Park source network load storage integrated power grid system and source network load storage management and control method thereof

Abstract

The present disclosure provides a park source network load storage integrated power grid system and a source network load storage management method thereof. The method of the present disclosure comprises: online scheduling, which comprises: when the net load prediction error does not exceed the rated power of the energy storage system, adjusting the charge and discharge power of the energy storage system to make the energy storage system fully compensate for the net load prediction error; when the net load prediction error exceeds the rated power of the energy storage system, adjusting the charge and discharge power of the energy storage system by a predetermined charging compensation coefficient and a discharging compensation coefficient to make the energy storage system partially compensate for the net load prediction error. The present disclosure realizes the compensation of the energy storage system for the net load prediction error in a combination of full compensation and partial compensation, can balance the SOC of the energy storage system on the premise of ensuring the prediction qualification rate, avoids the severe charge and discharge of the energy storage system, effectively prolongs the battery life of the energy storage system, and thus improves the stability of the park source network load storage integrated power grid system.

Description

Technical Field

[0001] This disclosure relates to the field of distributed power generation technology in industrial parks, and in particular to an integrated power grid system for power generation, grid, load and storage in industrial parks and its power generation, grid, load and storage management and control method. Background Technology

[0002] By constructing an integrated power grid system for source, grid, load, and storage in the park and connecting distributed power systems such as wind turbines and photovoltaic power generation equipment, the penetration rate of distributed power sources can be increased. It can also control the exchange power between the park's power grid common connection point and the public power grid, providing an important technical means for the public power grid to achieve load dispatching. At the same time, it realizes efficient management of the power grid and rational utilization of electricity, reducing the park's electricity costs.

[0003] The capacity and operational status of energy storage systems are crucial to the stable operation of the integrated power grid system of the industrial park. Currently, by fully compensating for the net load prediction error (i.e., the power prediction error of the superimposed load from new energy sources) through the charging and discharging of energy storage systems, the uneven distribution of the net load prediction error will make it difficult for the energy storage system to achieve a stable state of charge (SOC) over a long time scale, frequently triggering the power limit of the energy storage system, accelerating the battery degradation of the energy storage system, and affecting the operational stability of the integrated power grid system of the industrial park. Summary of the Invention

[0004] In view of this, this disclosure provides an integrated power grid system for source, grid, load and storage in a park and a method for source, grid, load and storage management.

[0005] According to a first aspect of this disclosure, a source-grid-load-storage management method is provided, the method being applied to an integrated source-grid-load-storage power grid system in a park, the integrated source-grid-load-storage power grid system comprising: wind turbine generators, photovoltaic power generation equipment, energy storage systems, a source-grid-load-storage control system, and loads; the method comprising: The actual values ​​of active power of the park load, actual output of wind turbine generators, and actual output of photovoltaic power generation equipment were collected. The net load prediction error is determined based on the pre-obtained predicted values ​​of active power, wind power output, and photovoltaic power output for the current scheduling cycle, as well as the actual values ​​of active power, wind turbine output, and photovoltaic power output for the park load, wind turbine output, and photovoltaic power generation equipment output. Online scheduling is performed based on the net load prediction error and the previously obtained unit operation plan for the current scheduling period. The online scheduling includes: when the net load prediction error does not exceed the rated power of the energy storage system, adjusting the charging and discharging power of the energy storage system to fully compensate for the net load prediction error; when the net load prediction error exceeds the rated power of the energy storage system, adjusting the charging and discharging power of the energy storage system through predetermined charging and discharging compensation coefficients to partially compensate for the net load prediction error.

[0006] In some embodiments of the first aspect of this disclosure, the method further includes: using a machine learning model or a time series model to obtain the predicted values ​​of active power, wind power output, photovoltaic power output, and power purchase and sale price vectors for the current scheduling period based on weather data, historical operating sequence data, and electricity market transaction information.

[0007] In some embodiments of the first aspect of this disclosure, the method further includes: using a mixed integer programming model to perform offline scheduling based on the predicted active power of the park load, the predicted wind power output, the predicted photovoltaic power output, and the predicted electricity purchase and sale price vector for the current scheduling period, so as to obtain the unit operation plan for the current scheduling period. The unit operation plan includes the start-up and shutdown plan and output plan of the controllable wind turbine generator, the charging and discharging plan of the energy storage system, and the electricity purchase and sale plan.

[0008] In some embodiments of the first aspect of this disclosure, the constraints of the mixed integer programming model include: the sum of the planned discharge power and the reserved discharge power of the energy storage system shall not exceed the rated discharge power of the energy storage system, wherein the reserved discharge power is determined according to the discharge compensation coefficient; and the sum of the planned charging power and the reserved charging power of the energy storage system shall not exceed the rated charging power of the energy storage system, wherein the reserved charging power is determined according to the charging compensation coefficient.

[0009] In some embodiments of the first aspect of this disclosure, the constraints of the mixed integer programming model include: the planned SOC trajectory of the energy storage system operates within a compressed feasible interval, wherein the upper limit of the feasible interval is the original upper limit of the energy storage system's SOC compressed downward by an upward offset safety margin, and the lower limit of the feasible interval is the original lower limit of the energy storage system's SOC increased upward by a downward offset safety margin, wherein the upward offset safety margin is determined according to the charging compensation coefficient, and the downward offset safety margin is determined according to the discharging compensation coefficient.

[0010] In some embodiments of the first aspect of this disclosure, adjusting the charging and discharging power of the energy storage system by pre-determined charging and discharging compensation coefficients allows the energy storage system to partially compensate for the net load prediction error, including: Adjust the charging and discharging power of the energy storage system according to the following formula:

[0011] in, This represents the charging and discharging power of the energy storage system at the i-th sampling point within the current scheduling cycle. This represents the net load prediction error for the i-th sampling point within the current scheduling period. This indicates the rated charging power of the energy storage system. This indicates the rated discharge power of the energy storage system. This represents the charging compensation coefficient. This represents the discharge compensation coefficient.

[0012] In some embodiments of the first aspect of this disclosure, the method further includes: The probability density function of the net load prediction error is obtained by fitting the t-location scale distribution based on historical prediction error data. Solve for the discharge compensation coefficient and charging compensation coefficient that minimize the absolute value of the expected charge and discharge power of the energy storage system or make the absolute value of the expected charge and discharge power of the energy storage system less than a preset allowable value. The absolute value of the expected charge and discharge power of the energy storage system is expressed by the following formula:

[0013] in, This represents the absolute value of the expected charge and discharge power of the energy storage system. This represents the charging compensation coefficient. Indicates the discharge compensation coefficient. The probability density function representing the net load forecast error. This indicates the pre-set discharge power threshold of the energy storage system. This indicates the charging power threshold of the energy storage system. This represents the lower limit of the pre-set dead zone threshold for the energy storage system. This indicates the upper limit of the dead zone threshold of the energy storage system as preset.

[0014] In some embodiments of the first aspect of this disclosure, the probability density function of the net load forecasting error is expressed as follows:

[0015]

[0016] in, The probability density function representing the net load forecast error. This indicates the net load forecast error. This represents the lower limit of the pre-set dead zone threshold for the energy storage system. This indicates the pre-set upper limit of the dead zone threshold for the energy storage system. Indicates position parameters, Indicates the scale parameter. Represents the degree of freedom parameter. This represents the gamma function.

[0017] According to a second aspect of this disclosure, an integrated power grid system for industrial parks is provided, comprising: wind turbine generators, photovoltaic power generation equipment, energy storage system, a power grid control system, and loads. The power grid control system is used to perform the above-described method to compensate for net load forecasting errors caused by the loads, wind turbine generators, and photovoltaic power generation equipment through the energy storage system.

[0018] In some embodiments of the second aspect of this disclosure, the source-grid-load-storage control system includes a station control layer, a general coordination control layer, and a local control layer.

