] n=find(||x b -x i || <r ρ )
[0115]
[0116] Among them, n is the optimal individual x b Number of artificial rabbits around; r ρ is the density radius; find means to find the artificial rabbit x within the density radius i number; ρ is the population density around the optimal individual; N is the total population.
[0117] The photovoltaic power generation prediction model based on the Bi-LSTM network is taken as model one, the photovoltaic power generation prediction model based on the Bi-LSTM network optimized by the artificial rabbit algorithm is taken as model two, and the photovoltaic power generation prediction model of this embodiment is taken as model three. The prediction results of the three models are compared, and the specific results can be referred to Table 1:
[0118] Table 1 Prediction results of different models
[0119] Model Accuracy / % Recall / % Accuracy / % <![CDATA[F1 / %]]> Model 1 68.51 47.83 83.02 60.69 Model 2 82.89 78.26 86.75 82.29 Model 3 94.48 93.88 93.81 93.85
[0120] As can be seen from the table, the prediction effect of model 3 is better than that of model 1 and model 2. Therefore, according to the power prediction results of model 3, real-time and dynamic intelligent power control can be realized. When it is predicted that the photovoltaic power generation power is about to increase, the output of the energy storage device or other adjustable power supply in the power grid is adjusted in advance to smooth the impact of photovoltaic power generation access on the power grid; when it is predicted that the photovoltaic power generation power is decreasing, the backup power supply is started in time or the load demand is adjusted to ensure the stable operation of the power grid.
[0121] For example, when the power grid is at peak power consumption, the system controls the inverter to increase active power output; when the power grid voltage is low, the inverter is controlled to output a certain amount of reactive power to support the stability of the power grid voltage. For example, in a microgrid system in an industrial park, when the power load in the park suddenly increases, the access control system quickly increases the active power output of the nearby 2MW photovoltaic power station from 1.5MW to 2MW, meeting the power demand of the park.
[0122] The power control unit 21 also builds a power grid fault diagnosis model based on the optimized Bi-LSTM network combined with the decision tree, and conducts real-time monitoring and analysis of electrical quantities such as voltage, current, power, and physical quantities such as temperature, vibration, and sound of the equipment in the power grid, so as to quickly and accurately diagnose the type, location, and cause of the power grid fault, and send out corresponding control signals and alarm signals.
[0123] For example, by installing a large number of intelligent sensors and monitoring devices in the power grid, real-time data on power grid operation is collected, and these data are input into a power grid fault diagnosis model based on an optimized Bi-LSTM network combined with a decision tree, to achieve rapid diagnosis and location of power grid faults. For example, when a short-circuit fault occurs in a power grid in a certain area, the fault diagnosis system can accurately determine within 100 milliseconds that the fault type is a three-phase short-circuit fault, and the fault location is in the middle section of a certain transmission line, and promptly notify the operation and maintenance personnel to carry out emergency repairs.
[0124] When the power control unit 21 detects a fault, the control module 2 automatically isolates the fault area, sends a control instruction to the protection module 3, and quickly adjusts the output of the control system. By reconstructing the power grid topology and transferring the load, the power grid can resume normal operation in a short time, reduce the time and scope of power outages, and improve the power supply reliability of the power grid.
[0125] Access point switching control unit 22: used to automatically find the best way to switch the access point of photovoltaic power generation in a power grid environment with multiple access points based on the improved artificial rabbit algorithm according to the operating conditions and needs of the power grid, so as to optimize the power grid flow distribution; for example, in an urban distribution network, when the load rate of a substation in a certain area is high, the access control system switches the access point of a nearby photovoltaic power station from the substation to an adjacent substation with a lower load rate, thereby realizing the optimal configuration of power grid resources.
[0126] Access point selection and optimization system based on improved artificial rabbit algorithm. Taking into account the topology of the power grid, line capacity, load distribution, voltage stability and other factors, the improved artificial rabbit algorithm automatically determines the best access point and access capacity of the photovoltaic power generation system to achieve the optimal distribution of power grid flow and minimize losses.
