Photovoltaic grid-connected voltage control system and method based on edge calculation

Abstract

The invention discloses a photovoltaic grid-connected voltage control system and method based on edge calculation, and belongs to the technical field of power grid control. The control system comprises a solar cell array, a photovoltaic inversion module and a protocol conversion module. The photovoltaic inversion module comprises a modulation module, a voltage sensor and a communication module; the protocol conversion module comprises a protocol conversion unit and a voltage regulation system, and the protocol conversion unit receives the alternating current voltage data sent by the communication module on one hand and obtains target grid-connected voltage data from a power grid on the other hand; according to the voltage regulation system, on one hand, the protocol conversion unit obtains alternating current voltage data output by the modulation module and target grid-connected voltage data of a power grid, on the other hand, voltage regulation parameters are generated based on model prediction control, and on the other hand, the voltage regulation parameters are sent to the communication module. According to the invention, a voltage regulation system based on edge calculation is additionally arranged on an existing protocol converter, so that real-time voltage monitoring and regulation are realized.

Description

Technical Field

[0001] The present invention relates to the field of power system control, and in particular to a photovoltaic grid-connected voltage control system and method based on edge computing. Background Art

[0002] The number of photovoltaic grid-connected users is increasing year by year. After some photovoltaic areas have newly installed photovoltaic users, the voltage has risen to more than 260V due to the quality of photovoltaic inverters, which seriously affects the electricity consumption of residents and even poses a hidden danger to the safe operation of the power grid. There are many types of photovoltaic inverters and different communication protocols. The inability to unify photovoltaic inverters has caused huge communication costs and equipment update costs in the process of grid connection.

[0003] In addition, the power grid has a large user base and a large number of photovoltaic installations each year. Due to the widespread phenomenon of photovoltaic voltage increase, the line loss in the substation area increases after the photovoltaic voltage increase, resulting in increased substation management costs and significantly reduced economic benefits. More seriously, the voltage rises too high, posing a hidden danger to the safe operation of the power grid. Summary of the invention

[0004] In view of the problem that the grid-connected voltages of different photovoltaic inverters in the prior art cannot be uniformly regulated, the present invention provides a photovoltaic grid-connected voltage control system and method based on edge computing. The present invention is oriented to the protocol terminal, and by independently controlling and calculating the output voltage of each modulation module, the edge computing method and the master station side development App remote debugging function are realized, thereby controlling the photovoltaic voltage in real time, preventing the grid-connected voltage from being too high, and ensuring the safe operation of the power grid.

[0005] The technical means adopted by the present invention are as follows: A photovoltaic grid-connected voltage control system based on edge computing, comprising: a solar cell array, a photovoltaic inverter module and a protocol conversion module; wherein: The photovoltaic inverter module includes a modulation module, a voltage sensor and a communication module. The modulation module is used to modulate the direct current output by the solar cell array into alternating current; the voltage sensor is used to detect the alternating current voltage data output by the modulation module; the communication module is used to receive the voltage data collected by the voltage sensor on the one hand, and send the received voltage adjustment parameters to the modulation module on the other hand to control the alternating current voltage output by the modulation module; The protocol conversion module includes a protocol conversion unit and a voltage regulation system. The protocol conversion unit receives the AC voltage data sent by the communication module on the one hand, and obtains the voltage allowed by the grid when the photovoltaic grid is connected to the grid on the other hand; the voltage regulation system obtains the AC voltage data output by the modulation module and the target grid-connected voltage data of the grid from the protocol conversion unit on the one hand, generates voltage regulation parameters based on model predictive control on the one hand, and sends the voltage regulation parameters to the communication module on the other hand; Generate voltage regulation parameters based on model predictive control, including: The voltage prediction model is constructed as follows:

[0006] in, The voltage output predicted by the voltage regulation system at time t+1, is the system state variable at time t, is the control variable in the voltage regulation system, For the voltage prediction model based on recurrent neural network, are the learnable parameters of the model; The predictive control objective function is constructed as follows:

[0007] in, Predicting voltage output for voltage regulation systems, The preset target value of the modulation module output voltage, is the length of the prediction time domain, To control the length of the time domain, , is the control input change at time t, which satisfies the constraint , is the control input lower limit, For control input upper limit, , To control the upper limit of input change rate, To control the lower limit of input change rate, Weight parameters for balancing tracking error and changes in control input; Solve the predictive control objective function to obtain the predicted value of the modulation module output voltage and the predicted value of the voltage regulation parameter.