[0019] This embodiment of the invention achieves compensation for net load prediction errors in the energy storage system by combining full compensation and partial compensation. This effectively avoids the problem of uneven distribution of net load prediction errors leading to difficulty in balancing the State of Charge (SOC) over long time, frequently triggering the power limit of the energy storage system and thus accelerating the degradation of the energy storage battery. When the net load prediction error exceeds the rated power of the energy storage system, charging compensation coefficients and discharging compensation coefficients are introduced to adjust the degree of partial compensation for the net load prediction error, maintaining the throughput balance of the energy storage system. While ensuring the prediction qualification rate, the SOC balance of the energy storage system is taken into account, avoiding violent charging and discharging of the energy storage system, effectively extending the battery life of the energy storage system, and thus improving the stability of the integrated power grid system of source, grid, load and storage in the park. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this disclosure or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of the integrated power grid system for source, grid, load and storage provided in an embodiment of the present disclosure; Figure 2 This is an example architecture diagram of the source-grid-load-storage control system involved in the embodiments of this disclosure; Figure 3 This is a diagram illustrating the deployment of the station control layer in the source-grid-load-storage control system according to an embodiment of this disclosure. Figure 4 A flowchart illustrating the source-grid-load-storage management method for an integrated power grid system in a park provided in this embodiment of the disclosure; Figure 5 Another flowchart illustrating the source-grid-load-storage management method for the integrated power grid system of the park provided in this embodiment of the disclosure; Figure 6 This is an example diagram showing the t-location scale distribution of net load prediction error in an embodiment of this disclosure. Detailed Implementation

[0022] The technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0023] The terminology used in the embodiments of this disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of this disclosure. The singular forms “a,” “the,” and “the” as used in the embodiments of this disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0024] Depending on the context, words such as "if," "when," etc., used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrases "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0025] Figure 1 A schematic diagram of the architecture of an integrated power grid system for source-grid-load-storage in a park, to which embodiments of this disclosure apply, is shown. See also: Figure 1 The integrated power grid system for source-grid-load-storage in the park, applicable to the embodiments of this disclosure, includes wind turbine generators, photovoltaic power generation equipment, energy storage systems, source-grid-load-storage control systems, and loads. The wind turbine power station, photovoltaic power station, loads, and energy storage systems are respectively connected to the source-grid-load-storage control system, while the source-grid-load-storage control system is externally connected to the public power grid.

[0026] The integrated power grid system of the industrial park is connected to the public power grid through a common connection point. The integrated power grid system is subject to the control and management of the public power grid and performs internal adjustments to ensure that the power exchange at the common connection point meets the control requirements of the public power grid. Under normal circumstances, the integrated power grid system is connected to the main public power grid and operates in a grid-connected manner.

[0027] Wind turbines convert wind energy into alternating current (AC), while photovoltaic (PV) power generation equipment converts solar energy into direct current (DC). Energy storage systems, including but not limited to batteries, store and release electrical energy, playing a role in peak shaving and valley filling, and mitigating fluctuations. The public power grid serves as a backup power source or outlet for absorbing excess electricity within the park's integrated power grid system, maintaining the stability of the system's frequency and voltage. Loads can include terminal electrical equipment responsible for consuming electrical energy. The integrated power grid control system collects operational data from various devices, performs comprehensive analysis and decision-making, issues control commands, and achieves optimized energy scheduling and safe, stable system operation.

[0028] See Figure 1 The wind turbine power station, photovoltaic power station, load, and energy storage system are each electrically connected to the power grid-grid-load-storage control system through their own control units. The power grid-grid-load-storage control system is also electrically connected to the public power grid through control units. The control units installed between the wind turbine power station, photovoltaic power station, energy storage system, and the power grid-grid-load-storage control system are used to receive control commands from the control system and execute specific switching and regulation actions.

[0029] See Figure 1 The wind turbine power station, photovoltaic power station, energy storage system, and source-grid-load-storage control system are all connected by power electronic converters. These converters, including AC / DC and DC / AC converters, are responsible for converting between different forms of electrical energy (i.e., alternating current / direct current) to ensure matching of voltage, frequency, and other electrical parameters across all components. Specifically, the AC power generated by the wind turbine is converted to DC by AC / DC converter, then back to AC by DC / AC converter, and finally connected to the main circuit, entering the source-grid-load-storage control system via the control unit. The DC power generated by the photovoltaic power generation equipment is converted to AC by DC / AC converter, connected to the main circuit, and then entered the source-grid-load-storage control system via the control unit. The DC power from the energy storage system is converted to AC by DC / AC converter and connected to the main circuit, allowing for bidirectional energy flow—charging and discharging. The AC output terminals of the wind turbine, photovoltaic power generation equipment, and energy storage system are connected in parallel to form an AC bus. This AC bus is connected to the public power grid through an interface for bidirectional energy exchange—drawing power from or sending power to the public grid. This AC bus also connects downstream to the loads, providing them with electrical energy.

[0030] See Figure 1The wind turbine generators, photovoltaic power generation equipment, energy storage systems, the public power grid, and the loads are all connected to the power generation, grid, load, and energy storage control system through their own control units. The control units at each end of the wind turbine generators, photovoltaic power generation equipment, and energy storage systems upload operating status data such as power generation, battery state of charge (SOC), grid frequency, and load size to the power generation, grid, load, and energy storage control system via the communication network. The power generation, grid, load, and energy storage control system then sends dispatch commands back to each control unit via the communication network to control the operating status of each power electronic converter and the on / off state of the grid connection switch.

[0031] The integrated power grid system for source, grid, load and storage in the park according to the present disclosure has access to renewable energy sources such as wind turbines and photovoltaic power generation equipment. The output of these renewable energy sources is uncontrollable and fluctuates. The present disclosure solves the problem of comprehensive utilization of various energy sources by integrating and utilizing various distributed resources, through source-grid-storage-load coordination and energy interconnection, and realizes unified scheduling of power sources, energy storage and various loads within the park.

[0032] In this embodiment of the disclosure, the source-grid-load-storage control system can realize intelligent energy management, intelligent regulation, and intelligent matching within the park, and achieve distributed wind-solar complementarity, automatic energy storage regulation, and intelligent matching of load and power generation.

[0033] Figure 2 A sample architecture diagram of a source-grid-load-storage control system is shown. See also... Figure 2 The source-grid-load-storage control system includes the station control layer (also known as the dispatch master station), the overall coordination control layer, and the local control layer.

[0034] See Figure 2 The station control layer may include, but is not limited to, hardware devices such as data acquisition servers with GPS clocks, application servers, historical servers, workstations, central switches, gateways, connections to the grid regulation end, web servers, and network isolation devices, as well as software such as Supervisory Control and Data Acquisition (SCADA), energy management, and energy trading systems, used to realize grid monitoring, dispatching, optimization, and energy trading of the integrated power grid system of the park's power generation, grid, load, and storage. Workstations may include, but are not limited to, printing workstations and operator workstations.

[0035] The station control layer may have one or more of the following functions: data acquisition and monitoring control, energy optimization, electricity metering and billing, energy consumption analysis, energy trading, and production management.

[0036] In practical applications, the configuration scheme of the station control layer can be determined according to the "Overall Plan for Security Protection of Secondary Power Systems" (Electric Power Supervision Safety

[2006] No. 34) and the basic principles of "security zoning, dedicated network, horizontal isolation, and vertical authentication". For example, the station control layer can be divided into three security zones: control zone, non-control zone, and management information zone, with a chain-like topology. SCADA and energy management belong to the control zone, while the hardware and software responsible for energy trading and electricity metering belong to the non-control zone. The hardware and software responsible for production management and energy services belong to the management information zone. The management information zone receives information from the non-control zone through a forward security isolation device and transmits messages to the non-control zone through a reverse security isolation device. The control zone and the non-control zone can interact through logical isolation.

[0037] The operational systems within the control area are a crucial link in power production, directly enabling real-time monitoring of the primary power system. Vertically, they utilize the power dispatch data network or dedicated channels, and are a key focus and core of security protection. These operational systems include power data acquisition and monitoring systems and energy management systems, primarily used by dispatchers and operators. Data transmission real-time is at the millisecond or second level, and data communication utilizes the real-time subnet of the power dispatch data network or dedicated channels.

[0038] Operating systems outside the controlled area are essential components of power production. They operate online but lack control functions, utilize the power dispatch data network, and are closely linked to the operating systems within the controlled area. These operating systems include relay energy metering systems and energy trading systems.

[0039] The management information zone refers to the collection of business systems outside the production control zone. Business systems outside the control zone include production management systems, integrated energy service systems, etc.

