Energy control method for photovoltaic system

Abstract

The invention discloses an energy regulation method for a photovoltaic system, and relates to the field of energy regulation for a photovoltaic system. The technical scheme thereof mainly includes: collecting comprehensive state data within a time period T, and processing the comprehensive state data to obtain time series state data; using a sliding average filter to preliminarily process the time series state data to obtain a smoothed initial data set, using the initial data set as an observation input of a Kalman filter, further smoothing the data using the Kalman filter to obtain a high-precision gain data set, and using an improved P&O algorithm to perform disturbance judgment on the gain data set; using an IncCond method to eliminate power oscillations of a voltage in a prediction interval, and performing segmented demand judgment, using a segmented search method to accurately locate a maximum power point, and outputting it as a system operating point, thereby realizing energy regulation of the photovoltaic system at the maximum power point and improving the energy efficiency of the photovoltaic system.

Description

Technical Field

[0001] The present invention relates to the field of photovoltaic system energy regulation, and more specifically, to a photovoltaic system energy regulation method. Background Art

[0002] The maximum power point plays a vital role in the energy regulation of photovoltaic systems. The maximum power point of a photovoltaic system is the maximum power that a photovoltaic module can output under specific light and temperature conditions. By ensuring that the system operates at this point, the energy output of the photovoltaic module can be maximized. In power plants or commercial photovoltaic systems, improving energy conversion efficiency means being able to produce more electricity under the same conditions, which directly translates into increased economic benefits. When integrated into a power grid or power system, ensuring that the photovoltaic system operates at the maximum power point can improve the efficiency of energy scheduling and make the use and distribution of photovoltaic power more efficient.

[0003] Especially when light conditions and environmental factors are constantly changing, continuous tracking of the maximum power point can ensure that the system always operates at the highest efficiency, thereby maximizing daily power generation. If the system does not correctly track the maximum power point, it will result in energy loss. This not only affects the total power generation of the system, but may also cause long-term instability of the system. Light and temperature changes will affect the performance of the photovoltaic modules. When the light changes rapidly or the system starts, the speed of tracking the maximum power point may be slow, resulting in the system being unable to reach or maintain the optimal power point in time. In addition, when accurately locating the maximum power point, oscillations or errors may occur, and the maximum power point cannot be accurately found. Therefore, when regulating the energy of the photovoltaic system through the maximum power point, innovative design is required. Summary of the invention

[0004] In view of the problems existing in the prior art, the purpose of the present invention is to provide an energy regulation method for a photovoltaic system, so as to achieve energy regulation of the photovoltaic system at a maximum power point.

[0005] To achieve the above object, the present invention provides the following technical solution: an energy control method for a photovoltaic system, the energy control method for a photovoltaic system comprising:

[0006] Step S1: Collecting comprehensive status data within a time period T, and processing the comprehensive status data to obtain time series status data;

[0007] Step S2: Optimize the time series state data to obtain a gain data set, and use the improved P&O algorithm to perform disturbance judgment on the gain data set;

[0008] Step S3: Use IncCond method to eliminate power oscillation of voltage in the prediction interval, make segmented demand judgment, use segmented search method to accurately locate the maximum power point, and output it as the operating point of the system.

[0009] Preferably, the comprehensive status data includes the voltage, current and power of the photovoltaic module, as well as the ambient temperature and light intensity. The voltage is measured by a voltage sensor, the current is measured by a current sensor or a Hall effect sensor, the power is obtained by calculation or a power sensor, the ambient temperature is measured by a temperature sensor, and the light intensity is measured using a light sensor.

[0010] Preferably, the method for processing the comprehensive status data comprises:

[0011] Set the time axis in chronological order, discretize the time period T into R_O time points on average, collect comprehensive status data at each time point, fill the comprehensive status data into the time point on the time axis, obtain R_O comprehensive status data with time series attributes, and aggregate R_O comprehensive status data to form time series status data.

[0012] Preferably, the method for optimizing the time series state data includes:

[0013] The time series state data is preliminarily processed using a sliding average filter to obtain a smoothed initial data set, which is used as the observation input of the Kalman filter. The Kalman filter is used to further smooth the data to obtain a high-precision gain data set.