[0008] Further, solving the predictive control objective function to obtain the predicted value of the modulation module output voltage and the predicted value of the voltage regulation parameter includes: The predictive control objective function is derived, and the optimal solution is iteratively solved by the gradient descent method.

[0009] Furthermore, the voltage regulation system communicates with the relay service in real time and uploads the acquired information to the management end, which facilitates information collection and control of the terminal and the master station.

[0010] Furthermore, the voltage control system also includes a remote control terminal, which can run an APP program to achieve 7x24 system monitoring and manual adjustment.

[0011] The present invention also discloses a photovoltaic grid-connected voltage control method based on edge computing, which is implemented based on any one of the control systems described above and is characterized in that it includes the following steps: S1. Obtain the output voltage of the current photovoltaic inverter module through a voltage sensor, and send the collected output voltage information to the protocol conversion unit of the protocol conversion module through a communication module; S2, the protocol conversion unit sends the output voltage information to the voltage regulation system on the one hand, and sends the voltage information to the power grid on the other hand; S3, the voltage regulation system predicts the future output voltage based on historical data and current status, sets the target voltage and the objective function of minimizing energy loss for optimization control, and the optimization control adopts PID control; S4. The output voltage predicted by the voltage regulation system is transmitted back to the modulation module of the photovoltaic inverter module, and the modulation module regulates the output voltage of the photovoltaic inverter module according to the predicted output voltage.

[0012] Compared with the prior art, the present invention has the following advantages: 1. The present invention adds a voltage regulation system based on edge computing to the existing protocol converter to achieve real-time voltage monitoring and regulation. Edge computing uses the mainstream real-time operating system (RTOS) to improve the response speed and ensure the real-time performance of the control algorithm.

[0013] 2. The communication module of the voltage system regulation system in the present invention adopts CAT4 protocol to improve the stability of communication and prevent the occurrence of packet loss problems.

[0014] 3. The edge computing module of the present invention establishes a dynamic model of the photovoltaic inverter and its grid, sets the target voltage and the objective function of minimizing energy loss to control the most appropriate voltage. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0016] Figure 1 This is an architecture diagram of a photovoltaic grid-connected voltage control system based on edge computing in the present invention.

[0017] Figure 2 This is a flow chart of a photovoltaic grid-connected voltage control method based on edge computing in the present invention.

[0018] Figure 3 Schematic diagram of fluctuation trends of voltage V, reactive power Q, active power P and voltage setting value V in an embodiment of the present invention.

[0019] Figure 4 Schematic diagram of voltage regulation simulation results in an embodiment of the present invention. DETAILED DESCRIPTION

[0020] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described 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, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0021] like Figure 1 As shown, the present invention provides a photovoltaic grid-connected voltage control system based on edge computing, including: a solar cell array, a photovoltaic inverter module and a protocol conversion module. Among them: The photovoltaic inverter module includes a modulation module, a voltage sensor and a communication module. The modulation module is used to modulate the direct current output by the solar cell array into alternating current; the voltage sensor is used to detect the alternating current voltage data output by the modulation module; the communication module is used to receive the voltage data collected by the voltage sensor on the one hand, and send the received voltage adjustment parameters to the modulation module on the other hand to control the alternating current voltage output by the modulation module.

[0022] Specifically, the modulation module is one of the core components of the photovoltaic inverter module, which can convert DC power into AC power through modulation technology. Its structure mainly includes a switching circuit, a filtering circuit and a controller unit. During operation, the controller unit generates a PWM signal to drive the switching device, and finally outputs high-quality AC power through filtering.

[0023] The protocol conversion module includes a protocol conversion unit and a voltage regulation system. The protocol conversion unit receives the AC voltage data output by the modulation module sent by the communication module, and obtains the voltage range allowed by the power grid when the photovoltaic grid is connected. The voltage regulation system obtains the AC voltage data output by the modulation module and the target grid-connected voltage data of the power grid from the protocol conversion unit, generates voltage regulation parameters based on model predictive control, and sends the voltage regulation parameters to the communication module.

[0024] Specifically, the voltage regulation system communicates with the photovoltaic inverter module, calculates the current appropriate voltage through model predictive control and control algorithm based on the modulation module output voltage collected in real time by the voltage sensor, and feeds back to the modulation module of the photovoltaic inverter module to change the output voltage to achieve voltage control.

