Semi-physical simulation platform and simulation method for load cluster tracking control

Abstract

This invention discloses a hardware-in-the-loop (HIL) simulation platform for load cluster tracking control. The master station simulation module simulates the issuance of load cluster control aggregation commands. Any smart energy unit module decomposes the load cluster control aggregation commands according to a distributed collaborative control algorithm, achieving distributed collaboration among each smart energy unit. Any adjustable load in each adjustable load cluster decomposes the load cluster control commands according to the distributed collaborative control algorithm, achieving distributed clustering of each adjustable load in the adjustable load cluster. Each adjustable load calculates its current adjustment amount based on the received load control commands and adjusts its operating state in real time according to this adjustment amount. This invention employs a distributed collaborative control strategy on the HIL platform and uses a synchronization mechanism to ensure the continuity of the distributed strategy, improving the system's stability and robustness.

Description

Technical Field

[0001] This invention relates to the field of power grid demand response user-side control technology, specifically to a hardware-in-the-loop simulation platform and simulation method for load cluster tracking control. Background Technology

[0002] With the vigorous development of new power systems driven by the "dual-carbon" strategy, distributed collaborative control methods for large-scale adjustable load clusters such as variable frequency air conditioning load clusters and energy storage system clusters have gradually become a research hotspot in academia. To more closely approximate real-world scenarios and reduce costs, hardware-in-the-loop (HIL) simulation has become a more realistic, reliable, and economical option. By conducting HIL simulations on large-scale adjustable loads, the effectiveness, practicality, and universality of distributed collaborative control strategies for adjustable loads based on multi-intelligent systems can be verified. This provides a reliable foundation and support for research on demand-side distributed collaborative control strategies.

[0003] However, real-world power systems not only contain multiple types of adjustable loads, but also need to meet the requirements of "real-time response," "plug-and-play," and "rapid disconnection" in practical engineering applications. In this context, truly applying the designed distributed cooperative control scheme to actual power system engineering is not easy. For example, due to clock delays, inconsistent updates of the states of various loads in the adjustable load cluster can lead to errors in calculation results. Furthermore, differences between different types of adjustable loads can also affect the system's real-time performance and stability. Therefore, it is necessary to design and optimize at both the software and hardware levels to address these different characteristics, and to design appropriate communication protocols and corresponding synchronization mechanisms to ensure the system's real-time performance and stability. Summary of the Invention

[0004] The purpose of this invention is to provide a hardware-in-the-loop simulation platform and simulation method for load cluster tracking control. This invention adopts a distributed cooperative control strategy on the hardware-in-the-loop platform and uses a synchronization mechanism to ensure the continuity of the distributed strategy, thereby improving the stability and robustness of the system.

[0005] To achieve this objective, the present invention provides a hardware-in-the-loop simulation platform for load cluster tracking control, which includes a master station simulation module, an adjustable load simulation module, and multiple smart energy unit modules.

[0006] The main station simulation module is used to simulate the issuance of load cluster control aggregation commands;

[0007] Any smart energy unit module is used to receive load cluster control aggregation instructions issued by the master station simulation module. Each smart energy unit decomposes the load cluster control aggregation instructions according to the distributed collaborative control algorithm, so that each smart energy unit module obtains the corresponding load cluster control instructions, thereby realizing distributed collaboration between each smart energy unit.

[0008] The adjustable load simulation module is used to simulate and aggregate dispersed adjustable loads to form an adjustable load cluster that corresponds one-to-one with each smart energy unit module. Any adjustable load in each adjustable load cluster receives the load cluster control command of the corresponding smart energy unit.

[0009] Each adjustable load in each adjustable load cluster decomposes the received load cluster control command according to the distributed cooperative control algorithm, so that each adjustable load in the adjustable load cluster receives the corresponding load control command, realizing the distributed clustering of each adjustable load in the adjustable load cluster. Each adjustable load calculates the current adjustment amount of the adjustable load according to the received load control command, and adjusts the working state of the corresponding adjustable load in real time according to the adjustment amount.

