CN-121076891-B - Dynamic energy storage scheduling method of modularized photovoltaic system

CN121076891BCN 121076891 BCN121076891 BCN 121076891BCN-121076891-B

Abstract

The invention discloses a dynamic energy storage scheduling method of a modularized photovoltaic system, which relates to the technical field of operation control of the photovoltaic power generation system and comprises the steps of dividing the photovoltaic power generation system into a plurality of modules with autonomous operation capability, wherein each module comprises a photovoltaic power generation unit, an energy storage unit and a local control unit, the local control unit is used for collecting current illumination intensity, battery charge state, temperature, power grid power flow direction and load power data of the module in real time and periodically uploading the parameters to a central coordination unit of the system, and the central coordination unit is used for constructing a nonlinear load prediction model based on module-level load rate change, external temperature sensitivity and day-night period characteristics so as to improve the response capability to temporary sudden load change.

Inventors

SHEN XIAOBO
DING MINGLIANG
JIANG RUI
HUA XUAN

Assignees

无锡申泰新能源科技有限公司

Dates

Publication Date: 20260512
Application Date: 20250808

Claims (8)

1. The dynamic energy storage scheduling method of the modularized photovoltaic system is characterized by comprising the following steps of: Dividing a photovoltaic power generation system into a plurality of modules with autonomous operation capability, wherein each module comprises a photovoltaic power generation unit, an energy storage unit and a local control unit, and the local control unit is used for collecting current illumination intensity, battery charge state, temperature, power grid power flow direction and load power data of the module in real time and uploading the parameters to a central coordination unit of the system periodically; Step two, the central coordination unit builds a nonlinear load prediction model based on module-level load rate change, external temperature sensitivity and day-night periodic characteristics so as to improve the response capability to temporary sudden load change; step three, when detecting that the modules exceeding the threshold number have excessive photovoltaic power generation power and have lower load in a certain time window, the central coordination unit judges whether the system enters a photovoltaic power impact area according to analysis and calculation results; Step four, after judging that the photovoltaic power impact zone is entered, performing priority scheduling on each energy storage unit by adopting an energy storage weight function constructed by a pressing and exciting composite function; Step five, the system sorts according to the energy storage weight function values of all the modules, and under the condition that the SoC limit, the battery temperature rise constraint and the local load demand are met, the high-weight module is preferentially scheduled to receive the surplus power for charging; Step six, when the system detects that the load is abnormally increased and the load predicted value is greatly higher than the current power generation value or the power grid gives a demand control instruction, the central coordination unit starts an emergency discharge scheduling flow to schedule the energy storage units with high SoC, proper temperature and long service life to participate in discharge support preferentially; in the fourth step, an energy storage weight function constructed by a pressing and exciting composite function is adopted, and the calculation expression is as follows: Wherein, the Representing the current state of charge of the ith energy storage unit; Representing the current battery temperature; indicating an optimal temperature setting; representing the cycle remaining proportion; representing the current real-time electricity price; representing the scheduling stimulus coefficients for adjusting the commercialized weights.
2. The method for dynamic energy storage scheduling of a modularized photovoltaic system according to claim 1, wherein in the first step, the local control unit periodically collects key operation state parameters of the module, and constructs a standardized data structure for unified uploading, and the method specifically comprises the following steps: The local control unit performs high-frequency sampling on parameters of photovoltaic output power, real-time voltage and current of the energy storage unit, temperature of the energy storage battery, load power and power feeding/extracting power of a power grid through a built-in illumination intensity sensor, a current and voltage detection module, a temperature sensor and a power meter, and performs average filtering and uploading in a fixed time window; The uploading data structure comprises a module unique identifier ID, a sampling time stamp, a parameter field set and a module state mark, wherein the parameter field set covers the current photovoltaic power generation power, the charge state of an energy storage unit, the current load power, the environment temperature, the battery temperature, the power grid interaction direction and the power value, and the state mark is used for indicating whether the module is in a communication interruption state, an energy storage fault state and a battery temperature rise early warning state; the data uploading adopts a mode of combining an event driving mechanism and a periodic synchronous mechanism, namely, when basic operation parameters are uploaded in each period, asynchronous quick uploading is immediately triggered when parameter mutation is detected, so that a central coordination unit can dynamically adjust a scheduling strategy according to abnormal changes.
