A control method and device of a high-voltage cascade energy storage system

Abstract

The application discloses a kind of control method and device of high-voltage cascaded energy storage system, it is related to electric power energy storage technical field, comprising: receiving the charge-discharge instruction of power grid dispatching center, and carry out decomposition and generate multistage power instruction set;Real-time monitoring is carried out to high-voltage cascaded energy storage system, obtains multiple unit state parameter set, and carries out dynamic balance analysis according to multistage power instruction set and multiple unit state parameter set, generates multistage collaborative scheduling instruction;Multistage collaborative scheduling instruction is mapped to high-voltage cascaded energy storage system and carries out prediction control, is converted into charge-discharge control instruction set and is executed, while monitoring record is carried out, generates system response data and carries out closed-loop feedback correction to multistage collaborative scheduling instruction, obtains system optimization control parameter set.The application solves the technical problem that charge-discharge control instruction in prior art cannot accurately adapt to power grid load fluctuation, achieves the effect of optimizing scheduling instruction, to improve the technical effect of system response speed and stability.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power energy storage, in particular to a control method and device of high-voltage cascaded energy storage system. BACKGROUND

[0002] In the control process of high-voltage cascaded energy storage system, the charge and discharge control instructions are usually adjusted according to the fluctuation of power grid load. However, the traditional charge and discharge control method cannot accurately respond to the actual situation of load fluctuation in time. The power grid load fluctuation has complex frequency characteristics, usually including high-frequency and low-frequency fluctuation components, which have different effects on the scheduling of energy storage system at different time scales. The traditional method fails to effectively identify and distribute these fluctuation components, resulting in insufficient response ability of the charge and discharge instructions, thereby affecting the overall efficiency and stability of the system. SUMMARY

[0003] The present application provides a control method and device of high-voltage cascaded energy storage system, which is used to solve the technical problem that the charge and discharge control instructions in the prior art cannot accurately adapt to the fluctuation of power grid load.

[0004] In view of the above problems, the present application provides a control method and device of high-voltage cascaded energy storage system.

[0005] The first aspect of the present application provides a control method of high-voltage cascaded energy storage system, the method comprising:

[0006] receiving the charge and discharge instructions of the power grid dispatching center, decomposing the charge and discharge instructions according to the load fluctuation characteristics, generating a multi-level power instruction set; real-time monitoring the high-voltage cascaded energy storage system to obtain a multi-unit state parameter set, performing dynamic balance analysis according to the multi-level power instruction set and the multi-unit state parameter set, generating a multi-level collaborative scheduling instruction; mapping the multi-level collaborative scheduling instruction to the high-voltage cascaded energy storage system for predictive control, converting it into a charge and discharge control instruction set, executing the charge and discharge control instruction set for monitoring and recording, generating system response data to close-loop feedback correct the multi-level collaborative scheduling instruction, and obtaining a system optimization control parameter set.

[0007] The second aspect of the present application provides a control device of high-voltage cascaded energy storage system, the device comprising:

[0008] The instruction decomposition module is used for receiving the charge-discharge instruction of the power grid dispatching center, calling the load fluctuation characteristics to decompose the charge-discharge instruction, and generating a multi-stage power instruction set; the real-time monitoring module is used for monitoring the high-voltage cascaded energy storage system in real time, obtaining a multi-unit state parameter set, performing dynamic balance analysis according to the multi-stage power instruction set and the multi-unit state parameter set, and generating a multi-stage collaborative scheduling instruction; the feedback correction module is used for mapping the multi-stage collaborative scheduling instruction to the high-voltage cascaded energy storage system for predictive control, converting it into a charge-discharge control instruction set, executing the charge-discharge control instruction set for monitoring and recording, generating system response data to perform closed-loop feedback correction on the multi-stage collaborative scheduling instruction, and obtaining a system optimization control parameter set.