[0010] The beneficial effects of this invention are:

[0011] 1) This invention utilizes a semi-physical platform to reduce the cost of manufacturing actual physical prototypes (virtual simulation technology can replace physical prototypes, simulating the performance and behavior of products, thereby reducing manufacturing costs), lower testing and debugging costs, shorten design iteration cycles, and improve the efficiency of problem discovery and resolution. It minimizes costs while closely approximating real-world scenarios, verifying the effectiveness and practicality of the distributed collaborative control strategy for adjustable loads in a novel power system (by deploying corresponding distributed collaborative control algorithms in the smart energy unit module and the adjustable load simulation module, instructions are decomposed according to predetermined rules), providing a reliable foundation and support for research on demand-side distributed collaborative control strategies.

[0012] 2) The hardware-in-the-loop simulation platform involved has the functions of "real-time response", "plug and play" and "rapid disconnection". Real-time response is the real-time decomposition of instructions, which is achieved through highly reliable communication. The distributed collaborative algorithm deployed by this invention and the highly reliable communication technology adopted can be applied to various practical engineering application scenarios, such as "emergency interaction", "real-time load access" and "real-time load disconnection", to carry out strategy effectiveness verification.

[0013] 3) A distributed collaborative control strategy is adopted on the semi-physical platform, and a synchronization mechanism is used to ensure the continuity of the distributed strategy, which improves the stability and robustness of the system (low communication redundancy and communication requirements, and the topology is not easily attacked by external forces). Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the structure of the present invention;

[0015] Figure 2 This is a functional diagram of a hardware-in-the-loop simulation platform;

[0016] Figure 3 It is a modular software architecture for a hardware-in-the-loop simulation platform;

[0017] Figure 4 It is a topology architecture of a 6-node smart energy unit in a hardware-in-the-loop simulation platform;

[0018] Figure 5 It is a modular hardware architecture for a hardware-in-the-loop simulation platform;

[0019] Figure 6 This is a schematic diagram of a large-scale adjustable load cluster architecture for a hardware-in-the-loop simulation platform.

[0020] Figure 7 This is a flowchart of the synchronization mechanism;

[0021] Figure 8 It is the message format in the synchronization mechanism;

[0022] Figure 9 This is a schematic diagram of the initial state of an energy storage system cluster according to an embodiment of the present invention;

[0023] Figure 10 This is a schematic diagram of the initial state of a variable frequency air conditioning cluster according to an embodiment of the present invention;

[0024] Figure 11 This is the first instruction issued according to an embodiment of the present invention;

[0025] Figure 12 This is a schematic diagram of the energy storage system cluster state after the first adjustment according to an embodiment of the present invention;

[0026] Figure 13 This is a schematic diagram of the inverter air conditioning cluster state after the first adjustment according to an embodiment of the present invention.

[0027] Figure 14 This is the second instruction issued according to an embodiment of the present invention;

[0028] Figure 15 This is a schematic diagram of the energy storage system cluster state after the second adjustment according to an embodiment of the present invention;

[0029] Figure 16 This is a schematic diagram of the variable frequency air conditioning cluster state after the second adjustment according to an embodiment of the present invention. Detailed Implementation

[0030] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:

[0031] like Figures 1-3 The hardware-in-the-loop simulation platform for load cluster tracking control shown includes a master station simulation module, an adjustable load simulation module, and multiple smart energy unit modules.

[0032] The main station simulation module is used to simulate the issuance of load cluster control aggregation commands;

[0033] Any smart energy unit module is used to receive load cluster control aggregation instructions issued by the master station simulation module. Each smart energy unit decomposes the load cluster control aggregation instructions according to the distributed collaborative control algorithm, so that each smart energy unit module obtains the corresponding load cluster control instructions, thereby realizing distributed collaboration between each smart energy unit.

[0034] The adjustable load simulation module is used to simulate and aggregate dispersed adjustable loads to form an adjustable load cluster that corresponds one-to-one with each smart energy unit module. Any adjustable load in each adjustable load cluster receives the load cluster control command of the corresponding smart energy unit.

[0035] Each adjustable load in each adjustable load cluster decomposes the load cluster control command received from the corresponding smart energy unit according to the distributed collaborative control algorithm, so that each adjustable load in the adjustable load cluster receives the corresponding load control command, realizing the distributed clustering of each adjustable load in the adjustable load cluster. Each adjustable load calculates the current adjustment amount of the adjustable load according to the received load control command, and adjusts the working state of the corresponding adjustable load in real time according to the adjustment amount.