[0127] For example, for a city distribution network, by establishing a mathematical model, taking the line impedance of the power grid, the location and size of the load nodes, the available access points, etc. as input parameters, an intelligent optimization algorithm is used to search for the photovoltaic power generation access point and access capacity combination that minimizes the grid loss and optimizes the voltage stability.
[0128] Realize the dynamic configuration function of access points. According to the real-time changes in the operation status of the power grid (such as dynamic changes in load, power grid failures, equipment maintenance, etc.) and the fluctuation of photovoltaic power generation, adjust the access points and access methods of the photovoltaic power generation system in real time to improve the flexibility and adaptability of the power grid.
[0129] Protection module 3 includes:
[0130] Overvoltage protection unit 31: used to cut off the connection between photovoltaic power generation and the grid when the grid voltage exceeds the set upper limit; for example: in a rural power grid, due to lightning and other reasons, the grid voltage instantly rises to 1.2 times the rated voltage, and the overvoltage protection device in the access control system immediately operates to disconnect the 100kW distributed photovoltaic power source from the grid, avoiding damage to equipment such as photovoltaic inverters due to overvoltage.
[0131] Overcurrent protection unit 32: used to monitor the photovoltaic power generation output current and the grid input current, and take protective measures when the discharge current exceeds the safety value; for example: during the operation of a rooftop photovoltaic power station of a factory, due to an inverter failure, the output current suddenly increased to 1.5 times the rated current, and the overcurrent protection device acted quickly to isolate the inverter from the grid, preventing the expansion of the fault range.
[0132] Island protection unit 33: used to detect whether the power grid is in an island operation state, and immediately stop the operation of the photovoltaic power generation system when the power grid is detected to be in an island operation state. For example: in an independent photovoltaic power supply system in a remote area, when the power grid is out of power due to a line fault, the island protection device detects the loss of power grid pressure and stops the operation of the photovoltaic inverter within 0.2 seconds, avoiding the safety threat caused by the island operation to user equipment and maintenance personnel.
[0133] The system also uses new power electronic technologies, such as modular multilevel converters (MMC), static synchronous compensators (STATCOM), active power filters (APF), etc., to provide real-time compensation and control for power quality problems (such as harmonics, voltage fluctuations and flicker, three-phase imbalance, etc.) caused by photovoltaic power generation access, ensuring that the power supply quality of the power grid meets national standards and user requirements.
[0134] For example, MMC or STATCOM devices are installed on the DC or AC side of the photovoltaic inverter to quickly adjust the output voltage or current to compensate for grid voltage fluctuations caused by photovoltaic power fluctuations in real time; APF devices are installed at the grid connection point to detect and filter the harmonic currents generated by the photovoltaic power generation system in real time.
[0135] It is also possible to develop photovoltaic inverter control strategies based on virtual synchronous generator (VSG) technology, so that the photovoltaic power generation system has inertia and damping characteristics similar to those of traditional synchronous generators when connected to the grid, enhance the photovoltaic power generation system's ability to support grid frequency and voltage, and improve the stability and anti-interference ability of the grid.
[0136] The system also builds a distributed energy management platform based on blockchain technology to achieve information sharing, transaction settlement and collaborative control between photovoltaic power generation systems and other distributed energy sources (such as wind power generation, micro gas turbines, energy storage systems, etc.). Through the decentralized, tamper-proof, and smart contract characteristics of blockchain, it ensures the security, transparency and efficiency of energy transactions.
[0137] For example, in a microgrid in an industrial park, each distributed energy device (including photovoltaic power generation, wind power generation, energy storage devices, etc.) acts as a node of the blockchain network, automatically executing energy transactions and collaborative control strategies through smart contracts. When the photovoltaic power generation is in excess, the excess electricity is sold to other users or energy storage devices in need through smart contracts; when the photovoltaic power generation is insufficient, it automatically obtains electricity from energy storage devices or other distributed power sources to achieve optimal configuration and efficient use of energy in the park.