[0009] The one or more technical solutions provided in the application have at least the following technical effects or advantages:

[0010] The application receives the charge-discharge instruction of the power grid dispatching center, calls the load fluctuation characteristics to decompose the charge-discharge instruction, and generates a multi-stage power instruction set; monitors the high-voltage cascaded energy storage system in real time, obtains a multi-unit state parameter set, performs dynamic balance analysis according to the multi-stage power instruction set and the multi-unit state parameter set, and generates a multi-stage collaborative scheduling instruction; maps the multi-stage collaborative scheduling instruction to the high-voltage cascaded energy storage system for predictive control, converts it into a charge-discharge control instruction set, executes the charge-discharge control instruction set for monitoring and recording, generates system response data to perform closed-loop feedback correction on the multi-stage collaborative scheduling instruction, and obtains a system optimization control parameter set. The application solves the technical problem that the charge-discharge control instruction in the prior art cannot accurately adapt to the power grid load fluctuation, achieves the effect of optimizing the scheduling instruction by decomposing the charge-discharge instruction and combining the multi-unit state parameter for dynamic balance analysis, thereby improving the system response speed and stability. BRIEF DESCRIPTION OF DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can also be obtained by those skilled in the art without any creative effort on the basis of these drawings.

[0012] Figure 1 A control method flow diagram of a high-voltage cascaded energy storage system is provided for the embodiments of the application.

[0013] Figure 2 A control device structure diagram of a high-voltage cascaded energy storage system is provided for the embodiments of the application.

[0014] Explanation of reference signs: instruction decomposition module 11, real-time monitoring module 12, feedback correction module 13. DETAILED DESCRIPTION

[0015] The present application provides a control method and device for a high-voltage cascade energy storage system, aiming to solve the technical problem that the charge-discharge control instruction in the prior art cannot accurately adapt to the load fluctuation of the power grid. By decomposing the charge-discharge instruction and combining dynamic balancing analysis of multiple unit state parameters, the effect of optimizing the dispatching instruction is achieved, thereby improving the response speed and stability of the system.

[0016] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0017] It should be noted that any variation of the terms "comprise" and "have" is intended to cover non-exclusive inclusion, for example, a process, method, device, product or server comprising a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or modules that are not clearly listed or inherent to these processes, methods, products or devices.

[0018] Embodiment one, as shown in the present application provides a control method for a high-voltage cascade energy storage system, the method comprising: Figure 1

[0019] Step S100: receiving the charge-discharge instruction from the power grid dispatching center, retrieving the load fluctuation characteristics to decompose the charge-discharge instruction, and generating a multi-stage power instruction set.

[0020] In the embodiments of the present application, first, the charge-discharge instruction from the power grid dispatching center is received, which contains the charge or discharge operation that the energy storage system should perform within a certain period of time, for guiding the operation of the energy storage system.

[0021] Next, the load fluctuation characteristics are retrieved to decompose the charge-discharge instruction. In this process, first, the power grid load fluctuation characteristics are identified by spectrum recognition, and the high-frequency fluctuation component and the low-frequency fluctuation component are identified. Then, the charge-discharge instruction is traversed, and the high-voltage cascade energy storage system is analyzed for response to obtain first and second response speed parameters. According to these response speed parameters, the first and second energy storage unit clusters are determined respectively, and the high-frequency fluctuation component is allocated to the first energy storage unit cluster, and the low-frequency fluctuation component is allocated to the second energy storage unit cluster. Finally, a multi-stage power instruction set is generated.

[0022] ​Further, the method provided by the application embodiment further comprises the following steps:

[0023] spectrum recognition is performed on the load fluctuation characteristics to determine high-frequency fluctuation components and low-frequency fluctuation components; response analysis is performed on the high-voltage cascade energy storage system by traversing the charge-discharge instructions to obtain a first response speed parameter and a second response speed parameter; a first energy storage unit cluster is determined according to the first response speed parameter, and a second energy storage unit cluster is determined according to the second response speed parameter; the high-frequency fluctuation components are distributed to the first energy storage unit cluster, and the low-frequency fluctuation components are distributed to the second energy storage unit cluster, to generate the multi-level power instruction set.