[0014] The process of using sliding average filtering to preliminarily process the time series state data is as follows: select a sliding window with a step size of O_P, calculate the sliding average in, represents the output of the sliding average filter at time point t, x t-i represents the data in the time series state data at time point i before time point t. The number of time points in the sliding window at time point t is x ro is the data in the time series state data at the roth time point in the time period T, is the mean value of the data in the time series state data within the time period T, α is the growth coefficient, N is the rated value, and the sliding average filter is applied to the voltage, current and power as well as the ambient temperature and light intensity data in the time series state data to obtain the initial data set after preliminary smoothing;

[0015] The process of further smoothing data using Kalman filtering is as follows: Set the data at the first time point after moving average filtering as the initial state estimate, set the initial covariance matrix, state transition matrix, control input matrix, and observation matrix, calculate the Kalman gain, update the state estimate and covariance matrix of the data until the covariance matrix gradually converges and the fluctuation error tends to be stable, and then stop. Apply Kalman filtering to the data of voltage, current, power, ambient temperature, and light intensity in the initial dataset to obtain the gain dataset.

[0016] Preferably, the method of using the improved P&O algorithm to judge the perturbation of the gain dataset includes:

[0017] Extract the voltage and power in the gain dataset, assume them as the initial voltage V(k) and initial power P(k) at time point k, apply a dynamic perturbation step ΔV_dt to the voltage V(k), the direction of the perturbation is to increase or decrease the voltage, denote the new obtained voltage as the measured voltage V(k + 1), and under the measured voltage V(k + 1), obtain the newly output measured power P(k + 1);

[0018] Compare the measured power P(k + 1) with the initial P(k). If P(k + 1) > P(k), it means the power is increasing, indicating that the voltage adjustment direction is correct. Take the measured voltage V(k + 1) and the measured power P(k + 1) as the new initial voltage and initial power, and continue to perform the perturbation along this direction using the new dynamic perturbation step ΔV_dt until the absolute difference between the measured power and the initial power is less than or equal to the threshold Q_w or no longer changes, and then obtain the prediction interval of the maximum power point. If P(k + 1) < P(k), it means the power is decreasing, indicating that the voltage adjustment direction is incorrect, and readjust in the opposite direction.

[0019] Preferably, the setting process of the dynamic perturbation step ΔV_dt is as follows:

[0020] Set the initial perturbation step ΔV_1, the maximum step ΔV_max, the minimum step ΔV_min, and the threshold ΔP_th of power change;

[0021] After each perturbation, calculate the real-time change amount ΔP_sb of the power. If ΔP_sb > ΔP_th, then update the perturbation step ΔV_dt = min(ΔV_1×β1, ΔV_max), where β1 is the increasing factor and β1 > 1. If ΔP_sb ≤ ΔP_th, then update the perturbation step ΔV_dt = max(ΔV_1×β2, ΔV_max), where β2 is the decreasing factor and 0 < β2 < 1.

[0022] During each perturbation process, use the updated dynamic perturbation step ΔV_dt to perform the perturbation;

[0023] The neural network model is used to obtain the initial perturbation step length ΔV_1, and the method includes:

[0024] Construct a sample set, import the sample set into the neural network model for training, use the mean square error as the loss function, obtain the trained neural network model until the loss function value no longer changes, then import the comprehensive state data at time point k into the trained neural network model, and obtain the initial perturbation step length ΔV_1 at time point k;

[0025] The construction of the sample set includes: collecting multiple historical data subsets, dividing them into multiple batches, and importing the data subsets in batches. In the data subsets, voltage, current and power difference as well as ambient temperature and light intensity are used as model inputs, and the disturbance step size ΔV_1 is used as the output of the model, where the power difference is the difference between the power at the current time point and the previous time point.

[0026] Preferably, the method of eliminating power oscillation of voltage in the prediction interval by using the IncCond method and performing segmented demand judgment includes:

[0027] Set the initial smoothed power value and select the smoothing factor θ, obtain the voltage V(z), current I(z) and power P(z) at time point z, and calculate the smoothed power at time point z in, represents the smoothed power at time point z-1 before time point z;

[0028] Calculating power changes Voltage change ΔV(z) = V(z) - V(z-1) and admittance increment

[0029] If ΔI≠0, it means that the current voltage has not reached the maximum power point, and the subdivision search method needs to be used to continue the analysis. If ΔI=0, it means that the current voltage has reached the maximum power point.

