A method for power management for a photovoltaic power source

Abstract

The application discloses a kind of photovoltaic power supply electricity management method, by obtaining traffic flow, weather forecast, energy storage module state and charging demand data, establishes multi-source fusion prediction model to calculate and predict electricity consumption and photovoltaic power generation, according to the state of charge, health state, temperature and remaining life of energy storage module, calculate priority index, and based on net load power demand and priority index generate energy distribution strategy, control photovoltaic, energy storage and grid collaborative power supply.The application improves prediction accuracy through multi-source data fusion, dynamically calculates the priority of energy storage module, realizes the optimized allocation of energy, can improve photovoltaic self-generation self-use rate, reduce dependence on power grid, reduce power grid access frequency and power transmission loss, prolong the service life of energy storage module, improve the overall operation efficiency of system, guarantee power supply stability and reliability, with good economy and practicality.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of artificial intelligence, and relates to a photovoltaic power supply electricity management method. BACKGROUND

[0002] At present, charging stations with photovoltaic power supply are facing severe challenges under the background of rapid popularization of new energy vehicles. On the one hand, the power generation of the photovoltaic system in the station is highly dependent on natural light conditions, while the charging demand of new energy vehicles has significant spatiotemporal distribution imbalance. During the low peak period in the afternoon, although the power generation of the photovoltaic system is high, the charging demand is relatively low, resulting in a large amount of photovoltaic electric energy being wasted; during the evening and night charging peak, the power generation capacity of the photovoltaic system decreases, while the charging load sharply rises, and the charging station has to rely on the power supply of the large power grid, so that the utilization rate of the photovoltaic power supply is greatly reduced.

[0003] On the other hand, most of the existing charging station energy management systems follow the traditional mode and can only simply distribute the photovoltaic energy storage according to the historical charging data, and cannot accurately capture abnormal electricity demand such as sudden changes in vehicle flow and temporary energy supplement of users. When large-scale activities are held, resulting in a sharp increase in vehicles in the surrounding area, or when vehicles are concentrated for charging due to extreme weather, the energy storage system of the charging station often cannot be optimized in advance due to the lack of fusion analysis of multidimensional data such as vehicle flow and weather changes, and thus the energy storage capacity is insufficient and the photovoltaic power cannot be effectively utilized, which not only affects the charging experience of users, but also increases the operating cost of the charging station. In addition, since the aging state of the photovoltaic components and the health condition of the energy storage batteries are not combined with real-time power generation prediction, there is a phenomenon of excessive dependence on energy storage devices in the energy scheduling process, which accelerates the battery wear and shortens the service life of the devices.

[0004] The traditional user microgrid has basic energy management logic, but only relies on fixed light duration and preset electricity consumption ratio for static calculation, lacks a response mechanism to real-time light intensity, temperature changes and other dynamic environmental factors, and is difficult to adapt to demand fluctuations in complex scenarios such as weather changes and changes in user electricity consumption habits. The management system based on the gradient battery constructs an energy storage module priority evaluation system through the state of charge, health state and temperature parameters, but does not fuse external environmental data such as vehicle flow and weather forecast for forward-looking scheduling, and when a sudden charging peak occurs, it is easy to cause overload of the energy storage module or lag of the power grid connection. The management system based on artificial intelligence introduces vehicle flow data for electricity demand prediction, but does not include battery health state and photovoltaic power generation prediction into a unified decision framework, resulting in insufficient scheduling accuracy of the energy distribution strategy in photovoltaic surplus / shortage scenarios, and making it difficult to realize the coordinated optimization of the power grid-photovoltaic- energy storage.

[0005] The prior art lacks an intelligent management scheme capable of fusing multi-dimensional data, dynamically optimizing energy distribution paths, and realizing coordinated scheduling of a power grid-photovoltaic- energy storage. Therefore, there is an urgent need for a photovoltaic power source electricity management method to solve the above problems. SUMMARY

[0006] To solve the problems in the background art, the present application provides a photovoltaic power source electricity management method, which can quickly and accurately process complex data in the power system, adapt to the dynamic changes of the power system operation state in real time, realize global optimal scheduling of multiple targets, and improve the reliability, economy and stability of the power system operation.