[0024] In the application embodiment, when the spectrum recognition is performed on the load fluctuation characteristics, first, the historical load data of the high-voltage cascade energy storage system is collected and change calculation is performed to obtain a load change rate. Then, fast Fourier transform (FFT) is performed according to the load change rate to extract a spectrum energy distribution proportion value. Subsequently, wavelet analysis is performed in combination with the fluctuation period and the spectrum energy distribution proportion value, and an appropriate decomposition layer number is set according to the analysis result. Finally, spectrum decomposition reconstruction is performed on the load fluctuation characteristics according to the set decomposition layer number, so as to identify the high-frequency fluctuation components and the low-frequency fluctuation components.

[0025] Then, according to the identified high-frequency and low-frequency fluctuation components, the received charge-discharge instructions are traversed, and response analysis is performed on the high-voltage cascade energy storage system. In this step, the dynamic response of the energy storage system is collected and analyzed in real time to obtain a first response speed parameter and a second response speed parameter. The first response speed parameter refers to the reaction speed of the energy storage system to the high-frequency fluctuation, and generally requires that the response time is not more than 100 ms, and is suitable for energy storage units (such as supercapacitors) that need to respond quickly. The second response speed parameter refers to the reaction speed of the energy storage system to the low-frequency fluctuation, and generally requires that the response time is greater than or equal to 1 second, and is suitable for energy storage units (such as lithium batteries and lead-acid batteries) that respond slowly.

[0026] Based on the first response speed parameter and the second response speed parameter, the matching of the energy storage unit cluster is performed. The first energy storage unit cluster is matched according to the first response speed parameter, and these energy storage units are mainly used to process high-frequency fluctuations, so the energy storage devices with fast response speed (such as supercapacitors) are selected. The second energy storage unit cluster is matched according to the second response speed parameter, and processes low-frequency fluctuations, so the energy storage devices with slow response speed (such as lithium batteries and lead-acid batteries) are selected.

[0027] Finally, the high-frequency fluctuation components are distributed to the first energy storage unit cluster, and the low-frequency fluctuation components are distributed to the second energy storage unit cluster. In this way, the multi-level power instruction set is generated.

[0028] Further, the method provided by the application embodiment further comprises:

[0029] The historical load data of the high-voltage cascade energy storage system is collected to calculate the load change rate; fast Fourier transform is performed according to the load change rate to extract the frequency spectrum energy distribution proportion; wavelet analysis is performed according to the fluctuation period and the frequency spectrum energy distribution proportion, and the decomposition layer number is set; the load fluctuation characteristics are decomposed and reconstructed according to the decomposition layer number to identify the high-frequency fluctuation component and the low-frequency fluctuation component.

[0030] In the application embodiment, first, historical load data is obtained from a high-voltage cascade energy storage system. The data is provided by a power grid dispatching center or real-time monitoring equipment of the energy storage system, and contains load demand records of the power grid within a certain period of time. To ensure the accuracy of the data, data is collected regularly from multiple monitoring points (such as substations, energy storage units, load devices, etc.), and abnormal values are removed through data cleaning and preprocessing. After the historical load data is collected, load change calculation is started. In this process, the load change rate is obtained by calculating the load difference between adjacent time points.

[0031] After obtaining the load change rate, fast Fourier transform (FFT) is performed to convert the time-domain load change data into frequency-domain signals. The frequency spectrum energy distribution proportion is extracted using FFT, that is, the energy proportion of each frequency component in the total load fluctuation. Through FFT, the frequency distribution of the load fluctuation is obtained. For example, it is found that low-frequency fluctuations (0-1 Hz) account for 70% of the total energy, and high-frequency fluctuations (1-5 Hz) account for 30%.