[0030] Preferably, the method for accurately locating the maximum power point using the subdivision search method comprises:

[0031] Extract the prediction interval of the maximum power point output by the improved P&O algorithm, set the prediction interval to [V_start, V_end], V_start represents the maximum voltage in the prediction interval, V_end represents the minimum voltage in the prediction interval, set the coarse step size ΔV_cl and the fine step size ΔV_jx, and the coarse step size ΔV_cl is greater than the fine step size ΔV_jx, the coarse step size ΔV_cl is used as the initial interval for coarse segmentation, and the fine step size ΔV_jx is used as the final interval for refined segmentation;

[0032] Roughly divide the voltage within the prediction interval with a rough step size ΔV_cl to obtain a voltage coarse point Vc_v=V_start+(c_v-1)×ΔV_cl, where Vc_v represents the c_vth voltage coarse point roughly divided within the prediction interval. For each selected voltage coarse point Vc_v, measure the corresponding coarse point power Pc_v;

[0033] Calculate the difference between any two coarse point powers Pc_v, select the voltage coarse point corresponding to the two coarse point powers Pc_v when the difference is the smallest, form a refined interval, and perform fine segmentation within the refined interval with the fine step size ΔV_jx as the final interval. Discretize the refined interval into voltage fine points, extract the corresponding fine point power, and select the maximum value of the fine point power as the maximum power point.

[0034] Preferably, the voltage and current corresponding to the maximum power point are recorded as the optimal voltage and optimal current, the reference voltage or control signal of the converter is adjusted, the parameters are output according to the optimal voltage, the current limit of the converter is set to the optimal current, and the switching frequency and duty cycle are adjusted so that the photovoltaic system can operate stably under the set optimal voltage and optimal current.

[0035] Preferably, during the regulation of the optimal voltage and the optimal current, feedback control is used to adjust the input and output in real time, so that the voltage and current of the system are stable and close to the maximum power point.

[0036] Compared with the prior art, the present invention has the following beneficial effects:

[0037] In the present invention, by comprehensively using P&O, IncCond and subdivision search method, the photovoltaic system can accurately track the maximum power point, significantly improving the power generation efficiency. The dynamic disturbance step and feedback control mechanism adopted reduce the power oscillation and tracking error of the system, and improve the operation stability of the overall system. The combined processing of sliding average filtering and Kalman filtering improves the smoothness and accuracy of the data, so that the MPPT algorithm can be optimized based on more accurate data. The dynamic disturbance step predicted by the neural network model improves the tracking speed of the system, so that it can adapt to environmental changes more quickly and enhance the response speed. Accurate MPP tracking and optimization strategy enables the photovoltaic system to maximize the power generation under different lighting and environmental conditions, directly increasing economic benefits. Improve the stability and reliability of the system, and reduce the maintenance cost and downtime caused by system instability. The system can quickly adapt to changes in light and temperature, and remain near the optimal working point, so as to achieve efficient power generation under various environmental conditions. The adaptive control optimization mechanism continuously improves the overall efficiency of intelligent analysis, and multiple modules cooperate with each other to achieve comprehensive data collection, processing and analysis, provide real-time and accurate analysis of detection data, and realize energy regulation of the photovoltaic system at the maximum power point. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 A schematic diagram of a method for energy regulation of a photovoltaic system proposed by the present invention;

[0039] Figure 2 It is a structural schematic diagram of the energy regulation method applied to the photovoltaic system in the present invention. DETAILED DESCRIPTION

[0040] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0041] Embodiment 1

[0042] Reference Figure 1 and Figure 2 , Example 1 further illustrates the energy control method of a photovoltaic system proposed by the present invention.

[0043] Maximum power point tracking (MPPT) provides several key benefits in the energy regulation of photovoltaic systems, including improving energy conversion efficiency, enhancing system economy, improving system stability and reliability, optimizing energy management efficiency, and enhancing adaptability to extreme environmental conditions. These benefits combined help ensure that photovoltaic systems can operate in the best state, thereby achieving maximum power generation benefits and system performance.

[0044] In the maximum power point tracking (MPPT) process of photovoltaic systems, in order to achieve efficient energy regulation, multiple key data need to be collected and processed in real time.