[0025] The protocol conversion unit is a bridge for the photovoltaic inverter module to achieve communication and interoperability. Its core function is to solve the problem of incompatible communication protocols between different devices, ensuring that data can be correctly transmitted and instructions can be accurately executed. The grid-connected voltage output by the photovoltaic manufacturer's system is collected through the inverter converter, and the allowable voltage range of the main grid is obtained at the same time, and the target value of voltage regulation is selected within the allowable voltage range. The data interaction between the photovoltaic inverter module and the voltage regulation system is realized through the protocol conversion unit, completing the closed-loop control of real-time voltage collection, analysis, feedback and execution.

[0026] Furthermore, the voltage regulation system uses filtering, correction, multi-parameter analysis, prediction optimization and other means to process the collected voltage information to ensure accurate control instructions. In this embodiment, it is preferred to use the MPC (model predictive control) control algorithm to predict the voltage change trend in the future period of time based on the collected real-time voltage information, combined with historical data and system models. For example, if the input voltage of the photovoltaic inverter module fluctuates greatly, MPC can use the predicted voltage regulation value as the goal, accurately predict the output voltage, generate control inputs, and make the output voltage meet the voltage requirements. For example, the output voltage meets the reasonable output range set in advance (220V±10%).

[0027] The present invention adopts an edge computing architecture and completes control calculations through local devices. Its core is to build a voltage regulation system based on a recurrent neural network model. The specific process is as follows.

[0028] The voltage prediction model is constructed as follows:

[0029] in, The voltage output predicted by the voltage regulation system at time t+1, is the system state variable at time t, including historical voltage, reactive power, active power of the photovoltaic inverter module, and other environmental characteristics. In this application, the main source of other environmental characteristics is randomly generated environmental noise (in the range of [-0.5, 0.5]), which is used to simulate external interference or measurement errors that may occur during system operation. The introduction of environmental noise is intended to enhance the robustness of the voltage prediction model so that it can better adapt to the uncertainty factors in the actual operating environment. ). is the control variable in the voltage regulation system, For the voltage prediction model based on recurrent neural network, are the learnable parameters of the model.

[0030] In this application, based on the recurrent neural network voltage prediction model Use the LSTM model to input the historical status data of the current system and obtain the future The predicted value of the time step, assuming that s(t) is the state of the solar array at time t, and u(t) is the control input variable at time t, together constitute the input feature [ ] and input it into the LSTM model.

[0031] The calculation formula of the forget gate is as follows.

[0032]

[0033] in, is the output of the forget gate, indicating the proportion of information that needs to be forgotten. σ is the sigmoid activation function, and the output range is [0,1]. f is the weight matrix of the forget gate. f is the bias term of the forget gate.

[0034] The input gate consists of two parts: the activation value of the input gate and the candidate value vector.

[0035] The activation values ​​of the input gate are shown below.

[0036]

[0037] The candidate value vector is shown below.

[0038]

[0039] in, is the activation value of the input gate, indicating the proportion of information that needs to be updated. W is a candidate value vector representing the new information to be written into the cell state. i and W c are the weight matrices of the input gate and candidate value vector respectively. b i and bc are the bias terms of the input gate and the candidate value vector respectively.

[0040] The calculation formula of the output gate is shown below.

[0041]

[0042] in, is the activation value of the output gate, indicating the proportion of information that needs to be output. o is the weight matrix of the output gate. o is the bias term of the output gate.

[0043] The update formula for the hidden state is shown below.

[0044]

[0045] in, is the updated hidden state. is the activation value of the cell state, which is used to control the output range of the hidden state. Represents element-wise multiplication.

[0046] After another fully connected layer, the hidden state of LSTM The mapping to the output target variable is shown below.

[0047]

[0048] in, is the learnable weight parameter of the LSTM model, which is randomly set by Gaussian initialization.

[0049] The constructed objective function is shown below.

[0050]

[0051] in, is the predicted value of the modulation module output voltage based on the LSTM model, is the preset target value of the modulation module output voltage, 220V is selected in this embodiment, To track the error, minimize the error and make the system output close to the set value; to reduce the dynamic changes of the control input and avoid overly drastic adjustments, , To control the change in input, is a weight factor used to balance the tracking error and the change of control input.

[0052] The constraints of the objective function are as follows.

[0053] ,

[0054] in, is the control input lower limit, Enter an upper limit for control. is the upper limit of the control input change rate, It is the lower limit of the control input change rate.