[0032] After frequency spectrum analysis, wavelet analysis is performed according to the periodic characteristics of the load fluctuation (such as daily fluctuations, seasonal changes, etc.) combined with the frequency spectrum energy distribution proportion. Wavelet analysis is a multi-scale signal processing method that can extract signal features at multiple time scales. In this step, the main period of the load fluctuation is first identified. For example, the power grid load usually shows obvious fluctuations within 24 hours, and appropriate decomposition layers are set according to these periodic characteristics. For short-period load fluctuations (such as daily fluctuations within 24 hours), fewer decomposition layers (such as 3 layers) are selected to capture higher-frequency fluctuations. For long-period fluctuations (such as seasonal fluctuations), more decomposition layers (such as 6 layers or more) are selected to capture low-frequency fluctuations.

[0033] According to the set decomposition layer, the load fluctuation signal is decomposed by using wavelet transform. Each layer of wavelet transform extracts different frequency components of the signal, and analyzes the load fluctuation on multiple scales. For example, through 3-layer wavelet decomposition, short-term changes (high-frequency components) and long-term trend changes (low-frequency components) of daily fluctuations are identified respectively.

[0034] Finally, the frequency components obtained by wavelet analysis are synthesized by frequency spectrum decomposition reconstruction to reconstruct the complete load fluctuation signal. In this way, high-frequency fluctuation components and low-frequency fluctuation components are identified and extracted. The high-frequency fluctuation component reflects the short-term fluctuation in the power grid, such as the instantaneous change of load or the start-stop of equipment; while the low-frequency fluctuation component represents the long-term change of the power grid load, such as the seasonal fluctuation of the power grid or the fluctuation during the peak electricity consumption period.

[0035] Step S200: Real-time monitoring of the high-voltage cascade energy storage system is performed to obtain a multi-unit state parameter set, and dynamic balancing analysis is performed according to the multi-level power instruction set and the multi-unit state parameter set to generate a multi-level collaborative scheduling instruction.

[0036] In the embodiments of the present application, the high-voltage cascade energy storage system is first monitored in real time, and the operation data of each energy storage unit is collected in real time. These data include the state parameter set of each energy storage unit, wherein the main parameters include the state of charge (SOC) of the battery, the temperature gradient and the voltage deviation rate, etc. Through this process, a multi-unit state parameter set is obtained. Among them, SOC represents the current charge level of the battery; the temperature gradient represents the temperature difference between different parts inside the energy storage unit; the voltage deviation rate refers to the deviation between the actual voltage of the energy storage unit and the designed voltage.

[0037] Next, dynamic balancing analysis is performed according to the multi-level power instruction set and the multi-unit state parameter set. In this process, first, the multi-level power instruction set is loaded and read in real time to determine the real-time frequency spectrum component of the power grid load. Then, the multi-level power instruction set and the multi-unit state parameter set (such as SOC, temperature gradient, voltage deviation rate) are synchronized to the health degree analysis unit to generate multiple health degree parameters. Based on these parameters, state transition evaluation is performed to generate a state evaluation matrix, and the state evaluation matrix is used for dynamic balancing analysis to obtain a state balancing coefficient. Then, the energy storage system is traversed to perform multi-level scheduling identification and generate a multi-level scheduling data tag. Finally, scheduling conflict analysis is performed to update the scheduling data tag, and finally a multi-level collaborative scheduling instruction is generated.

[0038] Further, in the method provided by the embodiments of the present application, the dynamic balancing analysis according to the multi-level power instruction set and the multi-unit state parameter set to generate the multi-level collaborative scheduling instruction further comprises:

[0039] The multi-stage power instruction set is loaded for real-time reading to determine real-time spectrum components; a health degree analysis unit is constructed, the multi-stage power instruction set and the multi-unit state parameter set are synchronized to the health degree analysis unit to generate a plurality of health degree parameters; state transition evaluation is performed on the multi-stage power instruction set and the multi-unit state parameter set according to the plurality of health degree parameters to generate a state evaluation matrix; dynamic balance analysis is performed according to the state evaluation matrix to generate a state balance coefficient; multi-stage scheduling identification is performed on the high-voltage cascade energy storage system based on the state balance coefficient to generate a multi-stage scheduling data tag; scheduling conflict analysis is performed based on the multi-stage scheduling data tag, and the multi-stage scheduling data tag is updated according to the analysis result to generate the multi-stage collaborative scheduling instruction.