[0040] For example, the software and hardware configuration deployment of each area of ​​the station control layer is shown in Table 1. Figure 3 A schematic diagram of the station control layer structure formed by deploying the configuration shown in Table 1 is presented.

[0041] Table 1

[0042] See Figure 2 The overall coordination and control layer may include, but is not limited to, the overall coordination controller and the overall protection device, which are used to complete the system-level real-time power balance and centralized fault protection of the source, grid, load and storage in the park.

[0043] The overall coordination and control layer ensures the safe and stable operation of the integrated power grid system within the park by controlling renewable energy, energy storage systems, loads, and connection points with the upper-level power grid PCC. The functions of the overall coordination and control layer may include one or more of the following: 1) Detecting and isolating incoming line faults in the integrated power grid; 2) Establishing electrical connections between the integrated power grid and the public power grid, disconnecting them after a fault to create an isolated grid operation state; 3) Collecting data from all electrical circuits of the integrated power grid, performing real-time calculations and logical judgments based on the grid's operating conditions, formulating control strategies for stable operation, and implementing real-time control to ensure the safe and stable operation of the integrated power grid system.

[0044] For example, the overall coordination control layer may include, but is not limited to, a central coordination controller for the integrated power grid (PGB) of the park, PCC point switches, regional protection devices for the integrated power grid (PGB), and mode controllers for the integrated power grid (PGB), to achieve real-time power balance and centralized fault protection at the system level. For instance, at the PCC interface, a PCC switch is set as the interface boundary between the integrated power grid system and the main grid. A dedicated integrated power grid coordination control cabinet for the integrated power grid (PGB) is installed at the substation. The central coordination control cabinet includes relay protection, measurement and control instruments, synchronization and grid connection devices, the integrated power grid coordination controller, communication processing devices, and other equipment.

[0045] The overall coordination controller is equipped with a dedicated logic control software package to realize real-time mode control of the integrated power grid system of the park, and real-time dynamic adjustment of power supply and load, so as to ensure the safe and stable operation of the integrated power grid system of the park. The local control layer is responsible for regulating the distributed generation power (DG) of the integrated power grid (source, grid, load, and storage) within the park, controlling the charging and discharging of energy storage, and managing the load. The local control layer includes local protection equipment and local controllers. It performs primary regulation of frequency and voltage by distributed generation and provides fault protection for the integrated power grid system. Through the coordination of local control and protection, it enables rapid self-healing of faults in the integrated power grid.

[0046] See Figure 2The local control layer may include, but is not limited to, local protection devices, distributed local controllers, energy storage converters, feeder acquisition devices, etc., and is responsible for completing the local protection and control of the equipment. Its main functions include data acquisition and protection, grid-connected inverter control of new energy power sources, maximum power point tracking (MPPT) control, and power conversion system (PCS) working mode switching control.

[0047] For example, the data acquisition hardware of the local control layer may include, but is not limited to, various terminal equipment such as intelligent data acquisition devices, remote terminals for switches, remote terminals for distribution transformers, remote terminals for switching stations / substations / ring network cabinets, fault acquisition terminals, and metering devices.

[0048] For example, the local protection hardware configuration of the local control layer may include: configuring a relay protection device with islanded and grid-connected mode switching function on the feeder. When the PCC switch of the integrated power grid system of the park is in the open position, the integrated power grid system of the park is disconnected from the public power grid, and the relay protection device receives the islanded mode signal and operates in islanded mode. When the PCC switch of the integrated power grid system of the park is in the closed position, the integrated power grid system of the park is connected to the public power grid, and the relay protection device operates in grid-connected mode, cooperating with the integrated power grid system of the park to monitor the operating status of the equipment and ensure the stable operation of the system.

[0049] For example, the local controller of the local control layer may include, but is not limited to, energy storage converters, photovoltaic inverters, wind turbine inverters, etc. The energy storage converter enables power point tracking (PPT) under normal operating conditions of the integrated power grid system (PGSHS) within the industrial park, managing battery charging and discharging, and facilitating power exchange between the designated PPSHS system and the public grid. The energy storage converter can also serve as a standard source in islanded operation. When the PPSHS system transitions from grid-connected to islanded operation, the support system provides reference voltage and frequency, achieving balanced power distribution among power sources, ensuring stable islanded system operation, providing resistance to short-term surges, smoothing power supply, storing energy, and peak shaving / valley filling. The energy storage converter supports uninterrupted transition from grid-connected to islanded mode. When the PPSHS system transitions from islanded to grid-connected, the energy storage converter can track the main grid voltage and frequency, seamlessly integrating the PPSHS system into the public grid.

[0050] The control strategies of energy storage converters include constant power control, droop control, and constant voltage and frequency control. They can be used to adjust the control strategies in real time according to the operating status of the integrated power grid system of the park, so as to ensure the safe and stable operation of the integrated power grid system of the park.

[0051] The equipment in the local control layer can be converted to Ethernet communication mode by a protocol conversion device and then connected to the switch. The switch is then connected to the centralized control layer via Ethernet. The unified DLT860-61850 communication protocol is used to achieve communication with the host computer, so that the entire park's integrated power grid system can achieve data interaction based on the IEC61850 standard.

[0052] In practical applications, the integrated power grid system for energy generation, grid, load, and storage in industrial parks adopts an optimized operation mode of "global optimization and regional autonomy." New energy sources achieve hierarchical and local balancing within the park, reducing energy transmission losses. The integrated power grid system can employ flexible load management strategies. For example, based on the nature of the loads carried by the integrated power grid system, they can be classified and categorized as interruptible, transferable, and critical loads, and different types of load management strategies can be considered in the park's power grid dispatch model.

[0053] The spatiotemporal distribution and dynamic characteristics of wind turbines and photovoltaic power generation equipment are correlated and complementary to a certain extent. Through continuous complementarity, the shortcomings of a single renewable energy source being susceptible to factors such as region, environment, and weather can be compensated. Furthermore, the correlation and complementarity can be used to overcome the inherent randomness and volatility of a single new energy source, thereby effectively improving the utilization efficiency of renewable energy, reducing spinning reserve, and enhancing the system's self-regulation capability.

[0054] In the integrated power grid system of the industrial park, power sources and loads such as wind turbines, photovoltaic power generation equipment, and energy storage systems can all participate in power supply and demand balance control as dispatchable resources. Both the power source side and the load side can be dispatchable resources. For example, by quickly and accurately controlling the interruptible loads of electricity customers, the emergency response time can be reduced from minutes to milliseconds. Interruptible loads are selected by electricity customers themselves, such as air conditioning and some lighting. After being disconnected, the power supply is guaranteed to be unaffected, maximizing the protection of enterprise production capacity and equipment safety. Through the friendly interaction between demand-side response and the active distribution system, peak shaving and valley filling of electricity and smart electricity use are achieved.

[0055] In the integrated power grid system of the park, advanced management and control technologies are used to optimize the combination of distributed and centralized energy supplies, highlight the complementarity and coordination between different combinations, give full play to the buffering role of the park's power grid technology, reduce the adverse effects of accepting new energy power on the safe and stable operation of the power grid, and improve the park's power grid's ability to accept diversified power sources.

[0056] In the integrated power grid system of the park, various energy conversion technologies and information flow and energy flow interaction technologies can be used to realize the development and utilization of energy resources and the coordination between resource transportation networks and energy transmission networks. The system can unify the various energy needs of users into a whole, further expand the power demand-side management into comprehensive energy management in the entire energy field, and further amplify the role of broad demand-side resources in promoting the consumption of clean energy and ensuring the safe and stable operation of the system.

[0057] In the integrated power grid system of the industrial park, the source-grid-load-storage control system can perform integrated management and control based on information such as the predicted values ​​of new energy sources and loads, market electricity prices, dispatch rules, load classification, and equipment characteristics. Since the basic premise of power grid operation is the balance between power generation and load, the output power and absorbed power of the energy storage system directly affect the stability of the integrated power grid system. The energy stored in the energy storage system is finite; when the energy storage system continuously charges and discharges to its SOC (State of Charge) limits, it can no longer operate in a charging and discharging state. Therefore, the energy storage capacity of the energy storage system is crucial to the stable operation of the integrated power grid system. In this embodiment, the source-grid-load-storage control system can be used to execute the following source-grid-load-storage management and control method to compensate for the net load prediction errors caused by loads, wind turbine generators, and photovoltaic power generation equipment, thereby maintaining the balance of the energy storage system's throughput, avoiding drastic charging and discharging of the energy storage system, extending the battery life of the energy storage system, and improving the overall stability of the integrated power grid system.