[0007] The first aspect of the present application provides a photovoltaic power source electricity management method, comprising:

[0008] Obtaining traffic flow data, weather forecast data, energy storage module state data and charging demand data;

[0009] Based on the data, a multi-source fusion prediction model is established to calculate the predicted power consumption and photovoltaic power generation; the priority index is calculated based on the state of charge, health state, temperature and remaining life of the energy storage module;

[0010] Based on the net load power demand and the priority index, an energy distribution strategy is generated to control the coordinated power supply of photovoltaic, energy storage and power grid.

[0011] Optionally, the construction method of the multi-source fusion prediction model is: for regular metering periods, a time series model is used , wherein is the predicted power consumption of period Dj, is the average power consumption of the historical T periods, is the previous cycle prediction value, is the adaptive weight; for the to-be-approved metering period, a traffic flow correlation model is used , wherein is the predicted traffic flow of the correlation line, is the correlation coefficient, is the user power consumption characteristic value, is the climate correction coefficient, , is the model parameter.

[0012] Optionally, the charging station is taken as the center to divide a radius The one-way road with a covered length greater than in the range is the intended road, the traffic flow and power consumption ratio of each intended road is calculated, the discrete value is selected, and The minimum road, wherein bp is the average of bs, es is the power consumption, vs is the traffic flow, and n is the data volume.

[0013] Optionally, the calculation model of the photovoltaic power generation is , is the power generation at time t, is the comprehensive efficiency, is the photovoltaic panel area, is the light intensity, is the ambient temperature, is the standard temperature, is the temperature coefficient.

[0014] Optionally, the calculation model of the energy storage module priority index is , is the priority of module s at time t, is the state of charge, is the health state, is the temperature, is the optimal working temperature, and L(s) is the remaining life index, , , , is the weight coefficient.

[0015] Optionally, the generation method of the energy distribution strategy is:

[0016] Calculate the net load power , is the predicted power consumption, is the photovoltaic power generation at time t;

[0017] When ≤0, the energy storage charging power is , is the average state of charge, is the charging efficiency;

[0018] When >0, if , only the energy storage is used for power supply, if , the grid power supply power is , is the real-time energy storage power, , is the threshold proportion, is the lower limit of safety.

[0019] Optionally, the step of optimizing the energy distribution strategy by reinforcement learning, the reward function of the reinforcement learning is wherein R is a reward value, Cost is an operation cost, Emission is a carbon emission, and Satisfaction is a user satisfaction, 、 、 is a weight coefficient.

[0020] Optionally, the energy storage module state data includes SOC, SOH, voltage, current and temperature parameters of the cascade utilization battery pack, the traffic volume data is collected through video recognition and RFID technology, and the weather forecast data includes light intensity, temperature and precipitation probability in a future preset time period.

[0021] Optionally, the energy distribution strategy further includes an energy storage module screening mechanism based on a priority index, and the energy storage modules are selected to participate in charging and discharging in descending order of P(s, t), and the modules are exited from operation when the SOC is lower than a preset threshold or the SOH is lower than a preset threshold.

[0022] The step of controlling the coordinated power supply of the photovoltaic power supply, the energy storage and the power grid comprises:

[0023] The instruction is sent to the switch management module through the high-speed controller area network to control the on-off state of the power grid and the inverter, and the output power of the charging pile is adjusted, wherein the inverter efficiency is used as a correction factor in power calculation.

[0024] Compared with the prior art, the present application has the following beneficial effects:

[0025] The present application provides a photovoltaic power supply power management method, which obtains traffic volume, weather forecast, energy storage module state and charging demand data, establishes a multi-source fusion prediction model to calculate and predict power consumption and photovoltaic power generation, calculates a priority index based on the state of charge, health status, temperature and remaining life of the energy storage module, generates an energy distribution strategy based on the net load power demand and the priority index, and controls the coordinated power supply of the photovoltaic power supply, the energy storage and the power grid. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 is a flowchart of the photovoltaic power supply power management method in an embodiment of the present application. DETAILED DESCRIPTION

[0027] With reference to the accompanying drawings, the technical solutions in the embodiments of the present application will be clearly and completely described below, obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work belong to the protection scope of the present application.