[0055] The prediction control objective function is solved based on the gradient descent method to obtain the predicted value of the modulation module output voltage and the predicted value of the voltage regulation parameter, thereby achieving real-time control and updating the prediction. The solution formula of the gradient descent method is as follows.

[0056]

[0057] Updating the control variables according to the gradient is shown below.

[0058]

[0059] Furthermore, the voltage regulation system and the relay service communicate in real time, and upload the acquired information to the management end, which facilitates the information collection and control of the terminal and the master station. The relay service is a bridge between the voltage regulation system and the management end, responsible for data forwarding, buffering, and protocol conversion tasks, and is mainly responsible for uploading the collected voltage data output by the modulation module to the database (master station) and background of the State Grid.

[0060] Furthermore, the voltage control system also includes a remote control terminal, which can run an APP program to achieve 7x24 system monitoring and manual adjustment. The APP is mainly responsible for displaying the collected information in real time on the UI interface, and can also adjust the voltage according to the threshold set in advance by the user. There is no need for manual verification and search of each photovoltaic component, which greatly reduces labor costs.

[0061] like Figure 2 As shown, the present invention also discloses a photovoltaic grid-connected voltage control method based on edge computing, which is implemented based on the above control system and includes the following steps.

[0062] S1. Obtain the current output voltage of the photovoltaic inverter module through a voltage sensor, and send the collected output voltage information to the protocol conversion unit of the protocol conversion module through a communication module.

[0063] S2. The protocol conversion unit sends the output voltage information to the voltage regulation system on the one hand, and sends the voltage information to the power grid on the other hand.

[0064] S3, the voltage regulation system predicts the future output voltage based on historical data and current status, sets the target voltage and the objective function of minimizing energy loss to perform optimization control. In this embodiment, the optimization control adopts PID control.

[0065] S4. The output voltage predicted by the voltage regulation system is transmitted back to the modulation module of the photovoltaic inverter module, and the modulation module regulates the output voltage of the photovoltaic inverter module according to the predicted output voltage.

[0066] like Figure 2 As shown, an application example of the present invention is given.

[0067] In this simulation experiment, we simulated the voltage fluctuations in the power system through simulation data, predicted the voltage using the LSTM model, and then calculated the adjusted voltage value through the adjustment formula, so that the voltage gradually approaches the target value.

[0068] First, the system settings of the voltage regulation system are simulated. The system parameters are set to filter inductance L=5mH, grid voltage V=220V, DC bus voltage V=400V, and sampling period , prediction time domain , control time domain 3. Weight parameters =0.01.

[0069] Based on the above parameters, the experimental input data s(t) consists of actual voltage V, reactive power Q, active power P, environmental noise and target voltage. The time simulation sequence of key variables such as voltage, reactive power and active power is as follows: Figure 3 As shown in the figure, the voltage range fluctuates between 205V and 235V, the reactive power and active power are between 40-60 kW and 50-70 kW respectively, and the ambient noise is randomly generated between [-0.5, 0.5]. These data are normalized and input into the LSTM model for learning. We constructed time series data, where each set of input contains data from the past 10 time steps, and the output is the voltage value of the current time step. When training the model, 80% of the data is used as the training set and 20% of the data is used as the test set.

[0070] After the prediction is completed, we adjust the predicted voltage using the following formula to simulate the working principle of the voltage regulation system: adjusted_voltage = y_pred + (y_tar- y_pred) * 0.5, where y_pred represents the predicted voltage output of the voltage regulation system according to the model prediction The storage variable of , y_tar is the target voltage (220V), and 0.5 is the adjustment coefficient, which means that each adjustment process will bring the voltage closer to the target value by half. This formula can be used to gradually correct the voltage deviation so that the adjusted voltage value gradually approaches the target voltage. In the simulation results, we draw four curves: actual voltage, predicted voltage, adjusted voltage, and target voltage. It can be seen from the observation that the actual voltage curve shows the dynamic changes of the original voltage fluctuation, and the predicted voltage curve is close to the actual voltage curve, indicating that the model can accurately capture the voltage change law. The adjusted voltage curve gradually approaches the target voltage and eventually stabilizes, indicating that the adjustment strategy is effective and can significantly improve the voltage deviation problem.