[0036] In the above technical solution, instruction decomposition is performed in both the smart energy unit and the adjustable load cluster to achieve multi-level instruction decomposition. Multi-level decomposition can accelerate the decomposition speed of instructions and improve the performance of the entire system. Furthermore, the distributed collaborative approach has low computational load, low communication requirements and redundancy, high reliability, and strong robustness.

[0037] In the above technical solution, the main function of the semi-physical simulation platform is to receive the aggregated control instructions issued by the master station or demand response center to the entire adjustable load group. Based on the information reported by the adjustable load group, and according to the principles of cost, adjustment constraints, economy and fairness reported by the adjustable load, the platform uses a distributed collaborative control algorithm deployed on the smart energy unit module and the adjustable load simulation module to achieve the collaborative decomposition of the aggregated control instructions of the adjustable load group.

[0038] In the above technical solution, the adjustable load simulation module is also used to monitor the working status of each adjustable load cluster in real time, and report the working status of each adjustable load cluster to the smart energy unit module to participate in the command and dispatch of the power system.

[0039] In the above technical solution, the smart energy unit module is also used to report the working status information of the adjustable load set to the main station simulation module, and the main station simulation module displays the changes in the working status of each adjustable load during the adjustment process.

[0040] In the above technical solution, when a new smart energy unit module is connected or a smart energy unit module is disconnected, the smart energy unit module performing distributed collaborative control calculations redistributes the load cluster control aggregation commands according to the distributed collaborative control algorithm. When a new adjustable load is added or an adjustable load is disconnected from an adjustable load cluster, the adjustable load module performing distributed collaborative control calculations redistributes the received load cluster control commands according to the distributed collaborative control algorithm. The semi-physical simulation platform is adaptable to the "connection" and "disconnection" of adjustable load resources. When a smart energy unit detects that the load device it manages has been "connected" or "disconnected," it quickly redistributes the total control deficit using the distributed collaboration of the smart energy unit. Local changes are mitigated by the entire adjustable load cluster participating in adaptive adjustment, maintaining the globally optimized configuration of the load cluster. The above design can handle the connection or disconnection of any load in the network cluster while ensuring command tracking and adaptive adjustment.

[0041] In the above technical solution, the multiple smart energy unit modules adopt a ring topology (e.g., Figure 4 As shown, each smart energy unit module communicates with two other adjacent smart energy unit modules. One smart energy unit module acts as the leader node, communicating with the master station's analog module. Compared to all smart energy units communicating with the master station, communication between only one smart energy unit and the master station reduces communication overhead, while communication with two other adjacent smart energy unit modules enables distributed collaboration.

[0042] In the above technical solution, each smart energy unit decomposes the load cluster control aggregation command according to the distributed collaborative control algorithm, so that each smart energy unit module receives the corresponding load cluster control command. The specific method for achieving distributed collaboration between each smart energy unit is as follows:

[0043]

[0044]

[0045] Among them, P all This indicates the command issued by the master station simulation module to the smart energy unit cluster, Q.i This represents the total electricity aggregated by the i-th smart energy unit module. This indicates the maximum capacity of each smart energy unit module. q i相邻 (l) is the average value, where i-adjacent represents the smart energy unit adjacent to the i-th smart energy unit, and l is a time series; q i (l) is an intermediate variable that has no actual physical meaning, p i (l) represents the power of energy storage unit i (i.e., the corresponding load cluster) at time l = 1, 2, ..., n, and ξ is the threshold for stopping iteration, selected according to system requirements. i (l+1) represents the power of energy storage unit i at time l+1, which is the corresponding load cluster control command.