[0028] In an embodiment, as shown in Figure 1 , a photovoltaic power supply power management method is provided, which is applied to Figure 1 for example, and includes the following specific steps:

[0029] S10: acquiring traffic flow data, weather forecast data, energy storage module state data and charging demand data;

[0030] Specifically, the traffic flow data refers to the traffic flow data of the motor vehicles on the road associated with the photovoltaic charging station, including the number of vehicles passing through per unit time, traffic density and vehicle type distribution and other parameters. The traffic flow data is used to predict the charging demand fluctuation and establish the correlation model of traffic flow and power consumption.

[0031] The weather forecast data refers to the meteorological parameter prediction data of the location where the photovoltaic power generation unit is located, including the light intensity, ambient temperature, precipitation probability, wind speed and the like. It is used to predict the photovoltaic power generation capacity and evaluate the influence of weather on the power generation efficiency.

[0032] The energy storage module state data is obtained by the real-time running parameters of each module in the energy storage system, including the state of charge, health status, working temperature, voltage, current and residual life index and the like. It is used to evaluate the available capacity, health status and priority ranking of the energy storage module.

[0033] The charging demand data refers to the data related to the user charging behavior, including the historical charging time, charging power, charging time length and power consumption mode feature vector and the like. It is used to predict the future charging demand and optimize the energy distribution strategy.

[0034] S20: based on the data, a multi-source fusion prediction model is established to calculate the predicted power consumption and photovoltaic power generation capacity.

[0035] Specifically, the multi-source fusion prediction model adopts a "time period-multiple dimension" modeling idea, divides the charging period into regular measurement period and to-be-approved measurement period, respectively uses the time series model and the traffic flow correlation model to predict the power consumption, and constructs the photovoltaic power generation prediction model through the weather forecast data to realize the bidirectional prediction of power consumption and power generation.

[0036] The construction method of the multi-source fusion prediction model is as follows: the time series model is used for the regular measurement period , wherein Predicted electricity consumption for time period Dj, Average electricity consumption for historical T periods, Last period prediction value, Adaptive weight; adopt traffic flow correlation model for approved metering period Wherein Correlation line predicted traffic flow, Correlation coefficient, User electricity consumption characteristic value, Climate correction coefficient, , Model parameters.

[0037] Centered on the charging station, the radius Range covers a length greater than One-way road as the intended road, calculate the traffic flow and electricity consumption ratio of each intended road Through discrete values Select The road with the smallest value, where bp is the mean value of bs, es is the electricity consumption, vs is the traffic flow, and n is the data quantity.

[0038] The calculation model of the photovoltaic power generation amount is Wherein The power generation amount at time t, The comprehensive efficiency, The photovoltaic panel area, The light intensity, The ambient temperature, The standard temperature, The temperature coefficient.

[0039] S30: Calculate the priority index according to the state of charge, health state, temperature and remaining life of the energy storage module.

[0040] Specifically, the priority index model calculates and quantifies the available priority of the energy storage module in the charging and discharging process through multi-dimensional parameter weighting, and the core parameters include state of charge (SOC), state of health (SOH), working temperature and remaining life index. Through weight distribution, the comprehensive evaluation of the module state is realized.

[0041] The calculation model of the priority index of the energy storage module is: Wherein, The priority of module s at time t, The state of charge, The health state, The temperature, The optimal working temperature, L(s) is the remaining life index, , , 、 is a weight coefficient.

[0042] The application takes an energy storage system containing three cascade battery modules as an example, gives the state of charge, health state, working temperature, residual life index and weight coefficient of each module, substitutes into the formula to calculate the priority index of each module and sorts, so as to activate the module with high priority first during discharging, limit the current of the low-priority module during charging, and realize the reduction of module life difference and the improvement of system overall operation efficiency in this way.

[0043] S40: generating an energy distribution strategy based on the net load power demand and the priority index, and controlling the collaborative power supply of photovoltaic, energy storage and power grid.

[0044] Specifically, the generation method of the energy distribution strategy is:

[0045] Calculate the net load power , wherein is the predicted power consumption, is the photovoltaic power at time t;

[0046] When ≤0, the energy storage charging power , wherein Z is the full capacity of the energy storage, is the average state of charge, is the charging efficiency;

[0047] When >0, if , only the energy storage is used for power supply, if , the power grid power , wherein, is the real-time energy storage power, , is the threshold proportion, is the lower limit of safety.