[0058] Figure 4 This diagram illustrates a flow chart of the power generation, grid, load, and energy storage (PGD) management method provided in this embodiment. This method can be executed by the PDD control system within the aforementioned integrated PDD power grid system in the industrial park. The PDD management method provided in this embodiment involves the station control layer in the PDD control system performing decision-making and coordination, while the overall coordination control layer and the local control layer are responsible for instruction decomposition and physical execution—a three-layer collaborative process. Specifically, the core decision-making logic operates at the station control layer, real-time power coordination operates at the overall coordination control layer, and data acquisition and instruction response are completed by the local control layer.

[0059] See Figure 4 The source-grid-load-storage management method of this disclosure includes the following steps 401 to 403: Step 401: Collect the actual values ​​of the active power of the park load, the actual values ​​of the output of the wind turbine generator set, and the actual values ​​of the output of the photovoltaic power generation equipment; Specifically, the source-grid-load-storage control system collects real-time data on the actual active power of the park's load, the actual output of wind turbine generators, and the actual output of photovoltaic power generation equipment. In some examples, the actual active power of the park's load may include the total active power of the park's total electrical load, which refers to the sum of the active power consumed in real time by all electrical terminals within the park, such as factory equipment, lighting, HVAC, and charging piles. If the park adopts zoned monitoring, the actual active power of the park's load may also include the actual load of a branch circuit of a certain production line or building.

[0060] In practical applications, local devices such as wind turbines, photovoltaic inverters, meters, and PCS in the local control layer of the power generation, grid, load, and energy storage control system can collect raw data such as current, voltage, power, and status through sensors. This data is then uploaded to the overall coordination control layer or station control layer via communication protocols such as Modbus and IEC 61850. The overall coordination control layer or station control layer uses this raw data to obtain the actual active power of the aforementioned park load, the actual output of wind turbine generators, and the actual output of photovoltaic power generation equipment. Alternatively, the SCADA system in the station control layer can centrally collect real-time active power and switch status information of the park to obtain the actual active power of the aforementioned park load, the actual output of wind turbine generators, and the actual output of photovoltaic power generation equipment.

[0061] Step 402: Determine the net load prediction error based on the pre-obtained predicted values ​​of active power, wind power output, and photovoltaic power output for the current scheduling cycle, as well as the actual values ​​of active power, wind turbine output, and photovoltaic power output for the park load, wind turbine output, and photovoltaic power generation equipment output. Specifically, the source-grid-load-storage control system subtracts the actual active power value of the park load collected in step 401 from the previously obtained predicted active power value of the park load for the current scheduling period to obtain the real-time load change; subtracts the actual output value of the wind turbine generators from the previously obtained predicted output value for the current scheduling period to obtain the real-time wind power output change; subtracts the actual output value of the photovoltaic power generation equipment from the previously obtained predicted output value for the current scheduling period to obtain the real-time photovoltaic output change; adds the real-time wind power output change to the real-time photovoltaic output change to obtain the renewable energy output change; and subtracts the renewable energy output change from the real-time load change to obtain the real-time net load prediction error. This net load prediction error must be compensated in a timely manner to maintain system power balance and frequency stability.

[0062] In other words, the net load forecast error can be calculated using the following formula:

[0063]

[0064]

[0065]

[0066] in, The discrete time series index is the sampling point number arranged in chronological order within the current scheduling period. A sampling point is a data point acquired at each discrete time interval (e.g., every 1 second, every 0.1 seconds). This discrete time interval can be 1 second, 0.1 seconds, or other values, and usually corresponds to the refresh rate of the data acquisition system (e.g., SCADA, PLC, etc.).

[0067] This represents the actual value of the active power of the park load at the i-th sampling point within the current scheduling cycle. It can be the active power value of the load collected in real time by SCADA or the park energy management system (EMS), and can be obtained through smart meters, RTUs, or load control terminals.

[0068] This represents the predicted active power value of the park load at the i-th sampling point within the current scheduling cycle, which is the corresponding point on the load prediction curve of the current scheduling cycle obtained previously.

[0069] This represents the real-time load change at the i-th sampling point within the current period. A positive value indicates that the actual load is higher than expected, while a negative value indicates that the actual load is lower than expected.

[0070] This represents the actual output value of the wind turbine generator at the i-th sampling point in the current period. It can be the real-time active power output of the wind turbine generator collected by the wind turbine monitoring system or SCADA.

[0071] This represents the predicted output value of the wind turbine at the i-th sampling point in the current period, which is the corresponding point on the expected output curve of the wind turbine in the current scheduling period obtained previously.

[0072] This represents the real-time wind power output change at the i-th sampling point within the current period. A positive value indicates that the wind conditions are better than predicted, while a negative value indicates that the power generation is insufficient.

[0073] This represents the actual output value of the photovoltaic power generation equipment at the i-th sampling point in the current period. It can be the real-time power of the photovoltaic power generation equipment collected from the photovoltaic inverter or SCADA.

[0074] This represents the predicted output value of the photovoltaic power generation equipment at the i-th sampling point in the current cycle, which is the corresponding point on the expected output curve of the photovoltaic power generation equipment in the current scheduling cycle obtained previously.

[0075] This represents the real-time change in photovoltaic power output at the i-th sampling point within the current period. A positive value indicates that the sunlight is stronger than expected, while a negative value indicates that it is weaker than expected.

[0076] This represents the net load prediction error at the i-th sampling point in the current period, i.e., the new energy superimposed load prediction error, which is the difference between the actual net load and the predicted net load. A value >0 indicates a positive value, meaning the actual net load is higher than the predicted net load, requiring increased power generation, such as starting up fast-moving generators or increasing energy storage discharge, or reduced power consumption, such as cutting off non-critical loads. A value <0 indicates that the actual net load is lower than the predicted net load, requiring a reduction in power generation, such as charging the energy storage system or reducing the unit output, or an increase in power consumption, such as activating adjustable loads.

[0077] In practical applications, the calculation of net load forecasting error can be performed by the station control layer or the overall coordination layer.

[0078] Step 403: Perform online scheduling based on the net load prediction error and the previously obtained unit operation plan for the current scheduling cycle. Online scheduling includes: when the net load prediction error does not exceed the rated power of the energy storage system, adjusting the charging and discharging power of the energy storage system to fully compensate for the net load prediction error; when the net load prediction error exceeds the rated power of the energy storage system, adjusting the charging and discharging power of the energy storage system through predetermined charging compensation coefficients and discharging compensation coefficients to partially compensate for the net load prediction error.

[0079] Specifically, based on the previously obtained unit operation plan for the current scheduling period (e.g., day-ahead plan), combined with the real-time net load forecast error obtained in step 402, and the current actual operating status of each device (e.g., unit start-up / shutdown status, unit output, energy storage unit charging / discharging, power purchased / sold, load shedding command, current SOC of the energy storage system, etc.), online scheduling is performed according to the pre-configured online scheduling strategy.

[0080] Online dispatch strategies are used to quickly decide how to allocate flexible resources to absorb net load forecasting errors. For example, online dispatch strategies may include: prioritizing the adjustment of the charging and discharging power of energy storage systems, adjusting the power purchased and sold from the external grid, and adjusting the output of controllable units (such as gas turbines); if this is still insufficient, demand-side response or load shedding commands are executed to ensure system safety.

[0081] Online dispatching adopts a stable control mode based on the comprehensive dynamic balance of load, energy storage, and power supply. It takes controllable power sources, energy storage, and flexible loads connected to the park's power grid as the management and control objects, takes the power exchange power of the grid's common connection point given by the main grid as the target, and takes power balance, voltage / frequency limit-free operation, equipment capacity limit, and energy storage SOC safety as safety constraints to optimize the operation mode of the park's power grid. Thus, under the hard boundary of the public grid's power exchange power not exceeding the limit, it achieves real-time local power balance and stable operation within the park by rapidly coordinating sources, grid, load, and storage, maintaining the reliable, stable, and economical operation of the park's power grid.