[0045] Collecting comprehensive status data within a time period T, and processing the comprehensive status data to obtain time series status data;

[0046] The comprehensive status data includes the voltage, current and power of the photovoltaic module, as well as the ambient temperature and light intensity. The voltage is measured by a voltage sensor, the current is measured by a current sensor or a Hall effect sensor, the power is obtained by calculation or a power sensor, the ambient temperature is measured by a temperature sensor, and the light intensity is measured using a light sensor.

[0047] Voltage is one of the electrical characteristics of the PV module output that is used to calculate power and perform maximum power point tracking. It is measured by a voltage sensor. Current is another electrical characteristic of the PV module output that is used together with voltage to calculate power. It is measured by a current sensor or a Hall effect sensor. Power is the product of the PV module voltage and current and is a key data for calculating the maximum power point. Power can be calculated in real time or directly measured by a power sensor.

[0048] Ambient temperature has a significant impact on the performance of PV modules. Understanding temperature variations helps adjust system parameters to maintain a stable power output. Light intensity directly affects the power generation ability of PV modules, as measured by temperature sensors such as thermocouples or digital temperature sensors. High light intensity generally increases the power output of PV modules. This is measured using light sensors such as photoelectric sensors or light intensity sensors.

[0049] By collecting data such as the voltage, current, power, ambient temperature and light intensity of the photovoltaic module and using appropriate sensors for data collection, the performance of the photovoltaic system can be monitored in real time and effective maximum power point tracking (MPPT) can be implemented. This data-driven approach can ensure that the photovoltaic system always operates in the best state under different environmental conditions, thereby improving energy conversion efficiency, system stability and overall performance.

[0050] Methods for processing comprehensive status data include:

[0051] Set the time axis in chronological order, discretize the time period T into R_O time points on average, collect comprehensive status data at each time point, fill the comprehensive status data into the time point on the time axis, obtain R_O comprehensive status data with time series attributes, and aggregate R_O comprehensive status data to form time series status data.

[0052] Optimize the time series state data to obtain the gain data set, and use the improved P&O algorithm to perform disturbance judgment on the gain data set;

[0053] Methods for optimizing the processing of time series state data include:

[0054] The time series state data is preliminarily processed using a sliding average filter to obtain a smoothed initial data set, which is used as the observation input of the Kalman filter. The Kalman filter is used to further smooth the data to obtain a high-precision gain data set.

[0055] The data processed by Kalman filtering will be the result of comprehensive consideration of preliminary smoothing and Kalman filtering, so that it can enjoy the smoothness of sliding average filtering and the estimation accuracy of Kalman filtering at the same time.

[0056] The process of using sliding average filtering to preliminarily process the time series state data is as follows: select a sliding window with a step size of O_P (the window step size is usually 3, 5, 7, etc.), calculate the sliding average in, represents the output of the sliding average filter at time point t, x t-i represents the data in the time series state data at time point i before time point t. The number of time points in the sliding window at time point t is x ro is the data in the time-series status data at the r0-th time point within the time period T. is the mean value of the data in the time-series status data within the time period T, α is the growth coefficient, N is the rated value. The moving average filter is applied to the data of voltage, current, power, ambient temperature, and light intensity in the time-series status data to obtain the initial data set after preliminary smoothing.

[0057] The process of further smoothing the data using the Kalman filter is as follows: Set the data at the first time point after the moving average filter as the initial state estimate, set the initial covariance matrix, state transition matrix, control input matrix, and observation matrix, calculate the Kalman gain, update the state estimate and covariance matrix of the data until the covariance matrix gradually converges and the fluctuation error tends to be stable, and then stop. Apply the Kalman filter to the data of voltage, current, power, ambient temperature, and light intensity in the initial data set to obtain the gain data set.

[0058] The basic idea of the P&O algorithm is to perturb the operating voltage of the photovoltaic module and observe the power change to determine whether it is close to the maximum power point. Adjust the voltage direction according to the power change to find the maximum power point.