[0046] In the above technical solution, the adjustable load cluster is an energy storage system cluster. The distributed control algorithm also differs for different adjustable load clusters. Taking an energy storage system cluster as an example, the basic principle of the distributed control algorithm is illustrated. The structural principles of other adjustable load clusters are similar. The system architecture of the adjustable system cluster can be divided into two layers: the physical layer and the communication layer. A schematic diagram of these two layers is shown below. Figure 6 As shown. At the communication layer, each energy storage unit within the energy storage system cluster can exchange charging and discharging power and current real-time capacity information with neighboring nodes through the communication network. In the entire system, at least one, but not all, energy storage units can receive external dispatch commands and sense the total charging and discharging power of the energy storage system cluster. At the physical layer, each energy storage unit is connected to a microgrid, capable of absorbing electrical energy from the microgrid and providing corresponding electrical energy when necessary. Considering that the energy storage system cluster consists of N energy storage units, the specific method for each energy storage unit in the energy storage system cluster to decompose the received load cluster control commands according to the distributed cooperative control algorithm is as follows:

[0047]

[0048] Where e represents the deviation between the total charging and discharging power of the energy storage system cluster and the external dispatch command at time t. During the control process, the energy storage system cluster uses the deviation e to enable the system to track the external power dispatch command in real time. Each energy storage unit i exchanges local charging and discharging power and real-time capacity information with neighboring nodes at the communication layer to achieve consistency in the system's SoC. c1, c2, and c3 represent the first, second, and third control parameters of the system, respectively. c1, c2, and c3 are constants set according to system requirements. i represents the i-th energy storage unit, and a ij P represents the communication weight between energy storage unit i and its neighbors. i P represents the power of the i-th energy storage unit. j E represents the power of the j-th energy storage unit.i E represents the energy of the i-th energy storage unit. j Let φ represent the energy level of the j-th energy storage unit, and let φ represent the set of energy storage units that can receive external dispatch commands and sense the total charging and discharging power of the energy storage system cluster. i This represents the control input for the i-th energy storage unit.

[0049] In the above technical solution, the distributed control algorithm for adjustable loads needs to ensure that the data exchanged by each adjustable load in the adjustable load cluster is synchronized, avoiding data inconsistencies that may occur during the calculation process due to different operating environments. Therefore, synchronization is performed in distributed computing. Figure 7 This section demonstrates the internal workings of this synchronization mechanism. First, adjustable loads directly connected to the load controller in the load cluster network are called central nodes. These nodes can obtain global information within the network. The remaining adjustable loads are called other nodes. The central nodes are virtual central nodes formed by the Gossip global broadcast algorithm. At the physical layer, they do not establish communication connections with every node. The Gossip algorithm refers to using random methods to allow nodes to communicate with each other, thereby achieving information propagation. Specifically, when a node has new information, it randomly selects another node to send this information to. The receiving node marks the information as received and then broadcasts it to its neighboring nodes. In this way, information continuously spreads through inter-node transmission until all nodes have received the information. Nodes communicate using specially encoded messages, such as... Figure 8 As shown, a message includes a frame header, function code, status information, number, and frame tail CRC checksum. The frame header uses a predefined 0xbadbeef byte sequence as the start position of the data frame. The function code is 4 bits long, where 0x1 indicates that this data is sent by each node to the central node and neighboring nodes after each calculation; 0x2 indicates that this data is an instruction message broadcast globally by the central node. The status information is 4 bytes long and contains the consistency variable Xi that the node needs to exchange with neighboring nodes after each calculation. The number is 4 bits long and indicates the sending node number of this data frame. Finally, there is a CRC checksum at the frame tail, which is used to indicate the end position of the data frame and to check whether the data frame is complete during transmission. After each calculation, the other nodes encode the calculated consistency status information Xi in the above format, set the function code to 1, and send it out through Gossip global broadcast to collect neighboring node interaction information. When all neighboring node information has been saved, the node sets its own global flag bit Flag2 to 1 and waits for the central node to send the instruction message for another calculation.

[0050] When a message with function code 2 is received, it means that all nodes have completed their calculations. The global flag flag 1 is then set to 1, and the next calculation is performed.

[0051] After the central node completes its own calculations and forwards them, it collects signals from all nodes indicating that their calculations are complete. Once all nodes have finished their calculations, the central node uses Gossip global broadcast to send a command to each node to begin the next calculation. This ensures that the clocks are synchronized on a single time scale, thereby guaranteeing the successful implementation of the distributed collaborative algorithm.