[0082] During online dispatching, the internal resources of the park are adjusted in real time according to the tie-line power tracking constraint to ensure that the exchange power with the public power grid does not exceed the limit of the day-ahead plan or dispatching agreement.

[0083] In online dispatching, the power adjustment of each controllable unit is dynamically allocated based on the real-time net load forecast error to maintain instantaneous power balance. Controllable units can include, but are not limited to: energy storage systems, dispatchable wind turbine generators, flexible loads, power exchange via interconnection lines with the external power grid, and renewable energy generation units. Flexible loads include, but are not limited to, interruptible loads that can be forcibly disconnected in a short time, such as some lighting, air conditioning, and non-critical production processes, as well as transferable loads that can be continuously / steppedly adjusted by adjusting the electricity consumption period or power setpoint, such as electric boilers, cold storage systems, and some charging piles. Renewable energy generation units include, but are not limited to, wind power inverters and photovoltaic inverters.

[0084] Optimizing the operation of the industrial park's power grid can include adjusting the charging and discharging power of the energy storage system to fully compensate for the net load forecasting error when the net load forecasting error does not exceed the rated power of the energy storage system; and adjusting the charging and discharging power of the energy storage system through predetermined charging and discharging compensation coefficients to partially compensate for the net load forecasting error when the net load forecasting error exceeds the rated power of the energy storage system. This allows for maximizing the absorption of wind and solar power output fluctuations through local regulation, reducing the impact on the public power grid, and ensuring system frequency and voltage stability.

[0085] Specifically, when the net load forecast error is within the rated power range of the energy storage system, i.e. At that time, This indicates the rated charging power of the energy storage system. This represents the rated discharge power of the energy storage system. The energy storage system fully compensates for the net load prediction error, and the grid-connected power (i.e., the power at the grid connection point) is equal to the predicted value.

[0086] That is, the charging and discharging power of the energy storage system is adjusted as follows:

[0087] in, This represents the charging and discharging power of the energy storage system at the i-th sampling point within the current scheduling cycle. Indicates energy storage system Specifically, when the net load forecast error is outside the rated power range of the energy storage system, i.e. Within, that is, outside the range If the net load prediction error is negative and its absolute value is greater than the rated charging power of the energy storage system, and if the net load prediction error is positive and greater than the rated discharging power of the energy storage system, the compensation power of the energy storage system is linearly reduced by multiples of the charging compensation coefficient and the discharging compensation coefficient. This allows the energy storage system to partially compensate for the net load prediction error, preventing the energy storage system from instantly reaching its power limit, while still partially mitigating the excess net load prediction error. In other words, the charging and discharging power of the energy storage system is adjusted according to the following formula:

[0088] in, This represents the charging and discharging power of the energy storage system at the i-th sampling point within the current scheduling cycle. This represents the net load prediction error for the i-th sampling point within the current scheduling period. This indicates the rated charging power of the energy storage system. This indicates the rated discharge power of the energy storage system. This represents the charging compensation coefficient. This represents the discharge compensation coefficient.

[0089] in, The charging compensation coefficient and the discharging compensation coefficient determine the extent to which the portion exceeding the rated power of the energy storage system is reduced. The larger the values ​​of the charging compensation coefficient and the discharging compensation coefficient, the less additional compensation the energy storage system bears.

[0090] When the net load forecast error exceeds the rated power of the energy storage system, if the full compensation method of the energy storage system is still used, the power of the energy storage system will be limited to the maximum charging and discharging power of the energy storage system. , If the net load prediction error exceeds the rated power of the energy storage system, the compensation ratio of the energy storage system to the net load prediction error is adjusted by the charging compensation coefficient and the discharging compensation coefficient. This avoids the energy storage system from instantly reaching its power limit, while still partially mitigating the excess error, thereby controlling the balance of the energy storage system's throughput and output. This effectively protects the energy storage system while smoothing grid-connected power fluctuations.

[0091] In practical applications, the online scheduling in this step can be executed collaboratively by the overall coordination control layer and the local control layer. For example, the overall coordination control layer can receive the online scheduling strategy rules parameters and the unit operation plan for the current scheduling cycle issued by the station control layer, quickly calculate the adjustment amount of each controllable unit (i.e., energy storage, controllable wind turbine generators, tie lines, loads, etc.) based on the real-time net load forecast error, and generate control commands (such as power adjustment commands, switching commands, etc.) and send them to the local control layer. The local control layer receives the control commands issued by the overall coordination control layer and completes power adjustment or switching actions at the millisecond to second level (such as energy storage charging and discharging adjustment, load shedding switch disconnection, etc.), thereby enabling the energy storage system to fully or partially compensate for the net load forecast error.

[0092] Further, after step 403, the online scheduling results of the current scheduling cycle can be saved as the initial values ​​for the next offline scheduling. In the offline scheduling of the next scheduling cycle, a snapshot of the actual physical state of the system at the end of the current scheduling cycle is read and used as the temporal boundary constraint for the mixed-integer programming model. Specifically, after the online scheduling of the current scheduling cycle ends, a snapshot of the actual physical state of the system at the end of the current scheduling cycle can be recorded. This snapshot includes, but is not limited to, the state of charge (SOC) of the energy storage system, unit operating status information, interruptible load status information, and the power exchange baseline with the external power grid. Here, the end time of online scheduling can be the last online control moment of the current scheduling cycle.

[0093] The State of Charge (SOC) of an energy storage system represents the percentage of its rated capacity remaining. Unit operating status information includes the start / stop status, output level, and continuous operation / stop duration of each controllable wind turbine. Interruptible load status information includes whether certain flexible loads have been disconnected and their duration of disconnection. The power exchange baseline with the external grid includes the current tie-line power and whether there are any unfulfilled power sales contracts.

[0094] Specifically, the SOC of the energy storage system, the start-up and shutdown status of the unit, the output level, and the status of interruptible load at the end of the online scheduling are written into the real-time database. These data will serve as the initial boundary conditions for the next offline scheduling (i.e., the unit operation plan for the next scheduling cycle), update the initial state of the equipment (such as the starting SOC of the energy storage for the next day), correct the deviation mode of the prediction model, and adjust the constraints or cost parameters to achieve rolling optimization and learning evolution.

[0095] In practical applications, after step 403 is completed, the EMS at the station control layer can determine whether the current scheduling cycle has ended and decide whether to start the next scheduling cycle. After the current scheduling cycle ends, the real-time / historical database server at the station control layer can write a snapshot of the actual physical state of the system at the end of the real-time phase into the historical database as the initial condition for the next scheduling cycle.

[0096] The method of this disclosure combines full compensation with partial compensation to compensate for net load prediction errors in the energy storage system. This effectively avoids the problem of uneven distribution of net load prediction errors leading to difficulty in balancing the State of Charge (SOC) over long time, frequently triggering the power limit of the energy storage system, and thus accelerating the degradation of the energy storage battery. When the net load prediction error exceeds the rated power of the energy storage system, charging compensation coefficients and discharging compensation coefficients are introduced to adjust the degree of partial compensation for the net load prediction error, maintaining the throughput balance of the energy storage system. While ensuring the prediction qualification rate, the method also takes into account the SOC balance of the energy storage system, avoids violent charging and discharging of the energy storage system, effectively extends the battery life of the energy storage system, and thus improves the stability of the integrated power grid system of source, grid, load and storage in the park.

[0097] Figure 5 Another schematic flowchart of the source-grid-load-storage management method provided in this disclosure embodiment is shown. See also Figure 5 The method of this disclosure embodiment may further include the following steps 404 and 405, both of which are performed before steps 401 to 403.

[0098] Step 404: Before the current scheduling cycle, based on weather data, historical operating sequence data and electricity market transaction information, use machine learning models or time series models to obtain the predicted values ​​of active power, wind power output, photovoltaic power output and electricity purchase and sale price vector for the current scheduling cycle. Load forecasting can include, but must first include, the following: based on historical electricity consumption data, weather, production plans, and other information, predicting the load forecast curves for all electricity-consuming terminals within the park, such as factories, buildings, and charging piles, for each time period of the current scheduling cycle. For example, the load forecast curves for every 15 minutes or hour within the current scheduling cycle. These load forecast curves describe the expected power consumption of electricity-consuming terminals such as factories, buildings, and charging piles for each time period within the current scheduling cycle. The load forecast curves include the predicted active power value for each moment of each time period within the current scheduling cycle. Weather information may include, but is not limited to, temperature, humidity, rainfall, wind speed, and radiation.