3. The method for dynamic energy storage scheduling of a modularized photovoltaic system according to claim 2, wherein the local control unit autonomously identifies and marks the running state of the module, and reports the state to the central coordination unit in real time through a state flag field for assisting the implementation of the subsequent energy storage scheduling and safety protection strategy, and the state flag at least comprises: The communication state mark is used for indicating whether the communication between the module and the central coordination unit is normal or not, and if the number of times of returning or sending failure of the scheduling instruction is not received in a plurality of continuous periods exceeds the limit, the communication state mark is automatically set to be in a communication interruption state; the energy storage health state sign is used for indicating whether the current energy storage unit has higher internal resistance, voltage drift, abnormal charge and discharge and cycle times reaching the early warning value fault sign, and when the health degree is lower than a set threshold value, the sign is automatically set as 'energy storage degradation' or 'energy storage failure'; the thermal state mark automatically generates a 'temperature rise early warning' or a 'thermal runaway risk' mark when the temperature is abnormal or the temperature rise trend is too fast through the joint analysis of the battery temperature acquisition value and the temperature rise rate; The load state mark is used for identifying whether the local load is in a peak value operation state, a sudden start-stop state or an irregular oscillation state, and if multiple power fluctuation overruns occur in a short time, the system marks as a load disturbance state; When any module reports an abnormality mark, the central coordination unit executes scheduling degradation, scheduling forbidden or local isolation strategies on the module in energy storage scheduling sequencing, and sends out maintenance suggestions, so that the stability and the safety of the overall operation of the system are ensured, and the system has good abnormality tolerance and online self-recovery capacity.
4. The method for dynamic energy storage scheduling of a modularized photovoltaic system according to claim 1, wherein in the second step, the calculation expression of the nonlinear load prediction model is: Wherein, the Representing actual load power in the current period; Representing the current load change rate; The temperature sensitive factor is represented to reflect the change trend of the load along with the temperature; Representing the current hours for simulating circadian behavior; The system self-tuning coefficients are dynamically adjusted according to the historical accuracy.
5. The method for dynamic energy storage scheduling of a modularized photovoltaic system according to claim 4, wherein the nonlinear prediction model for predicting future load power of each module in the central coordination unit has an initializing self-learning, rolling update and multi-cycle comparison mechanism in actual operation, and specifically comprises the following steps: When the system is operated for the first time or a module is added and accessed newly, the load prediction model constructs an initial prediction template based on time-by-time load power data in three historical days, and simultaneously, the illumination level, the ambient temperature and the actual electricity price of a corresponding time period are automatically associated to form an initial characteristic factor combination, and module-level electricity utilization behavior characteristics are fitted accordingly; When the model finishes one prediction period, triggering a parameter correction process, comparing the deviation between the predicted value and the actual load power of the period, and if the deviation exceeds a set threshold, automatically adjusting a weight factor and recalibrating part of the predicted factors; The system sets a multi-period comparison mechanism, dynamically aligns the predicted track of the current day with the actual load curve every 24 hours of operation period, calculates the average daily relative error value and the trend deviation value, and uses the average daily relative error value and the trend deviation value as the reference for updating the initialization template of the next day so as to form a predicted deviation correction closed loop taking the day as a unit.
6. The method for dynamic energy storage scheduling of a modular photovoltaic system according to claim 1, wherein in the third step, the analysis and calculation expression according to the central coordination unit is: Wherein, the Representing the total number of modules; representing the photovoltaic power generation power of the ith module at time t; representing the local load power of the ith module at time t; a critical threshold value determined by the combination of the system capacity and the maximum energy storage access power, when the sum of the total photovoltaic power generation amount minus the total local load amount of all modules in the system is greater than the threshold value At this point, the system is illustrated as having entered the photovoltaic power impingement zone.
7. The method for dynamic energy storage scheduling of a modularized photovoltaic system according to claim 6, wherein the determining mechanism for determining whether the system enters the photovoltaic power impact zone in the third step combines layered determination with a dynamic threshold, and specifically comprises: The preliminary judgment standard is that the total power of the module surplus photovoltaic exceeds 75% of the energy storage access capacity; Dynamic threshold Real-time adjustment is performed according to the following three factors: (1) The maximum charging power of all the energy storage units is currently achieved; (2) Predicting the ratio of load peak value to average value; (3) Grid connection points allow power access values and electricity price excitation coefficients; The sum of excess power exceeds a dynamic threshold, which is deemed to be the entry into the "photovoltaic power impingement zone".
8. The method for dynamic energy storage scheduling of a modularized photovoltaic system according to claim 1, wherein in the fifth step, when a scheduling failure occurs in a high-weight module due to a sudden anomaly or a physical access bottleneck, the system immediately starts an inter-module power redundancy migration mechanism, and the mechanism comprises: Introducing a 'power receiving capability real-time query interface', and uploading the maximum acceptable power value per module per cycle When the system detects that the original target module is not available, the system is used for controlling the current modules The value is subjected to redundant power redistribution; to the original redundant power Divided into a plurality of sub-power units The slave module can safely receive the power allowance and issue the power allowance in batches; If the total receiving capacity of all modules in the current scheduling period is still smaller than The system starts a power abstract scheduling cache pool, temporarily records the residual redundant power, and performs priority allocation on the newly added adjustable module when starting the next scheduling period; if multiple modules make an acceptance request for the same redundant power in the synchronous time window, the coordination unit orders the conflict modules according to the latest dispatching participation frequency of the modules, the energy storage health factor and the response speed.