[0040] In the embodiments of the present application, first, the multi-stage power instruction set is loaded and read in real time, and the multi-stage power instruction set is analyzed by fast Fourier transform (FFT) to convert load data in the time domain into frequency domain signals, thereby determining real-time spectrum components.

[0041] Next, a health degree analysis unit is constructed, and the multi-stage power instruction set and the multi-unit state parameter set are synchronized to the health degree analysis unit. Specifically, a health degree analysis thread is created by the main controller of the high-voltage cascade energy storage system, and the multi-stage power instruction set and the multi-unit state parameter set are synchronized to the health degree analysis unit. In this process, first, a triple data buffering mechanism is set to ensure efficient and accurate data processing. The first buffer layer temporarily stores the multi-stage power instruction set and the multi-unit state parameter set to generate a first buffer result; the second buffer layer aligns the time stamp of the first buffer result to ensure the time synchronization of the data and generates a second buffer result; and the third buffer layer performs unit matching identification on the second buffer result to generate a third buffer result. Finally, priority parameter update compensation is performed based on the third buffer result to generate a plurality of health degree parameters.

[0042] Subsequently, state transition evaluation is performed on the multi-stage power instruction set and the multi-unit state parameter set according to the plurality of health degree parameters. In this step, first, dynamic normalization is performed on the health degree parameters to obtain a standard health degree parameter vector. Then, by calling unoccurred events, an initial probability table is obtained, and state transition rules are defined to construct a discrete state space. The multi-stage power instruction set and the multi-unit state parameter set are mapped into the discrete state space for state transition analysis, and node transition statistics are performed according to the initial probability table to obtain node transition frequency information. Combined with the standard health degree parameter vector and the node transition frequency information, transfer path backtracking matching is performed to determine the distribution data of state transition. Finally, these data are structured and stored, and a state evaluation matrix is constructed.

[0043] Next, dynamic balance analysis is performed according to the state evaluation matrix. The state balance coefficient is a quantitative value obtained by dynamic balance analysis of the state evaluation matrix of the energy storage unit. Specifically, the health status distribution of the current system is evaluated by counting the transition frequency between each state (such as healthy, pre-warning, and failure). When calculating the state balance coefficient, first, the state transition frequency is calculated, and the number of transitions between each health state is counted. Then, the state balance degree is calculated, i.e., the ratio of the transition frequency of each health state to the total transition frequency, to obtain the stability of each state. For example, the health state stability = health to pre-warning transition frequency + health to failure transition frequency divided by the total transition frequency. The pre-warning state stability and the failure state stability are calculated in a similar manner to the health state stability. Finally, the health state stability, the pre-warning state stability, and the failure state stability are averaged to generate the state balance coefficient.

[0044] Once the state balance coefficient is obtained, a multi-level scheduling label is generated based on the coefficient. The purpose of this process is to determine the scheduling priority of each energy storage unit based on the health status of the energy storage unit and the state balance coefficient. Energy storage units with higher state balance coefficients generally have more stable health states and should be prioritized for scheduling. By traversing each energy storage unit in the system, a multi-level scheduling data label is generated, which identifies the scheduling priority, scheduling period, and charging and discharging tasks of each energy storage unit. The label divides energy storage units into different priority levels based on their health status and scheduling needs to ensure that the healthiest energy storage units can respond to grid load demand first.

[0045] Next, scheduling conflict analysis is performed based on the generated multi-level scheduling data label. In this analysis process, it is detected whether there is a resource conflict or scheduling conflict between energy storage units. For example, if two energy storage units are scheduled to perform charging tasks at the same time, and their charging demand exceeds the maximum power output of the system, the system will find a scheduling conflict. The key to conflict analysis is to determine whether there is an overload or competition when multiple energy storage units are running in parallel, especially when the battery charging and discharging capacity is limited. Scheduling conflicts can affect the stability of the entire system.