[0099] Furthermore, if there are multiple types of loads within the park, such as production lines, HVAC, lighting, and charging piles, corresponding load forecast curves can be provided for each type of load to indicate the predicted active power value of each type of load in each time period within the current scheduling week, which facilitates subsequent fine-tuning.

[0100] Specifically, the expected output curves of wind turbine generators in each period of the current scheduling cycle can be calculated based on numerical weather forecasts (e.g., wind speed, wind direction, etc.) for the current scheduling cycle. The expected output curves of wind turbine generators include the predicted output values ​​of wind turbine generators in each period of the current scheduling cycle.

[0101] Specifically, the expected output curves of photovoltaic power generation equipment in each period of the current scheduling cycle can be calculated based on meteorological forecast information such as irradiance, temperature, and cloud cover in the current scheduling cycle. The expected output curves of photovoltaic power generation equipment include the predicted output values ​​of photovoltaic power generation equipment in each period of the current scheduling cycle.

[0102] Specifically, the electricity price information of the current dispatch cycle in the external electricity market can be obtained to form a prediction vector of the purchase and sale price of electricity in the current dispatch cycle. This electricity price information of the current dispatch cycle is used to optimize the purchase and sale strategy of electricity in the current dispatch cycle. The electricity price information of the current dispatch cycle may include, but is not limited to, time-of-use electricity price, spot market prediction price, etc.

[0103] In practical applications, the predicted values ​​of active power output, wind power output, photovoltaic power output, and electricity purchase and sale price vector for the park load can be obtained using different models, which can be, but are not limited to, machine learning models or time series models. This disclosure does not impose any limitations on these aspects.

[0104] In practical applications, the station control layer can be used to obtain the predicted values ​​of active power, wind power output, photovoltaic output, and electricity purchase and sale price for the current scheduling period. For example, the station control layer's prediction server, historical database, and meteorological data interface can be used to run NWP parsing, load prediction models, new energy prediction models, and electricity price prediction models to generate a prediction sequence for the next 24 hours (i.e., the current scheduling period). This prediction sequence includes the predicted values ​​of active power, wind power output, photovoltaic output, and electricity purchase and sale price.

[0105] Step 405: Before the current scheduling cycle, based on the predicted active power, wind power output, photovoltaic power output, and electricity purchase and sale price vector of the park load for the current scheduling cycle, offline scheduling is performed using a mixed integer programming (MILP) model to obtain the unit operation plan for the current scheduling cycle. The unit operation plan may include the start-up and shutdown plan and output plan of the controllable wind turbine generators, the charging and discharging plan of the energy storage system, and the electricity purchase and sale plan.

[0106] The power purchase and sale plan includes: power purchase plan, power sale plan, net exchange power curve, and power purchase / sale cost / revenue estimation information. The power purchase plan indicates the active power purchased from the grid in each time period, the power sale plan indicates the surplus power sold to the grid in each time period, the net exchange power curve indicates the net value of power purchase minus power sale in each time period, reflecting the overall power interaction between the park and the main grid, and the power purchase / sale cost / revenue estimation information represents the expected transaction amount based on the predicted electricity price.

[0107] In practical applications, the station control layer can construct and solve the MILP model to output the globally optimal unit operation plan.

[0108] Among them, the mixed integer programming model can be used to quantify all aspects of the purchase and sale of electricity between wind turbine generators, photovoltaic power generation equipment, energy storage systems and the external power grid, as well as interruptible loads (load shedding), into decision variables. For example, when to charge and release energy storage, and whether to start the backup power station. Some of these variables are integers, such as start-up and shutdown status, while others are continuous values, such as output magnitude. The goal of offline scheduling is to minimize the total operating cost or maximize the revenue while satisfying load and safety constraints. The total operating cost includes the cost of purchasing electricity, the cost of starting and stopping the generators, energy storage losses, and load shedding penalties, while the revenue may include, but is not limited to, electricity sales revenue.

[0109] Offline scheduling provides a complete day-ahead operational plan, including the start-up and shutdown plans and output plans of controllable wind turbines, the charging and discharging plans of energy storage systems, power purchase and sale plans, and load shedding commands for each time period within the current scheduling cycle. Specifically, for small wind turbines (controllable units) that can be scheduled by the power grid-grid-load-storage control system, offline scheduling provides their start-up and shutdown plans and output plans. The start-up and shutdown plans of controllable units indicate whether they will be connected to the grid for each time period the following day, with values ​​of 0 or 1. The output plans of controllable units indicate the target output of the controllable units during the grid-connected period. The energy storage system charging and discharging plan includes: the charging and discharging status, charging and discharging power, and state of charge (SOC) trajectory for each time period the following day. The charging and discharging status indicates whether the energy storage system is operating in a charging, discharging, or idle state; the charging and discharging power indicates the precise power value for each time period; and the SOC trajectory is simply the SOC change curve from the current initial SOC through each charging and discharging period.

[0110] The objective function of the mixed-integer programming model comprehensively considers power supply operating costs, energy storage depreciation costs, environmental protection costs, load shedding costs, and network loss costs. Constraints simultaneously satisfy power balance, reserve requirements, ramp rate, upper and lower output limits, minimum start-up and shutdown times, upper and lower limits of energy storage capacity, upper and lower limits of energy storage charging and discharging power, and upper limits of power purchase and sale. Based on a comprehensive consideration of the above cost factors and constraints, a mixed-integer programming model for the park power grid was established. This model is applicable to various operating modes, including typical grid-connected economic operation mode, long-term stable islanded operation mode, and short-term islanded operation mode. During operation, the appropriate operating mode and optimization strategy can be selected as needed.

[0111] For example, the objective function of the mixed integer programming model comprehensively considers various cost items such as operating costs, energy storage depreciation, environmental protection, load shedding, and network losses. It adopts a weighted summation to transform multiple objectives into a single objective. The core objective is to minimize the total life cycle cost while ensuring the aforementioned constraints.

[0112] To ensure reliable power supply to the industrial park and avoid wasting renewable energy, it is necessary to reserve energy storage capacity to cope with real-time net load forecasting errors. This is to prevent the energy storage system from instantly reaching its power limit or capacity limit in the event of extreme net load forecasting errors, which could lead to power outages, curtailment of solar power, or significant backfeeding to the grid, resulting in penalties. Therefore, in this embodiment, the charging compensation coefficient and discharging compensation coefficient are used as key inputs in the mixed-integer programming model to provide constraints for the capacity configuration and charging / discharging strategy optimization of the energy storage system. This reserve capacity prevents the energy storage system from instantly reaching its power limit or capacity limit in the event of extreme net load forecasting errors.

[0113] In some implementations, the constraints of the mixed-integer programming model, in addition to those mentioned above such as power balance, reserve, ramp rate, output upper and lower limits, minimum start-up and shutdown time, energy storage capacity upper and lower limits, energy storage charging and discharging power upper and lower limits, and power purchase and sale upper limits, may also include: the sum of the planned discharge power and the reserved discharge power of the energy storage system shall not exceed the rated discharge power of the energy storage system, and the reserved discharge power is determined according to the discharge compensation coefficient; and the sum of the planned charging power and the reserved charging power of the energy storage system shall not exceed the rated charging power of the energy storage system, and the reserved charging power is determined according to the charging compensation coefficient. The reserved power ensures that, in most scenarios, additional discharge commands triggered by positive net load prediction errors in the real-time phase will not cause the energy storage power to exceed the limit, and the energy storage system still has sufficient capacity to execute additional charging commands triggered by negative net load prediction errors in the real-time phase.

[0114] Specifically, the aforementioned constraints concerning the reserve amount can be expressed as follows:

[0115]

[0116] This represents the planned discharge power of the energy storage system at the i-th sampling point. Indicates the discharge compensation coefficient. This represents the power reserve constant in the discharge direction. This indicates the rated discharge power of the energy storage system. This represents the binary variable indicating the discharge state of the energy storage system at the i-th sampling point. , indicating discharge, , indicating other. This represents the planned charging power of the energy storage system at the i-th sampling point. This represents the charging compensation coefficient. This represents the power reserve constant in the charging direction. This indicates the rated charging power of the energy storage system. A binary variable representing the charging status of energy storage. , indicating charging. , indicating other.