[0052] In the above technical solution, the master station simulation module uses a TI AM3352 CPU based on the ARM Cortex A8 core to meet the requirements of fast management and control functions. The AM3352 development board of the simulation master station can be developed for specific scenarios, displaying the changes of the entire adjustable load during the adaptive adjustment process on the monitor, and allowing interaction on the monitor's UI interface to issue simulated aggregation commands. The simulation master station layer can receive monitoring and reporting information uploaded by the adjustable load group.

[0053] In the main station module, the UI interface can interactively send cluster scheduling commands to the entire cluster control framework, display the overall cluster's operating status and the tracking of aggregation commands, and also view the specific adjustability of each load cluster participating in the system's control. In the smart energy unit module, the UI interface can display the real-time status of adjustable load clusters.

[0054] In terms of communication, the smart energy unit modules use the IEEE 802.11 local area network wireless network communication protocol and the IEEE 802.3 local area network communication protocol to simulate the remote wireless link of the smart energy units by artificially increasing the latency (the communication latency of 5G private network can usually reach the level of 10 milliseconds).

[0055] The adjustable load simulation module utilizes six STM32F429 embedded processor platforms and a Raspberry Pi 4B computing platform to build an adjustable load layer to simulate an adjustable load cluster, supporting physical simulation and pure digital simulation using external small devices. Depending on the type of adjustable load, adjustable resources can be aggregated into the same load cluster, forming a large-scale load cluster network through a dedicated communication network and information on adjacent adjustable loads.

[0056] In this invention, the UI (User Interface) module is deployed in both the simulation master station module and the smart energy unit module. In the master station module, the UI can interactively send cluster scheduling commands to the entire cluster control framework, display the overall cluster's operating status and the tracking of aggregated commands, and view the specific adjustability of each load cluster participating in the control process. In the smart energy unit module, the UI can display the real-time status of the adjustable load clusters. A distributed coordination module is deployed in each smart energy unit module, and each smart energy unit sends scheduling commands to the adjustable load clusters by executing the instructions from the distributed coordination module. A distributed computing module is deployed in each adjustable load module, distributing the commands sent by the smart energy units to each adjustable resource according to a distributed algorithm. Different distributed control algorithms are used for different adjustable load clusters. A reporting and feedback module is deployed in both the smart energy unit module and the adjustable load simulation module, reporting the status of the adjustable load clusters and the status of individual resources within each adjustable load cluster to the simulation master station module and the smart energy unit module, respectively.

[0057] The hardware-in-the-loop simulation platform involves the processing of a large amount of data, so a communication architecture needs to be designed to meet the requirements of the control framework for high reliability and stability. Figure 5 This paper demonstrates the communication architecture between various hardware-in-the-loop (HIL) modules in a hardware-in-the-loop simulation platform. Multiple communication protocols and technologies are employed to address the data transmission characteristics between different layers. Each layer has clearly defined uplink and downlink interfaces and protocols to ensure data transmission and processing proceed according to regulations. Specifically, the communication between the simulation master station and the smart energy unit requires information transmission that is not easily affected by external interference, ensuring the security and reliability of control commands. Therefore, they can be interconnected via network cables and switches, using the Socket communication protocol for data transmission. Since the adjustable loads are scattered throughout the building, establishing physical connections is inconvenient. Furthermore, to reduce the cost of user access to aggregated control, wireless communication technologies such as Wi-Fi are used to connect the adjustable loads and smart energy units, using the more reliable Modbus communication protocol for data transmission. This communication network architecture meets the requirements of the HIL platform, which involves large-scale data processing and exhibits high reliability and stability.

[0058] A hardware-in-the-loop simulation method for load cluster tracking control includes the following steps:

[0059] Step 1: The main station simulation module simulates the issuance of load cluster control aggregation commands;

[0060] Step 2: Any smart energy unit module is used to receive the load cluster control aggregation command issued by the master station simulation module. Each smart energy unit decomposes the load cluster control aggregation command according to the distributed collaborative control algorithm, so that each smart energy unit module obtains the corresponding load cluster control command, thereby realizing distributed collaboration between each smart energy unit.