[0138] Develop a multi-energy complementary and coordinated control strategy. Comprehensively consider the characteristics and advantages of different energy sources such as photovoltaic power generation, wind power generation, and energy storage systems, and optimize control algorithms to achieve seamless switching and coordinated operation among multiple energy sources, thereby improving the reliability and stability of energy supply.
[0139] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.
Claims
1. A power grid access control system based on photovoltaic power generation, characterized in that: include: Monitoring module (1), control module (2) and protection module (3); The monitoring module (1) comprises: Photovoltaic array monitoring unit (11): used to monitor relevant parameters of the photovoltaic array in real time and obtain the power generation performance and status of the photovoltaic array; A power grid parameter monitoring unit (12): used for monitoring power grid related parameters; The control module (2) comprises: A power control unit (21): used to control the photovoltaic inverter according to the demand of the power grid and the output capacity of photovoltaic power generation; An access point switching control unit (22): used to automatically search for and switch the access point of photovoltaic power generation based on an improved artificial rabbit algorithm in a power grid environment with multiple access points according to the operation status and demand of the power grid, thereby optimizing the power grid flow distribution; The protection module (3) comprises: Overvoltage protection unit (31): used to cut off the connection between photovoltaic power generation and the grid when the grid voltage exceeds a set upper limit value; Overcurrent protection unit (32): used to monitor the photovoltaic power generation output current and the grid input current, and take protective measures when the current exceeds the safety value; Island protection unit (33): used to detect whether the power grid is in an island operation state, and immediately stop the operation of the photovoltaic power generation system when the power grid is detected to be in an island operation state.
2. The photovoltaic power generation-based power grid access control system according to claim 1, characterized in that: The power control unit (21) establishes a photovoltaic power prediction model based on artificial intelligence and big data technology, and realizes short-term, medium-term and long-term prediction of photovoltaic power through in-depth analysis of historical meteorological data, photovoltaic panel performance data and power grid load data; The photovoltaic power generation prediction model is built based on the optimized Bi-LSTM network; the optimized Bi-LSTM network transforms the unidirectional optimized LSTM network into a bidirectional forward layer and reverse layer; the forward layer forwardly calculates once from the initial moment to the τ moment, obtains and maintains the output of the forward hidden layer at each moment; the reverse layer reversely calculates once from the initial moment to the τ moment, obtains and maintains the output of the reverse hidden layer at each moment, and finally combines the output results of the forward layer and the reverse layer at the corresponding moment to obtain the final output. Each neuron of the optimized LSTM network is a memory cell with three gates, namely, the forget gate f τ , input gate i τ and output gate o τ , as follows: f τ =σ(w f ×[h τ-1 ,y τ ]+b f ) i τ =σ(w i ×[h τ-1 ,y τ ]+b i ) the τ =σ(w o ×[h τ-1 ,y τ ]+b o ) Among them, h τ-1 is the output of the previous moment; y τ is the input data; σ is the Sigmoid function; w f 、w i and w o is the model weight parameter; b f 、b i and b o is the model bias parameter; Through the forget gate and the previous cell state c t-1 Multiplication determines the information that needs to be discarded in the memory cell at the previous moment, through the input gate i τ and the alternative content that needs to be updated C t Multiply to get the newly added information, and get the current cell state c by discarding the information and the root information t : C τ =tanh(w C ×[h τ-1 ,y τ ]+b C ) c τ =f τ ×c τ-1 +i τ ×C τ Among them, w C is the model weight parameter; b C is the model bias parameter; Finally, the cell state value c is converted to t Perform linear mapping and output gate o τ Multiply to get the current output h t : h τ =o τ ×tanh(c τ ) The model parameters are optimized based on the improved artificial rabbit algorithm; the improved artificial rabbit algorithm is as follows: Search phase: The search individual randomly selects another search individual in the population to update its position and add disturbance: X i,t+1 =ω t ·x i,t +R·(x i,t -x j,t )+round(0.5·(0.05+r1))·z R=L·α g(l)=ranfperm(D) Among