[0048] The optimization strategy further includes the step of optimizing the energy distribution strategy by reinforcement learning, and the reward function of the reinforcement learning is , wherein R is the reward value, Cost is the operation cost, Emission is the carbon emission, Satisfaction is the user satisfaction, , , is a weight coefficient.

[0049] The energy storage module state data includes the state of charge, health state, voltage, current and temperature parameters of the cascade utilization battery pack, the traffic flow data is collected through video recognition and RFID technology, and the weather forecast data includes the light intensity, temperature and precipitation probability in the future preset time period.

[0050] wherein the energy storage module state data covers the state of charge, state of health, voltage, current and temperature parameters of the battery pack for cascade utilization, wherein the state of charge reflects the proportion of the current power to the rated capacity, the state of health represents the health degree and remaining service life of the battery, the voltage and current parameters are used to monitor the charging and discharging state of the battery, and the temperature parameter is related to the safety and performance stability of the battery. The traffic flow data is collected by video recognition and RFID technology, specifically, the high-definition camera deployed on the road combined with computer vision algorithm (such as YOLO target detection model) identifies vehicles and counts the traffic flow, and at the same time, the RFID card reader reads the new energy vehicle electronic tag to obtain the vehicle model, driving trajectory and other information, providing data support for the charging demand prediction. The weather forecast data includes the light intensity, temperature and precipitation probability in the future preset time period (such as 72 hours), wherein the light intensity directly affects the photovoltaic power generation, the temperature parameter is used to modify the photovoltaic panel power generation efficiency model, and the precipitation probability is used as a climate influencing factor for the electricity demand prediction. By accessing the national meteorological data API and real-time monitoring of the on-site weather station, the accuracy and timeliness of the data are ensured.

[0051] The energy distribution strategy also includes a storage module screening mechanism based on priority index, and the storage modules are selected to participate in charging and discharging according to P(s, t) from high to low, and when the state of charge of the module is lower than the preset threshold or the state of health is lower than the preset threshold, the module is exited from operation.

[0052] The step of controlling the coordinated power supply of photovoltaic, energy storage and power grid includes:

[0053] The instructions are sent to the switch management module through the high-speed controller area network to control the on-off state of the power grid and the inverter, and at the same time, the output power of the charging pile is adjusted, wherein the inverter efficiency is used as a correction factor in power calculation.

[0054] Specifically, the energy distribution strategy instructions are transmitted to the switch management module through the high-speed controller area network to realize precise control of the on-off state of the power grid and the inverter. When the system needs the power grid to intervene in power supply, the switch management module closes the corresponding circuit breaker after receiving the instructions, so that the power grid is connected to the power supply loop; if the photovoltaic and energy storage can meet the electricity demand, the power grid is disconnected, and clean energy is used preferentially. At the same time, the output power of the charging pile is adjusted according to the energy distribution strategy to ensure that each charging pile charges the vehicle according to the preset power. In the power calculation process, the inverter efficiency is used as a correction factor in the calculation, for example, in the calculation of the available power of photovoltaic power generation, the "photovoltaic power generation x inverter efficiency" is used as the actual available power to avoid the deviation of power calculation caused by ignoring the inverter loss, so as to realize the coordinated and efficient power supply of photovoltaic, energy storage and power grid, and improve the overall energy utilization rate of the system.