[0061] Step 3: The adjustable load simulation module is used to simulate and aggregate the dispersed adjustable loads to form an adjustable load cluster corresponding to each smart energy unit module. Any adjustable load in each adjustable load cluster receives the load cluster control command of the corresponding smart energy unit.

[0062] Step 4: Each adjustable load in each adjustable load cluster decomposes the received load cluster control command according to the distributed cooperative control algorithm, so that each adjustable load in the adjustable load cluster receives the corresponding load control command, realizing the distributed clustering of each adjustable load in the adjustable load cluster. Each adjustable load calculates the current adjustment amount of the adjustable load according to the received load control command, and adjusts the working state of the corresponding adjustable load in real time according to the adjustment amount.

[0063] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method described above.

[0064] This embodiment sets up a variable frequency air conditioning load cluster and a homogeneous energy storage system cluster, and observes the power point tracking (PPT) performance of each load cluster by inputting continuously changing adjustment commands from an external source. The specific steps are as follows:

[0065] S1: Assume a variable frequency air conditioning load cluster contains 10 variable frequency air conditioners with rated powers of 800W, 800W, 800W, 1200W, 1200W, 1200W, 1800W, 1800W, 2780W, and 2780W, respectively. Their natural frequency upper and lower limits are 130Hz and 30Hz, respectively. Initially, their power varies due to factors such as room size and set temperature, at 200W, 300W, 420W, 1000W, 700W, 1120W, 1500W, 1000W, 2000W, and 1200W, respectively. Therefore, assume the adjustable power range of the variable frequency air conditioning load cluster is approximately 2000W, and the initial consistency state variable Xi for each air conditioner is 0.

[0066] S2: Assume a homogeneous energy storage system cluster contains 10 identical energy storage units, each with a maximum capacity of 40 kWh. Initially, each energy storage unit has a power output of 0 W and capacities of 5 kWh, 10 kWh, 8 kWh, 3 kWh, 1 kWh, 15 kWh, 7 kWh, 19 kWh, 13 kWh, and 11 kWh, respectively. Considering the building's power supply capacity and quality, the maximum charging and discharging power of each energy storage unit is set to 500 W. Assume the adjustable power range of the homogeneous energy storage system cluster is approximately 5000 W.

[0067] S3: The initial state of each load cluster at the initial moment is as follows: Figure 9 and Figure 10 As shown. Figure 11 As shown, the load aggregation unit first issues a dispatch command to the system to increase power by 3500W. Based on the adjustability of the two load clusters, the load aggregation unit decomposes the command into: an instruction to increase power by 2500W for the energy storage system cluster and an instruction to increase power by 1000W for the variable frequency air conditioning load cluster.

[0068] S4: The status of each load cluster after adjustment is as follows Figure 12 and Figure 13 As shown, within a short period of time, each adjustable load can reach the predetermined target through the internal command tracking algorithm. In the energy storage system cluster, the SoC of each energy storage unit reaches a consistent state, with a capacity of 26.554 kWh. The charging power of each energy storage unit is proportionally allocated, at 250 W based on its own capacity. The total charging power of the entire energy storage system cluster is 2500 W, capable of tracking external dispatch commands. In the variable frequency air conditioning cluster, the utilization level of each different variable frequency air conditioner is similar. Specifically, after adjustment, the consistency state variable X of each variable frequency air conditioner... i All values ​​are 0.1486. ​​Furthermore, the total power variation of each inverter air conditioner meets the system's total upward adjustment requirement of 1000W, thus achieving power balance.

[0069] S5: Next, continue to issue a dispatch command to the system from the load aggregation unit to reduce the power by 3500W, such as... Figure 14 As shown. Similarly, the overall dispatch command will be broken down into a 2500W power reduction command for the energy storage system cluster and a 1000W power reduction command for the variable frequency air conditioning load cluster.