[0059] The method of using the improved P&O algorithm to perform perturbation judgment on the gain data set includes:

[0060] Extract the voltage and power in the gain data set, assume them as the initial voltage V(k) and initial power P(k) at time point k. Apply a dynamic perturbation step ΔV_dt to the voltage V(k), and the perturbation direction is to increase or decrease the voltage (i.e., -ΔV or +ΔV). Denote the obtained new voltage as the measured voltage V(k + 1). Under the measured voltage V(k + 1), obtain the newly output measured power P(k + 1).

[0061] Compare the measured power P(k + 1) with the initial P(k). If P(k + 1) > P(k), it means the power is increasing, indicating that the voltage adjustment direction is correct. Take the measured voltage V(k + 1) and measured power P(k + 1) as the new initial voltage and initial power, and continue to perform perturbation along this direction using the new dynamic perturbation step ΔV_dt until the absolute difference between the measured power and the initial power (i.e., the change in power before and after the perturbation) is less than or equal to the threshold Q_w or no longer changes, and then stop to obtain the prediction interval of the maximum power point. If P(k + 1) < P(k), it means the power is decreasing, indicating that the voltage adjustment direction is wrong, and readjust in the opposite direction.

[0062] As the core of the improved P&O algorithm, the perturbation step size cannot be too large, otherwise the maximum power point may be missed; nor can it be too small, otherwise the tracking speed will be slow. Therefore, it is dynamically adjusted according to the speed of real-time power changes to improve tracking efficiency.

[0063] The setting process of dynamic disturbance step length ΔV_dt is:

[0064] Set the initial disturbance step size ΔV_1, the maximum step size ΔV_max, the minimum step size ΔV_min and the power change threshold ΔP_th to determine the speed of power change;

[0065] After each disturbance, calculate the real-time change of power ΔP_sb. If ΔP_sb>ΔP_th, update the disturbance step size ΔV_dt=min(ΔV_1×β1,ΔV_max), where β1 is the increase factor and β1>1. If ΔP_sb≤ΔP_th, update the disturbance step size ΔV_dt=max(ΔV_1×β2,ΔV_max), where β2 is the reduction factor and 0<β2<1.

[0066] In each disturbance process, the updated dynamic disturbance step size ΔV_dt is used for disturbance;

[0067] The neural network model is used to obtain the initial perturbation step length ΔV_1, and the method includes:

[0068] Construct a sample set, import the sample set into the neural network model for training, use the mean square error as the loss function, obtain the trained neural network model until the loss function value no longer changes, then import the comprehensive state data at time point k into the trained neural network model, and obtain the initial perturbation step length ΔV_1 at time point k;

[0069] The construction of the sample set includes: collecting multiple historical data subsets, dividing them into multiple batches, and importing the data subsets in batches. In the data subsets, voltage, current and power difference as well as ambient temperature and light intensity are used as model inputs, and the disturbance step length ΔV_1 is used as the output of the model, where the power difference is the difference between the power at the current time point and the previous time point;

[0070] The IncCond method is used to eliminate the power oscillation of the voltage in the prediction interval, and the segmented demand judgment is performed. The segmented search method is used to accurately locate the maximum power point and output it as the operating point of the system;

[0071] The IncCond method determines the maximum power point by comparing the admittance (conductance) increment of the photovoltaic array. The basic principle is that at the maximum power point, the derivative of power is zero, that is, the conductance (negative value of the derivative) and the increment (derivative) of voltage are equal.

[0072] Methods for using the IncCond method to eliminate power oscillations in the voltage prediction interval and to make segmented demand judgments include:

[0073] Set the initial smoothed power value and select the smoothing factor θ, obtain the voltage V(z), current I(z) and power P(z) at time point z, and calculate the smoothed power at time point z in, represents the smoothed power at time point z-1 before time point z;

[0074] Calculating power changes Voltage change ΔV(z) = V(z) - V(z-1) and admittance increment

[0075] If ΔI≠0, it means that the current voltage has not reached the maximum power point, and the subdivision search method needs to be used to continue the analysis. If ΔI=0, it means that the current voltage has reached the maximum power point;

[0076] The subdivision search method is a method for more accurate search when approaching the maximum power point. The basic step is to divide the voltage range into smaller intervals for fine search after approaching the maximum power point.