[0046] In the scheduling conflict analysis, the charging and discharging tasks of the energy storage units are adjusted according to the scheduling priority. For example, if energy storage unit A and energy storage unit B have a power conflict at the same time, and energy storage unit A has a better health status and a higher priority, the system will adjust the charging task of energy storage unit B, delay its charging task, or redistribute the power load, thereby solving the scheduling conflict.

[0047] Finally, the multi-level scheduling data tag is updated according to the result of the scheduling conflict analysis, and the final multi-level collaborative scheduling instruction is generated. The updated data tag adjusts the scheduling tasks and priorities of the energy storage units, ensuring that the system can be flexibly scheduled according to the current health status, load demand and scheduling priority of the energy storage units. Based on these adjustments, the multi-level collaborative scheduling instruction is generated, which specifies specific charging and discharging tasks, scheduling time periods, and related charging and discharging powers for each energy storage unit.

[0048] Further, the method provided by the application embodiment further comprises:

[0049] The health degree analysis thread is created based on the main controller of the high-voltage cascaded energy storage system, the health degree analysis unit is obtained, and the multi-level power instruction set and the multi-unit state parameter set are synchronized to the health degree analysis unit: S1: a three-level data buffering mechanism is set, which includes a first buffering layer, a second buffering layer and a third buffering layer; S2: the multi-level power instruction set and the multi-unit state parameter set are temporarily stored through the first buffering layer to generate a first buffering result; S3: the first buffering result is timestamped through the second buffering layer to generate a second buffering result; S4: the second buffering result is matched and identified through the third buffering layer to generate a third buffering result; and S5: the third buffering result is used to update and compensate the priority parameters to generate the plurality of health degree parameters.

[0050] In the application embodiment, first, the health degree analysis thread is created based on the main controller of the high-voltage cascaded energy storage system, and the health degree analysis thread is responsible for real-time analysis of the health status of each energy storage unit. The health degree analysis unit receives the multi-level power instruction set and the multi-unit state parameter set obtained from the energy storage system.

[0051] To ensure smooth processing and synchronization of data, a triple data buffering mechanism is set up, including a first buffer layer, a second buffer layer, and a third buffer layer. In the first buffer layer, a multi-level power instruction set and a multi-unit state parameter set are temporarily stored to generate a first buffer result, ensuring that the system can quickly receive and store real-time data from the grid dispatch center and the energy storage system. The second buffer layer aligns the timestamps of the first buffer result. Since the multi-level power instruction set and the multi-unit state parameter set may be received at different time points, the timestamp alignment technology ensures that they can be synchronized to generate a second buffer result, ensuring the consistency of power instructions and state parameters in time. The third buffer layer performs unit matching identification on the second buffer result to generate a third buffer result. The core task at this stage is to match the power instruction of each energy storage unit with its corresponding state parameter, ensuring that the dispatch instruction of each energy storage unit is synchronized with its operating state, avoiding misdispatching or overloading.

[0052] Subsequently, based on the third buffer result, priority parameter update compensation is performed. When the update frequency of the power instruction is higher than that of the state parameter, the historical data interpolation compensation method is used to fill in the gap of state parameter update. The interpolation compensation method uses linear interpolation or polynomial interpolation technology to calculate the state value at the intermediate time based on the state data at the last time and the current time. This compensation step ensures that even if the state parameter is not updated in real time, the charging and discharging instruction can be dispatched based on the latest health degree evaluation.