[0070] S6: Figure 15 , Figure 16The status of each load cluster after the load reduction is shown, demonstrating that each adjustable load can immediately track changes in external commands. Specifically, the SoC of each energy storage unit in the energy storage system cluster is consistent, with a capacity of 9.206 kWh. The charging power of each energy storage unit is proportionally allocated according to its capacity, all at -250W. The total discharge power of the energy storage system cluster is -2500W, enabling it to track external dispatch commands. Furthermore, the utilization level of the adjustment capabilities of each different variable frequency air conditioner in the variable frequency air conditioning cluster is similar, reflected in the consistency of the state variable X. i After adjustment, all values ​​were 0.1863. Furthermore, the total power change of each inverter air conditioner met the system's total power reduction requirement of 1000W, achieving power balance.

[0071] Based on the experimental results of the two adjustments above, it can be concluded that the tracking control algorithms for large-scale energy storage system cluster regulation and large-scale variable frequency air conditioning S-regulation proposed in this paper can track external control commands well in the hardware-in-the-loop simulation experimental platform, and have certain practical application value. Therefore, they can be further applied to practical engineering.

[0072] The contents not described in detail in this specification are existing technologies known to those skilled in the art.

Claims

1. A hardware-in-the-loop simulation platform for load cluster tracking control, characterized in that: It includes a master station simulation module, an adjustable load simulation module, and multiple smart energy unit modules; The main station simulation module is used to simulate the issuance of load cluster control aggregation commands; Any smart energy unit module is used to receive load cluster control aggregation instructions issued by the master station simulation module. Each smart energy unit decomposes the load cluster control aggregation instructions according to the distributed collaborative control algorithm, so that each smart energy unit module obtains the corresponding load cluster control instructions, thereby realizing distributed collaboration between each smart energy unit. The adjustable load simulation module is used to simulate and aggregate dispersed adjustable loads to form an adjustable load cluster that corresponds one-to-one with each smart energy unit module. Any adjustable load in each adjustable load cluster receives the load cluster control command of the corresponding smart energy unit. Each adjustable load in each adjustable load cluster decomposes the received load cluster control command according to the distributed cooperative control algorithm, so that each adjustable load in the adjustable load cluster receives the corresponding load control command, realizing the distributed clustering of each adjustable load in the adjustable load cluster. Each adjustable load calculates the current adjustment amount of the adjustable load according to the received load control command, and adjusts the working state of the corresponding adjustable load in real time according to the adjustment amount.

2. The hardware-in-the-loop simulation platform for load cluster tracking control according to claim 1, characterized in that: The adjustable load simulation module is also used to monitor the working status of each adjustable load cluster in real time and report the working status of each adjustable load cluster to the smart energy unit module.

3. The hardware-in-the-loop simulation platform for load cluster tracking control according to claim 2, characterized in that: The smart energy unit module is also used to report the working status information of the adjustable load set to the main station simulation module, which displays the changes in the working status of each adjustable load during the adjustment process.

4. The hardware-in-the-loop simulation platform for load cluster tracking control according to claim 1, characterized in that: When a new smart energy unit module is connected or a smart energy unit module is disconnected, the smart energy unit module performing distributed collaborative control calculations redistributes the load cluster control aggregation instructions according to the distributed collaborative control algorithm.

5. The hardware-in-the-loop simulation platform for load cluster tracking control according to claim 1, characterized in that: When a new adjustable load is added to an adjustable load cluster or an adjustable load is removed, the adjustable loads performing distributed collaborative control calculations redistribute the received load cluster control commands according to the distributed collaborative control algorithm.

6. The hardware-in-the-loop simulation platform for load cluster tracking control according to claim 1, characterized in that: The multiple smart energy unit modules adopt a ring topology. Each smart energy unit module communicates with two other adjacent smart energy unit modules. Among them, one smart energy unit module serves as the leader node uplink and master station simulation module.

7. The hardware-in-the-loop simulation platform for load cluster tracking control according to claim 1, characterized in that: The adjustable load cluster is an energy storage system cluster. The specific method by which each energy storage unit of the energy storage system cluster decomposes the received load cluster control commands according to the distributed cooperative control algorithm is as follows: Where e represents the deviation between the total charging and discharging power of the energy storage system cluster and the external dispatch command at time t. During the control process, the energy storage system cluster uses the deviation e to enable the system to track the external power dispatch command in real time. Each energy storage unit i exchanges local charging and discharging power and real-time capacity information with neighboring nodes at the communication layer to achieve consistency in the system's SoC. c1, c2, and c3 represent the first, second, and third control parameters of the system, respectively. c1, c2, and c3 are constants set according to system requirements. i represents the i-th energy storage unit, and a ij P represents the communication weight between energy storage unit i and its neighbors. i P represents the power of the i-th energy storage unit. j E represents the power of the j-th energy storage unit. i E represents the energy of the i-th energy storage unit. j Let φ represent the energy level of the j-th energy storage unit, and let φ represent the set of energy storage units that can receive external dispatch commands and sense the total charging and discharging power of the energy storage system cluster. i This represents the control input for the i-th energy storage unit.