[0071] Figure 4 The simulation results of the voltage regulation system are shown, including four curves: actual voltage, predicted voltage, adjusted voltage, and target voltage. The blue dotted line represents the actual voltage value of the system, that is, the real operating voltage data without any adjustment. It can be seen that it fluctuates over time, reflecting the dynamic behavior of the system affected by load changes and environmental factors. The orange dotted line represents the voltage value predicted by the model. It is based on the historical input characteristics (such as voltage, reactive power, active power, etc.) for time series prediction. The overall trend should be close to the actual voltage, indicating the model's ability to capture dynamic changes in voltage. The green solid line shows the result after adjusting the predicted voltage, simulating the role of the voltage regulation system, with the goal of gradually adjusting the predicted voltage to the target voltage (220V). As can be seen from the figure, the green solid line gradually approaches the red dotted line, indicating that the regulation strategy has worked and successfully converged the voltage to the target value. The red dotted line represents the target voltage (220V), which is used as a reference benchmark to measure the degree of deviation of the actual voltage, predicted voltage, and adjusted voltage. Overall, the simulation results verify the predictive ability of the model proposed in this patent and the effectiveness of the regulation strategy.

[0072] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A photovoltaic grid-connected voltage control system based on edge computing, characterized in that: include: Solar cell array, photovoltaic inverter module and protocol conversion module; wherein: The photovoltaic inverter module includes a modulation module, a voltage sensor and a communication module. The modulation module is used to modulate the direct current output by the solar cell array into alternating current; the voltage sensor is used to detect the alternating current voltage data output by the modulation module; the communication module is used to receive the voltage data collected by the voltage sensor on the one hand, and send the received voltage adjustment parameters to the modulation module on the other hand to control the alternating current voltage output by the modulation module; The protocol conversion module includes a protocol conversion unit and a voltage regulation system. The protocol conversion unit receives the AC voltage data sent by the communication module on the one hand, and obtains the voltage allowed by the grid when the photovoltaic grid is connected to the grid on the other hand; the voltage regulation system obtains the AC voltage data output by the modulation module and the target grid-connected voltage data of the grid from the protocol conversion unit on the one hand, generates voltage regulation parameters based on model predictive control on the one hand, and sends the voltage regulation parameters to the communication module on the other hand; Generate voltage regulation parameters based on model predictive control, including: The voltage prediction model is constructed as follows: in, The voltage output predicted by the voltage regulation system at time t+1, is the system state variable at time t, is the control variable in the voltage regulation system, For the voltage prediction model based on recurrent neural network, are the learnable parameters of the model; The predictive control objective function is constructed as follows: in, Predicting voltage output for voltage regulation systems, The preset target value of the modulation module output voltage, is the length of the prediction time domain, To control the length of the time domain, , is the control input change at time t, which satisfies the constraint , is the control input lower limit, For control input upper limit, , To control the upper limit of input change rate, To control the lower limit of input change rate, Weight parameters for balancing tracking error and changes in control input; Solve the predictive control objective function to obtain the predicted value of the modulation module output voltage and the predicted value of the voltage regulation parameter.

2. A photovoltaic grid-connected voltage control system based on edge computing according to claim 1, characterized in that: Solving the predictive control objective function to obtain the predicted value of the modulation module output voltage and the predicted value of the voltage regulation parameter includes: The predictive control objective function is derived, and the optimal solution is iteratively solved by the gradient descent method.

3. A photovoltaic grid-connected voltage control system based on edge computing according to claim 1, characterized in that: The voltage regulation system communicates with the relay service in real time and uploads the acquired information to the management end, which facilitates information collection and control of the terminal and the master station.

4. A photovoltaic grid-connected voltage control system based on edge computing according to claim 1, characterized in that: The voltage control system also includes a remote control terminal, which can run an APP program to achieve 7x24 system monitoring and manual adjustment.

5. A photovoltaic grid-connected voltage control method based on edge computing, implemented based on the control system according to any one of claims 1 to 4, characterized in that: The following steps are involved: S1. Obtain the output voltage of the current photovoltaic inverter module through a voltage sensor, and send the collected output voltage information to the protocol conversion unit of the protocol conversion module through a communication module; S2, the protocol conversion unit sends the output voltage information to the voltage regulation system on the one hand, and sends the voltage information to the power grid on the other hand; S3, the voltage regulation system predicts the future output voltage based on historical data and current status, sets the target voltage and the objective function of minimizing energy loss for optimization control, and the optimization control adopts PID control; S4. The output voltage predicted by the voltage regulation system is transmitted back to the modulation module of the photovoltaic inverter module, and the modulation module regulates the output voltage of the photovoltaic inverter module according to the predicted output voltage.

Images