[0077] Methods for using the segmentation search method to accurately locate the maximum power point include:

[0078] Extract the prediction interval of the maximum power point output by the improved P&O algorithm, set the prediction interval to [V_start, V_end], V_start represents the maximum voltage in the prediction interval, V_end represents the minimum voltage in the prediction interval, set the coarse step size ΔV_cl and the fine step size ΔV_jx, and the coarse step size ΔV_cl is greater than the fine step size ΔV_jx, the coarse step size ΔV_cl is used as the initial interval for coarse segmentation, and the fine step size ΔV_jx is used as the final interval for refined segmentation;

[0079] Roughly divide the voltage within the prediction interval with a rough step size ΔV_cl to obtain a voltage coarse point Vc_v=V_start+(c_v-1)×ΔV_cl, where Vc_v represents the c_vth voltage coarse point roughly divided within the prediction interval. For each selected voltage coarse point Vc_v, measure the corresponding coarse point power Pc_v;

[0080] Calculate the difference between any two coarse point powers Pc_v, select the voltage coarse point corresponding to the two coarse point powers Pc_v when the difference is the smallest, form a refined interval, and perform fine segmentation within the refined interval with the fine step size ΔV_jx as the final interval. Discretize the refined interval into voltage fine points, extract the corresponding fine point power, and select the maximum value of the fine point power as the maximum power point.

[0081] After determining the maximum power point (MPP), the system parameter settings are adjusted to stably operate at this point to ensure that the efficiency of the PV system is maximized.

[0082] The voltage and current corresponding to the maximum power point are recorded as the optimal voltage and optimal current. The reference voltage or control signal of the converter is adjusted, the parameter output is performed according to the optimal voltage, the current limit of the converter is set to the optimal current, and the switching frequency and duty cycle are adjusted to make the photovoltaic system operate stably under the set optimal voltage and optimal current.

[0083] In the process of regulating the optimal voltage and current, a feedback control system (e.g., proportional-integral-differential controller (PID)) is used to adjust the input and output in real time, so that the voltage and current of the system are stable and close to the maximum power point, and the gain and parameters of the controller are adjusted to make the system respond faster and reduce overshoot and steady-state errors. When the lighting conditions change, the operating point is adjusted dynamically. The system should be able to update the maximum power point in real time and adjust the operating point to keep it near the latest maximum power point.

[0084] By combining P&O, IncCond and the segmentation search method, a dynamic switching mechanism is formed. Based on power changes, the power change threshold is set, and the P&O method and IncCond method are dynamically switched; based on derivative fluctuations: the fluctuations of incremental conductance and conductance value are detected to determine whether it is necessary to switch to the segmentation search method, which can realize the dynamic MPPT of the photovoltaic system and ensure efficient and stable operation under different conditions. In the initial search stage, the P&O method is used for rapid positioning, and then the IncCond method is used for precise adjustment, and finally the segmentation search method is used for fine optimization when necessary. This combined method combines the advantages of each algorithm and improves the response speed and accuracy of the system.

[0085] Embodiment 2

[0086] Reference Figure 1 and Figure 2 , Example 2 further illustrates the energy control method of a photovoltaic system proposed by the present invention.

[0087] The energy control method of a photovoltaic system comprises the following steps:

[0088] Step S1: Collecting comprehensive status data within a time period T, and processing the comprehensive status data to obtain time series status data;

[0089] Step S2: Optimize the time series state data to obtain a gain data set, and use the improved P&O algorithm to perform disturbance judgment on the gain data set;

[0090] Step S3: Use IncCond method to eliminate power oscillation of voltage in the prediction interval, make segmented demand judgment, use segmented search method to accurately locate the maximum power point, and output it as the operating point of the system.

[0091] The energy control system of a photovoltaic system, applied to an energy control method of a photovoltaic system, comprises:

[0092] Data acquisition module: used to collect raw data from various areas of the city, perform time series processing on the raw data, and obtain comprehensive data packets in a time series state;

[0093] Data optimization module: Use sliding average filtering to preliminarily process the time series state data to obtain a smoothed initial data set, use the initial data set as the observation input of the Kalman filter, use the Kalman filter to further smooth the data, obtain a high-precision gain data set, and use the improved P&O algorithm to perform disturbance judgment on the gain data set;

[0094] Control point module: Use IncCond method to eliminate power oscillation of voltage in the prediction interval, make segmented demand judgment, use segmented search method to accurately locate the maximum power point, and use it as the system's operating point for voltage output;

[0095] The modules are connected to each other via wired and / or wireless means.