8. The hardware-in-the-loop simulation platform for load cluster tracking control according to claim 1, characterized in that: Distributed control algorithms for adjustable loads need to ensure that the data exchanged between each adjustable load in the adjustable load cluster is synchronized, avoiding data inconsistencies that may occur during computation due to different operating environments. Therefore, synchronization in distributed computing involves first designating the adjustable loads directly connected to the load controller in the load cluster network as central nodes. These central nodes can obtain global information about the network. The remaining adjustable loads are called the other nodes. The central nodes are virtual central nodes formed by the Gossip global broadcast algorithm. They do not establish communication connections with every node at the physical layer. The Gossip algorithm refers to using random methods to allow nodes to communicate with each other, thereby achieving the purpose of information propagation. Specifically, when a node has a new message, it randomly selects another node to send this message to. The receiving node marks the message as received and then broadcasts it to its neighboring nodes. In this way, the information is continuously spread through the transmission between nodes until all nodes have received the message. Nodes communicate with each other using a specially encoded message. Communication involves a message consisting of a frame header, function code, status information, number, and frame tail CRC checksum. The frame header uses a predefined 0xbadbeef byte sequence as the start position of the data frame. The function code is 4 bits long, where 0x1 indicates that this data is sent by each node to the central node and neighboring nodes after each calculation; 0x2 indicates that this data is an instruction message broadcast globally by the central node. The status information is 4 bytes long and contains the consistency variable Xi that the node needs to exchange with neighboring nodes after each calculation. The number is 4 bits long and indicates the sending node number of this data frame. Finally, there is a CRC checksum at the frame tail, used to indicate the end position of the data frame and to check whether the data frame is complete during transmission. After each calculation, the remaining nodes encode the calculated consistency status information Xi in the above format, set the function code to 1, and send it out through Gossip global broadcast to collect the exchanges from neighboring nodes. After all the neighboring node information has been saved, the node sets its own global flag bit Flag2 to 1, waiting for the central node to send the instruction message for another calculation. When a message with function code 2 is received, it means that all nodes have completed their calculations. The global flag flag 1 is then set to 1, and the next calculation is performed. After the central node has calculated its own information and forwarded it, it collects the signals from all nodes indicating that the calculation is complete. Once all nodes have finished calculating, it sends a command to each node to proceed with the next calculation via Gossip global broadcast.

9. A hardware-in-the-loop simulation method for load cluster tracking control, characterized in that, It includes the following steps: Step 1: The main station simulation module simulates the issuance of load cluster control aggregation commands; Step 2: Any smart energy unit module is used to receive the load cluster control aggregation command issued by the master station simulation module. Each smart energy unit decomposes the load cluster control aggregation command according to the distributed collaborative control algorithm, so that each smart energy unit module obtains the corresponding load cluster control command, thereby realizing distributed collaboration between each smart energy unit. Step 3: The adjustable load simulation module is used to simulate and aggregate the dispersed adjustable loads to form an adjustable load cluster corresponding to each smart energy unit module. Any adjustable load in each adjustable load cluster receives the load cluster control command of the corresponding smart energy unit. Step 4: Each adjustable load in each adjustable load cluster decomposes the received load cluster control command according to the distributed cooperative control algorithm, so that each adjustable load in the adjustable load cluster receives the corresponding load control command, realizing the distributed clustering of each adjustable load in the adjustable load cluster. Each adjustable load calculates the current adjustment amount of the adjustable load according to the received load control command, and adjusts the working state of the corresponding adjustable load in real time according to the adjustment amount.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, it implements the steps of the method as described in claim 9.

Images