[0117] This can be calculated from the probability density function f(x) of the net load forecast error, as described below. Specifically, it is obtained using a historical net load forecast error sample sequence. Construct a sample sequence of two random variables and obtain empirical quantiles. in, This represents the total number of net load forecast error samples in the historical net load forecast error sample sequence. This indicates the sequence number of the historical net load forecast error sample.

[0118] Specifically, it can be obtained through the following formula. :

[0119]

[0120]

[0121]

[0122] in, This represents a sample sequence of historical net load forecast errors. The j-th historical net load forecast error sample in the data. The power threshold representing the net load forecast error for the full response of an energy storage system is typically taken as a percentage of or equal to the rated power of the energy storage system. This represents the total number of net load forecast error samples in the historical net load forecast error sample sequence. To indicate the confidence level, for example, we can take... The value is 0.95, which can be preset according to the engineering reliability requirements. The rounding up symbol, Represents a sequence The numbers after sorting in ascending order A series of ordinal statistics Represents a sequence The numbers after sorting in ascending order A series of ordinal statistics.

[0123] In some implementations, the constraints of the mixed-integer programming model may further include: the planned SOC trajectory of the energy storage system operates within a compressed feasible interval, wherein the upper limit of the feasible interval is the original upper limit of the energy storage system's SOC compressed downwards by an upward offset safety margin, and the lower limit of the feasible interval is the original lower limit of the energy storage system's SOC increased upwards by a downward offset safety margin, wherein the upward offset safety margin is determined based on the charging compensation coefficient, and the downward offset safety margin is determined based on the discharging compensation coefficient. This ensures that during continuous charging to compensate for negative net load prediction errors in the real-time phase, the SOC does not exceed the upper limit, and during continuous discharging to compensate for positive net load prediction errors, the SOC does not exceed the lower limit.

[0124] Specifically, the planned SOC trajectory of an energy storage system operating within the compressed feasible range can be expressed as follows:

[0125]

[0126]

[0127] in, This indicates the lower limit of the original SOC of the energy storage system, for example, 10% or 20%; This indicates the lower limit of the original SOC of the energy storage system, for example, 90% or 95%; Indicates the downward offset safety margin. This indicates an upward offset safety margin. This means that the constraint satisfies the formula for all time periods i (i.e., all sampling points i). This indicates the planned load point status of the energy storage system at the end of the scheduling period corresponding to the i-th sampling point within the current scheduling cycle.

[0128] in, This represents the charging compensation coefficient. Indicates the discharge compensation coefficient; This indicates the maximum number of time periods that consecutive extreme errors may last. It can be set based on historical error persistence statistics, and N can be set to 3~6. This indicates the power reserve in the charging direction. This indicates the power reserve in the discharge direction. This indicates the charging efficiency of the energy storage system, with a typical value of 0.90~0.95; This indicates the discharge efficiency of the energy storage system, with a typical value of 0.90~0.95. This indicates the rated capacity of the energy storage system. This indicates the duration of the time interval between adjacent sampling points (i.e., a single online scheduling period).

[0129] In the field of power system dispatching, a dispatching cycle is typically defined as 24 hours, or one day. However, in practical applications, the length of a single dispatching cycle can be flexibly adjusted as needed. In this embodiment, steps 401-403 of the online dispatching process are executed cyclically at minute intervals within a dispatching cycle. For example, if a dispatching cycle is set to one day, online dispatching can be performed every 5-15 minutes or other fixed intervals within a single operating day. Steps 404-405 of the prediction and offline dispatching processes are executed once every dispatching cycle. For example, if a dispatching cycle is set to one day, the prediction in step 404 and the offline dispatching in step 405 are executed once daily, with the specific execution time set between 22:00 and 24:00 the day before the operating day or at 00:00 on the operating day itself.

[0130] Furthermore, the method in this embodiment may further include: step 406, determining the discharge compensation coefficient and charging compensation coefficient of the energy storage system. Specifically, based on historical prediction error data, the probability density function of the net load prediction error is obtained by fitting a t-location-scale distribution (T-LS or T-TLS); the discharge compensation coefficient and charging compensation coefficient that minimize the absolute value of the expected charging and discharging power of the energy storage system, or the discharge compensation coefficient and charging compensation coefficient that make the absolute value of the expected charging and discharging power of the energy storage system less than a preset allowable value, are solved.

[0131] Specifically, the probability density function of the net load forecast error can be expressed as follows:

[0132]

[0133] in, The probability density function representing the net load forecast error. This indicates the net load forecast error. This represents the lower limit of the pre-set dead zone threshold for the energy storage system. This indicates the pre-set upper limit of the dead zone threshold for the energy storage system. Indicates position parameters, Indicates the scale parameter. Represents the degree of freedom parameter. This represents the gamma function.

[0134] in, <0< , It is the dead zone boundary allowed by the system. It can be set to the rated power range of the energy storage system, or it can be a range smaller than the rated power of the energy storage system.

[0135] , , All of these can be obtained by fitting historical forecast error data using statistical estimation methods such as maximum likelihood estimation. Specifically, historical forecast error data is collected; for example, net load forecast errors for each time period over the past 30 days (e.g., 15-minute intervals, 96 time periods per day) are collected to obtain historical forecast error data containing 2880 net load forecast errors. Using this historical forecast error data, a likelihood function for the t-location scale distribution is obtained, which contains... , , Automatically find a set of values ​​that maximizes the likelihood function. , , The value is the one used in the aforementioned probability density function. , , In practical applications, the aforementioned t-location scale distribution can also be updated periodically (e.g., daily or weekly) by refitting the updated net load prediction error dataset. , , The value of .

[0136] Figure 6 This diagram illustrates an example of the t-location scale distribution of net load forecast error. The horizontal axis represents the net load forecast error, measured in units of pu (pu). pu is the per-unit value of the net load forecast error relative to the total installed capacity of the park (or the rated capacity of the park's transformers, or the park's maximum load). 0 indicates that the net load forecast error perfectly matches the actual situation. The vertical axis represents the probability density, i.e., the probability of the net load forecast error occurring around a certain value. The red curve indicates the t-location scale distribution. This curve shows that the t-distribution has a sharp peak and thick tails, indicating that the system's net load forecast error is indeed very small most of the time, concentrated around 0. Figure 6As can be seen from the distribution example diagram, the extent to which the actual net power demand of the park will deviate from the predicted value at a certain point in the future directly determines how much power and capacity the energy storage system needs to be configured to smooth out such fluctuations.

[0137] The absolute value of the expected charge / discharge power of an energy storage system is the absolute value of the expected average net discharge / charge power over long-term operation. It is used to assess energy balance and capacity allocation requirements. The absolute value of the expected charge / discharge power directly reflects the energy storage system's ability to mitigate net load forecast errors. The smaller the value, the more balanced the charging and discharging of the energy storage system is statistically, and the less likely the SOC is to drift.

[0138] The absolute value of the expected charge and discharge power of an energy storage system can be expressed as follows:

[0139] in, This represents the absolute value of the expected charge and discharge power of the energy storage system. This represents the charging compensation coefficient. Indicates the discharge compensation coefficient. The probability density function representing the net load forecast error. This indicates the pre-set discharge power threshold of the energy storage system. This indicates the charging power threshold of the energy storage system. This represents the lower limit of the pre-set dead zone threshold for the energy storage system. This indicates the upper limit of the dead zone threshold of the energy storage system as preset.

[0140] The embodiments of this disclosure can be determined in the following ways. :1) Solve for smallest ;2) Solve for Less than or equal to the preset allowable value Specifically, the preset tolerance value δ is a threshold pre-set based on physical constraints and operational requirements. Here, grid search and local gradient descent can be used to efficiently solve for this threshold. Minimum or make Less than or equal to the preset allowable value .

[0141] In some examples, it is possible to use As a hard constraint To preset the allowable value, in Within the range, find all that satisfy the hard constraint. If the hard constraints are satisfied From multiple sets, the set that optimizes other preset indicators (such as minimizing grid interaction fluctuations) is selected. .