them, X i,t+1 is the candidate position of the i-th rabbit at the t+1th iteration; ω t is the weight factor at the tth iteration; ω max and ω min are the maximum and minimum values of the weight factor respectively; x i,t is the current position of the i-th rabbit at the t-th iteration; x j,t is the current position of the jth rabbit at the tth iteration; R is the running operator; L is the running length; α(d) is the mapping function; g(l) is the random function; ranfperm(D) represents a random integer between 1 and D; z is a random number that obeys the standard normal distribution of N(0,1); r1, r2 and r3 are random numbers between [0,1]; e is a mathematical constant; T is the maximum number of iterations; Development phase: In each iteration, individuals in the population generate b holes around themselves along the dimension of the search space to reduce the probability of being preyed upon. The location of the kth hiding hole of the i-th rabbit is as follows: Among them, hole i,k,t represents the hiding cave location randomly selected from K caves; r4 is a random number between [0,1]; p(b) is the cave mapping function; The random hidden positions are updated as follows: X i,t+1 =x j,t +R·(r4·hole i,v,t -x j,t ) Among them, hole i,v,t is a cave randomly selected from K caves; p r (b) is the cave random mapping function; r5 is a random number between [0,1]; After implementing the two strategies, the rabbit's position is updated x i,t+1 as follows: Moving from the exploration phase to the development phase: By energy factor β t Simulate the transition from exploration to exploitation; Among them, r6 is a random number between [0,1]; when β t >1, the rabbit takes a detour to find food and enters the global exploration stage; when β t When ≤1, the rabbit randomly chooses a cave to hide in, and the local development stage is carried out; The population density around the optimal individual is also monitored based on the population density detection mechanism. When the population density reaches the threshold, the iteration is stopped; the details are as follows: n=find(||x b -x i ||<r ρ ) Among them, n is the optimal individual x b Number of artificial rabbits around; r ρ is the density radius; find means to find the artificial rabbit x within the density radius i number; ρ is the population density around the optimal individual; N is the total population.
3. The photovoltaic power generation-based power grid access control system according to claim 2, characterized in that: The power control unit (21) also builds a power grid fault diagnosis model based on an optimized Bi-LSTM network combined with a decision tree, performs real-time monitoring and analysis on physical quantities in the power grid, diagnoses the type, location and cause of the power grid fault, and issues corresponding control signals and alarm signals.
4. The photovoltaic power generation-based power grid access control system according to claim 3, characterized in that: When the power control unit (21) detects a fault, the control module (2) automatically isolates the fault area, sends a control instruction to the protection module (3), controls the output of the system, and restores the normal operation of the power grid.
5. The photovoltaic power generation-based power grid access control system according to claim 4, characterized in that: The access point switching control unit (22) selects the best access point and access capacity based on the system operation model and the improved artificial rabbit algorithm to achieve optimal distribution of power grid current and minimize losses.
6. The photovoltaic power generation-based power grid access control system according to claim 1, characterized in that: The protection module (3) is also equipped with a converter, a compensation device and a filter on the system circuit, which are used to adjust the output voltage or current, and to compensate in real time for the voltage fluctuation of the power grid caused by the power fluctuation of the photovoltaic generating end; and to detect and filter the harmonic current generated by the system in real time.
7. The photovoltaic power generation-based power grid access control system according to claim 1, characterized in that: The system also uses blockchain technology to achieve information sharing and collaborative control between the system and other distributed energy sources.
Citation Information
Cited By
Intelligent filament voltage stabilization control system and method thereof
CN119421281A
Intelligent filament voltage stabilization control system and method thereof
CN119421281B
Distributed photovoltaic fusion control method and device
CN120073874A
Multi-energy complementary micro-grid coordination control method
CN120357566A
Regional distributed photovoltaic fault diagnosis method accessed to power distribution network
CN120825123A