Description

Dynamic energy storage scheduling method of modularized photovoltaic system Technical Field The invention relates to the technical field of operation control of photovoltaic power generation systems, in particular to a dynamic energy storage scheduling method of a modularized photovoltaic system. Background With the great development of global energy structure transformation and renewable energy sources, photovoltaic power generation is widely applied to scenes such as industrial parks, commercial buildings, residential communities and the like as a clean and renewable energy source form. Especially in the industrial park that possesses great roof resource, through deploying distributed photovoltaic system, realize generating electricity on spot, the on-the-spot consumption, can effectively reduce the energy cost of using, promote energy utilization efficiency. However, photovoltaic power generation has intermittent and fluctuating characteristics, and is susceptible to weather and time variations. In order to improve the stability and the self-utilization rate of the system, an energy storage system is generally configured in a matched manner. Currently, modular photovoltaic and energy storage designs are widely adopted in industrial parks or micro-grids, i.e. the system is divided into a plurality of functional modules or areas, and each module is provided with an independent photovoltaic module and an independent energy storage unit. This modular design facilitates system expansion and maintenance, but also presents new challenges in energy storage scheduling. In the prior art, most photovoltaic + energy storage systems still rely on static or semi-static scheduling strategies, such as based on timed charge-discharge rules, or simple "peak clipping and valley filling" logic. The method can not be flexibly adjusted according to actual running conditions, so that the problems of low running efficiency of the system, uneven energy storage utilization rate, inaccurate load prediction, delayed scheduling response and the like are caused. Disclosure of Invention The invention aims to provide a dynamic energy storage scheduling method of a modularized photovoltaic system, which aims to solve the problems in the background technology. In order to solve the technical problems, the invention provides the following technical scheme that the dynamic energy storage scheduling method of the modularized photovoltaic system comprises the following steps: Dividing a photovoltaic power generation system into a plurality of modules with autonomous operation capability, wherein each module comprises a photovoltaic power generation unit, an energy storage unit and a local control unit, and the local control unit is used for collecting current illumination intensity, battery charge state, temperature, power grid power flow direction and load power data of the module in real time and uploading the parameters to a central coordination unit of the system periodically; Step two, the central coordination unit builds a nonlinear load prediction model based on module-level load rate change, external temperature sensitivity and day-night periodic characteristics so as to improve the response capability to temporary sudden load change; step three, when detecting that the modules exceeding the threshold number have excessive photovoltaic power generation power and have lower load in a certain time window, the central coordination unit judges whether the system enters a photovoltaic power impact area according to analysis and calculation results; Step four, after judging that the photovoltaic power impact zone is entered, performing priority scheduling on each energy storage unit by adopting an energy storage weight function constructed by a pressing and exciting composite function; Step five, the system sorts according to the energy storage weight function values of all the modules, and under the condition that the SoC limit, the battery temperature rise constraint and the local load demand are met, the high-weight module is preferentially scheduled to receive the surplus power for charging; And step six, when the system detects that the load is abnormally increased and the load predicted value is greatly higher than the current power generation value or the power grid gives a demand control instruction, the central coordination unit starts an emergency discharge scheduling flow to schedule the energy storage unit with high SoC, proper temperature and long service life to participate in discharge support preferentially. According to the above technical solution, in the first step, the local control unit periodically collects key operation state parameters of the module, and constructs a standardized data structure for unified uploading, which specifically includes: The local control unit performs high-frequency sampling on parameters of photovoltaic output power, real-time voltage and current of the energy storage unit, temperature of