[0142] In some examples, it can be As the target, directly Minimize, find Lowest Then verify whether it satisfies If the requirements are not met, the configuration of the energy storage system can be adjusted, for example, by increasing the size of the system. Or capacity, find the one that satisfies and Minimized .

[0143] In practical applications, the calculation of charging compensation coefficients and discharging compensation coefficients can occur at the following times: 1) during the initial stage of system commissioning, when the charging compensation coefficients and discharging compensation coefficients are estimated using historical similar data; 2) when the distribution of prediction error changes significantly, such as during seasonal transitions or prediction model upgrades; 3) during regular maintenance, for example, by setting the charging compensation coefficients and discharging compensation coefficients to be updated monthly or quarterly. After the calculation is completed, the charging compensation coefficients and discharging compensation coefficients can be stored as fixed parameters in the station control layer's parameter library, which can be directly called without repeated optimization.

[0144] In this embodiment, the charging compensation coefficient and discharging compensation coefficient are determined by the station control layer in the source-grid-load-storage control system. Specifically, the station control layer is responsible for calculating the charging compensation coefficient and discharging compensation coefficient, and issuing the results as strategy parameters. Simultaneously, in step 405, the charging compensation coefficient and discharging compensation coefficient are used to form the constraints of the MILP (Multi-Level Product Parameter). The overall coordination control layer only uses the charging compensation coefficient and discharging compensation coefficient. In steps 402-403, the overall coordination control layer calculates the charging and discharging power of the energy storage system based on the charging compensation coefficient and discharging compensation coefficient issued by the station control layer and the real-time calculated net load prediction error. The local control layer is completely unaware of the charging compensation coefficient and discharging compensation coefficient, and only executes the specific power commands issued by the overall coordination control layer.

[0145] In related technologies, even when the net load prediction error exceeds the rated power of the energy storage system, full compensation is still used. The power of the energy storage system is limited to the maximum charging and discharging power, and compensation cannot continue. The grid connection deviation exceeds the standard, making it difficult to balance the SOC of the energy storage system over a long time scale, frequently triggering the power limit of the energy storage system, and accelerating the degradation of the energy storage system battery.

[0146] The embodiments disclosed herein employ a strategy combining full and partial compensation. A compensation coefficient is introduced to adjust the degree of partial compensation for net load prediction errors in the energy storage system, maintaining a balance between the energy storage system's throughput and output. This approach ensures both prediction accuracy and SOC balance, preventing drastic charging and discharging and extending battery life. Therefore, this energy storage control strategy, combining full and partial compensation, effectively protects the energy storage system while smoothing grid-connected power fluctuations.

[0147] The technical solutions provided in this disclosure have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this disclosure. The descriptions of the embodiments above are only for the purpose of helping to understand the methods and core ideas of this disclosure. Furthermore, those skilled in the art will recognize that, based on the ideas of this disclosure, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this disclosure.

[0148] The above description is merely a preferred embodiment of this disclosure and is not intended to limit this disclosure. Any modifications or equivalent substitutions made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A source-grid-load-storage management method, characterized in that, The method is applied to an integrated power grid system for energy generation, grid, load, and storage in an industrial park. This system includes: wind turbine generators, photovoltaic power generation equipment, energy storage systems, a power generation, grid, load, and storage control system, and loads. The method includes: The actual values ​​of active power of the park load, actual output of wind turbine generators, and actual output of photovoltaic power generation equipment were collected. The net load prediction error is determined based on the pre-obtained predicted values ​​of active power, wind power output, and photovoltaic power output for the current scheduling cycle, as well as the actual values ​​of active power, wind turbine output, and photovoltaic power output for the park load, wind turbine output, and photovoltaic power generation equipment output. Online scheduling is performed based on the net load prediction error and the previously obtained unit operation plan for the current scheduling period. The online scheduling includes: when the net load prediction error does not exceed the rated power of the energy storage system, adjusting the charging and discharging power of the energy storage system to fully compensate for the net load prediction error; when the net load prediction error exceeds the rated power of the energy storage system, adjusting the charging and discharging power of the energy storage system through predetermined charging and discharging compensation coefficients to partially compensate for the net load prediction error.

2. The method according to claim 1, characterized in that, The method further includes: Based on weather data, historical operating sequence data, and electricity market transaction information, machine learning models or time series models are used to obtain the predicted values ​​of active power, wind power output, photovoltaic power output, and electricity purchase and sale price vectors for the current dispatch cycle in the park.

3. The method according to claim 1, characterized in that, The method further includes: Based on the predicted active power, wind power output, photovoltaic power output, and electricity purchase and sale price vector of the park load for the current scheduling period, offline scheduling is performed using a mixed integer programming model to obtain the unit operation plan for the current scheduling period. The unit operation plan includes the start-up and shutdown plan and output plan of the controllable wind turbine generators, the charging and discharging plan of the energy storage system, and the electricity purchase and sale plan.

4. The method according to claim 3, characterized in that, The constraints of the mixed integer programming model include: the sum of the planned discharge power and the reserved discharge power of the energy storage system shall not exceed the rated discharge power of the energy storage system, and the reserved discharge power is determined according to the discharge compensation coefficient; and the sum of the planned charging power and the reserved charging power of the energy storage system shall not exceed the rated charging power of the energy storage system, and the reserved charging power is determined according to the charging compensation coefficient.

5. The method according to claim 3, characterized in that, The constraints of the hybrid integer programming model include: the planned SOC trajectory of the energy storage system operates within the compressed feasible interval. The upper limit of the feasible interval is the original upper limit of the energy storage system's SOC compressed downward by an upward offset safety margin. The lower limit of the feasible interval is the original lower limit of the energy storage system's SOC increased upward by a downward offset safety margin. The upward offset safety margin is determined based on the charging compensation coefficient, and the downward offset safety margin is determined based on the discharging compensation coefficient.

6. The method according to claim 1, characterized in that, Adjusting the charging and discharging power of the energy storage system by pre-determined charging and discharging compensation coefficients allows the energy storage system to partially compensate for the net load prediction error, including: Adjust the charging and discharging power of the energy storage system according to the following formula: in, This represents the charging and discharging power of the energy storage system at the i-th sampling point within the current scheduling cycle. This represents the net load prediction error for the i-th sampling point within the current scheduling period. This indicates the rated charging power of the energy storage system. This indicates the rated discharge power of the energy storage system. This represents the charging compensation coefficient. This represents the discharge compensation coefficient.

7. The method according to claim 1, characterized in that, The method further includes: The probability density function of the net load prediction error is obtained by fitting the t-location scale distribution based on historical prediction error data. Solve for the discharge compensation coefficient and charging compensation coefficient that minimize the absolute value of the expected charge and discharge power of the energy storage system or make the absolute value of the expected charge and discharge power of the energy storage system less than a preset allowable value. The absolute value of the expected charge and discharge power of the energy storage system is expressed by the following formula: in, This represents the absolute value of the expected charge and discharge power of the energy storage system. This represents the charging compensation coefficient. Indicates the discharge compensation coefficient. The probability density function representing the net load forecast error. This indicates the pre-set discharge power threshold of the energy storage system. This indicates the charging power threshold of the energy storage system. This represents the lower limit of the pre-set dead zone threshold for the energy storage system. This indicates the upper limit of the dead zone threshold of the energy storage system as preset.

8. The method according to claim 7, characterized in that, The probability density function of the net load forecast error is expressed as follows: in, The probability density function representing the net load forecast error. This indicates the net load forecast error. This represents the lower limit of the pre-set dead zone threshold for the energy storage system. This indicates the pre-set upper limit of the dead zone threshold for the energy storage system. Indicates position parameters, Indicates the scale parameter. Represents the degree of freedom parameter. This represents the gamma function.

9. An integrated power grid system for source-grid-load-storage in an industrial park, characterized in that, The integrated power grid system of the park includes: wind turbine generators, photovoltaic power generation equipment, energy storage system, power generation, grid-load-storage control system and load. The power generation, grid-load-storage control system is used to execute the method described in any one of claims 1 to 8 to compensate for the net load prediction error caused by the load, wind turbine generators and photovoltaic power generation equipment through the energy storage system.

10. The system according to claim 9, characterized in that, The source-grid-load-storage control system includes a station control layer, a general coordination control layer, and a local control layer.

Images