[0096] In addition, according to the implementation of the present application, the process described in the attached figure of a method for regulating energy of a photovoltaic system can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, which stores machine-readable instructions, and the machine-readable instructions can be executed by a processor to execute instructions corresponding to the method steps provided in the present application. Of course, the architecture shown in the attached figure of a method for regulating energy of a photovoltaic system is only exemplary, and when implementing different devices, adaptive selection or adjustment can be performed according to actual needs.

[0097] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and thresholds in the formula are set by technicians in this field according to actual conditions.

[0098] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technical users in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A method for regulating energy of a photovoltaic system, characterized in that: The energy control method of the photovoltaic system includes: Step S1: Collecting comprehensive status data within a time period T, and processing the comprehensive status data to obtain time series status data; Step S2: Optimize the time series state data to obtain a gain data set, and use the improved P&O algorithm to perform disturbance judgment on the gain data set; Step S3: Use IncCond method to eliminate power oscillation of voltage in the prediction interval, make segmented demand judgment, use segmented search method to accurately locate the maximum power point, and output it as the operating point of the system; The method of using the IncCond method to eliminate power oscillation of voltage in the prediction interval and to perform segmented demand judgment includes: Set the initial smoothed power value and select the smoothing factor θ, obtain the voltage V(z), current I(z) and power P(z) at time point z, and calculate the smoothed power at time point z in, represents the smoothed power at time point z-1 before time point z; Calculating power changes Voltage change ΔV(z) = V(z) - V(z-1) and admittance increment If ΔI≠0, it means that the current voltage has not reached the maximum power point, and the subdivision search method needs to be used to continue the analysis. If ΔI=0, it means that the current voltage has reached the maximum power point; The method for accurately locating the maximum power point using the subdivision search method comprises: Extract the prediction interval of the maximum power point output by the improved P&O algorithm, set the prediction interval to [V_start, V_end], V_start represents the maximum voltage in the prediction interval, V_end represents the minimum voltage in the prediction interval, set the coarse step size ΔV_cl and the fine step size ΔV_jx, and the coarse step size ΔV_cl is greater than the fine step size ΔV_jx, the coarse step size ΔV_cl is used as the initial interval for coarse segmentation, and the fine step size ΔV_jx is used as the final interval for refined segmentation; Roughly divide the voltage within the prediction interval with a rough step size ΔV_cl to obtain a voltage coarse point Vc_v=V_start+(c_v-1)×ΔV_cl, where Vc_v represents the c_vth voltage coarse point roughly divided within the prediction interval. For each selected voltage coarse point Vc_v, measure the corresponding coarse point power Pc_v; Calculate the difference between any two coarse point powers Pc_v, select the voltage coarse point corresponding to the two coarse point powers Pc_v when the difference is the smallest, form a refined interval, and perform fine segmentation within the refined interval with the fine step size ΔV_jx as the final interval. Discretize the refined interval into voltage fine points, extract the corresponding fine point power, and select the maximum value of the fine point power as the maximum power point.

2. The energy control method of a photovoltaic system according to claim 1, characterized in that: The comprehensive status data includes the voltage, current and power of the photovoltaic module, as well as the ambient temperature and light intensity. The voltage is measured by a voltage sensor, the current is measured by a current sensor or a Hall effect sensor, the power is obtained by calculation or a power sensor, the ambient temperature is measured by a temperature sensor, and the light intensity is measured using a light sensor.

3. The energy control method of a photovoltaic system according to claim 2, characterized in that: The method for processing the comprehensive status data comprises: Set the time axis in chronological order, discretize the time period T into R_O time points on average, collect comprehensive status data at each time point, fill the comprehensive status data into the time point on the time axis, obtain R_O comprehensive status data with time series attributes, and aggregate R_O comprehensive status data to form time series status data.