[0055] Although the present application has been described in detail with reference to the foregoing embodiments, the technical solutions recorded in the foregoing embodiments can be modified, or some of the technical features can be replaced by equivalent features, by those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for power management for a photovoltaic power source, the method comprising: The application relates to a method for controlling the collaborative power supply of photovoltaic, energy storage and power grid. The method comprises the following steps: acquiring traffic flow data, weather forecast data, energy storage module state data and charging demand data; the traffic flow data is collected through video recognition and RFID technology, and the weather forecast data comprises illumination intensity, temperature and precipitation probability in a future preset time period; According to the state of charge, the state of health, the temperature and the remaining life of the energy storage module, a priority index is calculated; the calculation model of the priority index of the energy storage module is P(s, t) = ω1* SOC(s, t) + ω2* SOH(s, t) - ω3* |Ta(s, t) - T opt + ω4*L(s), wherein P(s, t) is the priority of the module s at the time t, SOC(s, t) is the state of charge, SOH(s, t) is the state of health, Ta(s, t) is the temperature, T opt is the optimal working temperature, L(s) is the remaining life index, and ω1, ω2, ω3 and ω4 are weight coefficients. establishing a multi-source fusion prediction model based on the data, and calculating predicted power consumption and photovoltaic power generation; the multi-source fusion prediction model adopts a time series model for regular metering periods, adopts a traffic flow correlation model for to-be-approved metering periods, and selects target line traffic flow data associated with a charging station through a specific method; generating an energy distribution strategy based on net load power demand and a priority index, and controlling the collaborative power supply of photovoltaic, energy storage and power grid; the energy distribution strategy is generated by the following method: C(t) = D(t) * Pmax charge (t) = min(D(t), Z * (1 - SOC avg (T)) η ch ), where Z is the energy storage full capacity, SOC avg is the average state of charge, and η ch is the charging efficiency. calculating a net load power D(t)=E(t)-G(t), wherein E(t) is predicted power consumption, and G(t) is photovoltaic power generation; P grid (t) = max(0, D(t) - C(t) - G(t) + Z(q1 - SOC min )), where C(t) is the real-time energy storage power, q0, q1 are threshold ratios, SOC min is the lower limit of safety.

2. The photovoltaic power source electricity management method according to claim 1, characterized by, The method for adopting a time series model for regular metering periods is: wherein ep is the predicted power consumption of period Dj j is the average power consumption of the historical T periods, is the previous cycle prediction value, and a is an adaptive weight. Traffic flow correlation model for approved metering period where V j is the predicted traffic flow of the correlation line, bp is the correlation coefficient, U j is the user power consumption characteristic value, C j is the climate correction coefficient, and β and γ are model parameters.

3. The photovoltaic power source electricity management method according to claim 2, characterized by, The determination method of the association line is as follows: taking the charging station as the center, a one-way road with a covering length greater than r2 in a radius r1 range is defined as an intended road, and a traffic flow and power consumption ratio of each intended road is calculated By discrete values Select the road with the minimum F2, where bp is the mean of bs, es is the power consumption, vs is the traffic flow, and n is the data quantity.

4. The photovoltaic power source electricity management method of claim 1, wherein, when D(t)>0, if C(t)+G(t)-D(t) is greater than or equal to Z*q1, only energy storage is used for power supply, if Z*q1>C(t)+G(t)-D(t) is greater than or equal to Z*q0, power grid power supply is used 5. The photovoltaic power source electricity management method of claim 1, wherein, Further comprising the step of optimizing the energy allocation strategy by reinforcement learning, the reward function of the reinforcement learning being R = ω5*Cost -1 + ω6*Emission -1 + ω7*Satisfaction, wherein R is the reward value, Cost is the operating cost, Emission is the carbon emission, Satisfaction is the user satisfaction, and ω5, ω6, ω7 are weight coefficients.

6. The photovoltaic power source electricity management method of claim 1, wherein, the calculation model of the photovoltaic power generation is G(t)=eta(t)*S*I(t)*[1+Kappa(T(t)-T0)], wherein G(t) is photovoltaic power at t moment, eta(t) is comprehensive efficiency, S is photovoltaic panel area, I(t) is illumination intensity, T(t) is environmental temperature, T0 is standard temperature, and Kappa is a temperature coefficient.

7. The photovoltaic power source electricity management method of claim 4, wherein, The energy storage module state data comprises state of charge, health state, voltage, current and temperature parameters of a battery pack for step utilization.

8. The photovoltaic power source electricity management method of claim 1, wherein, The energy distribution strategy further comprises an energy storage module screening mechanism based on a priority index, and energy storage modules are selected to participate in charging and discharging according to P(s,t) from high to low, and the energy storage modules are exited from operation when the state of charge is lower than a preset threshold or the health state is lower than a preset threshold. The step of controlling the collaborative power supply of photovoltaic, energy storage and power grid comprises the following steps: sending an instruction to a switch management module through a high-speed controller area network, controlling the on-off state of the power grid and an inverter, and adjusting the output power of a charging pile, wherein the inverter efficiency is used as a correction factor in power calculation.

Images