4. The energy control method of a photovoltaic system according to claim 3, characterized in that: The method for optimizing the time series state data includes: The time-series state data is preliminarily processed using a moving average filter to obtain a smoothed initial data set. The initial data set is used as the observation input for the Kalman filter to further smooth the data and obtain a high-precision gain data set. The process of using sliding average filtering to preliminarily process the time series state data is as follows: select a sliding window with a step size of O_P, calculate the sliding average in, represents the output of the sliding average filter at time point t, x t-i represents the data in the time series state data at time point i before time point t. The number of time points in the sliding window at time point t is x ro is the data in the time series state data at the roth time point in the time period T, is the mean value of the data in the time series state data within the time period T, α is the growth coefficient, N is the rated value, and the sliding average filter is applied to the voltage, current and power as well as the ambient temperature and light intensity data in the time series state data to obtain the initial data set after preliminary smoothing; The process of further smoothing the data using the Kalman filter is as follows: Set the data at the first time point after the moving average filter as the initial state estimate. Set the initial covariance matrix, state transition matrix, control input matrix, and observation matrix. Calculate the Kalman gain and update the state estimate and covariance matrix of the data until the covariance matrix gradually converges and the fluctuation error tends to be stable. Apply the Kalman filter to the data of voltage, current, power, ambient temperature, and light intensity in the initial data set to obtain the gain data set.

5. The energy control method of a photovoltaic system according to claim 4, characterized in that: The method of using the improved P&O algorithm to judge the perturbation of the gain data set includes: Extract the voltage and power from the gain data set, assumed to be the initial voltage V(k) and initial power P(k) at time point k. Apply a dynamic perturbation step ΔV_dt to the voltage V(k), and the perturbation direction is to increase or decrease the voltage. Denote the new voltage obtained as the measured voltage V(k + 1). Under the measured voltage V(k + 1), obtain the newly output measured power P(k + 1). Compare the measured power P(k + 1) with the initial P(k). If P(k + 1) > P(k), it means the power is increasing, indicating that the voltage adjustment direction is correct. Take the measured voltage V(k + 1) and the measured power P(k + 1) as the new initial voltage and initial power, and continue to perform the perturbation in this direction using the new dynamic perturbation step ΔV_dt until the absolute difference between the measured power and the initial power is less than or equal to the threshold Q_w or no longer changes, and obtain the prediction interval of the maximum power point. If P(k + 1) < P(k), it means the power is decreasing, indicating that the voltage adjustment direction is incorrect, and readjust in the opposite direction.

6. The energy control method of a photovoltaic system according to claim 5, characterized in that: The setting process of the dynamic perturbation step ΔV_dt is as follows: Set the initial perturbation step ΔV_1, the maximum step ΔV_max, the minimum step ΔV_min, and the threshold ΔP_th of power change. After each perturbation, calculate the real-time change amount ΔP_sb of the power. If ΔP_sb > ΔP_th, update the perturbation step ΔV_dt = min(ΔV_1×β1, ΔV_max), where β1 is the increasing factor and β1 > 1. If ΔP_sb ≤ ΔP_th, update the perturbation step ΔV_dt = max(ΔV_1×β2, ΔV_max), where β2 is the decreasing factor and 0 < β2 < 1. During each perturbation process, use the updated dynamic perturbation step ΔV_dt for perturbation. Among them, the method of using a neural network model to obtain the initial perturbation step ΔV_1 includes: Construct a sample set, import the sample set into the neural network model for training, use the mean square error as the loss function, obtain the trained neural network model until the loss function value no longer changes, then import the comprehensive state data at time point k into the trained neural network model, and obtain the initial perturbation step length ΔV_1 at time point k; The construction of the sample set includes: collecting multiple historical data subsets, dividing them into multiple batches, and importing the data subsets in batches. In the data subsets, voltage, current and power difference as well as ambient temperature and light intensity are used as model inputs, and the disturbance step size ΔV_1 is used as the output of the model, where the power difference is the difference between the power at the current time point and the previous time point.

7. The energy control method of a photovoltaic system according to claim 6, characterized in that: The voltage and current corresponding to the maximum power point are recorded as the optimal voltage and optimal current, the reference voltage or control signal of the converter is adjusted, the parameters are output according to the optimal voltage, the current limit of the converter is set to the optimal current, and the switching frequency and duty cycle are adjusted to make the photovoltaic system operate stably under the set optimal voltage and optimal current.

8. The energy control method of a photovoltaic system according to claim 7, characterized in that: Feedback control is used in the process of regulating the optimal voltage and the optimal current to adjust the input and output in real time, so that the voltage and current of the system are stable and close to the maximum power point.

Images