[0053] After that, the compensated state parameter is used to calculate the health degree parameter. First, the electrical health degree parameter is calculated, which is based on the voltage deviation rate and the actual power tracking error. The voltage deviation rate reflects the difference between the actual voltage of the energy storage unit and the target voltage, and the power tracking error represents the deviation degree of the energy storage unit in tracking the grid load. In the calculation, the voltage deviation rate and the actual power tracking error are first normalized to ensure that they are in the same quantization range. The specific method is to divide the voltage deviation rate and the power tracking error by their maximum values (for example, the maximum value of the voltage deviation rate is 10%, and the maximum value of the power tracking error is 5%). Then, the normalized data is weighted to obtain the electrical health degree score of the energy storage unit. The weights of the voltage deviation rate and the actual power tracking error are pre-set, which are 0.7 and 0.3 respectively, and the electrical health degree parameter is obtained by calculation.

[0054] Secondly, the thermal health degree parameter is calculated, and the energy storage unit is scored in multiple levels according to the temperature gradient and the local temperature rise rate. The temperature gradient represents the temperature difference between different parts of the energy storage unit, and an excessive temperature difference may cause local overheating or efficiency reduction of the battery. The local temperature rise rate reflects the temperature rise speed of the local area of the energy storage unit during the charging and discharging process, and a fast temperature rise may damage the battery. The thermal health degree score is generated according to these parameters. Assuming that the temperature gradient of the energy storage unit is 3℃ and the temperature rise rate is 0.5℃ / s, a multi-level scoring mechanism is used, and the thresholds of the temperature gradient and the temperature rise rate are set, which correspond to different health degree scores. For example, a temperature gradient of 3℃ may get 80 points, and a temperature rise rate of 0.5℃ / s gets 70 points. The thermal health degree parameter is calculated by weighting the two scores according to the pre-set weighting coefficients. This parameter reflects the health condition of the energy storage unit in terms of thermal management.

[0055] Finally, the aging health degree parameter is calculated, which is a life attenuation index generated by combining the cycle number increment of the energy storage unit and the resistance change trend. The increase of the internal resistance is an important sign of battery degradation, and the cycle number increment indicates the frequency of use of the energy storage unit. Assuming that the cycle number increment of the energy storage unit is 500 times and the resistance change trend is 15%, the cycle number increment and the resistance change trend are first normalized, and then the weights of these two factors are set to 0.5 and 0.5 respectively. Through weighted calculation, the life attenuation index is obtained, which is the aging health degree parameter.

[0056] Through the above calculation, multiple health degree parameters are obtained, including the electrical health degree parameter, the thermal health degree parameter and the aging health degree parameter.

[0057] Further, the method provided by the application embodiment further comprises:

[0058] The multiple health degree parameters are dynamically normalized to determine a standard health degree parameter vector; the state transition probability initialization is called to obtain an initial probability table, and the state transition rules are defined to construct a discrete state space in combination with the initial probability table; the multiple-level power instruction set and the multiple-unit state parameter set are mapped to the discrete state space for state transition, the node transition frequency information is obtained by performing node transition statistics according to the initial probability table; the state transition distribution data is determined by performing transition path backtracking matching according to the standard health degree parameter vector in combination with the node transition frequency information; and the state transition distribution data is stored in a structured manner to construct the state evaluation matrix.

[0059] In the embodiments of the present application, first, multiple health parameters are dynamically normalized, and the electrical health parameter, the thermal health parameter and the aging health parameter are unified into the same quantization range. Specifically, by normalizing each health parameter, their values are mapped to a standard range (usually between 0 and 1). For example, if the maximum value of the thermal health parameter is 100 points, and the current thermal health parameter of the energy storage unit is 80 points, then through normalization, the normalized value of the thermal health parameter is 0.8. The same method applies to the electrical health parameter and the aging health parameter. After normalization, these values are summarized into a standard health parameter vector, such as [electrical score, thermal score, aging score].

[0060] Next, the state transition probability initialization is performed for the events that do not occur, and the initial probability table is established for subsequent state transition analysis. The initial probability table is initialized according to the frequency of state changes of the energy storage unit in the historical operation data. In the historical data, the cases of the energy storage unit transferring from one health state to another health state are analyzed. For example, it is found that the probability of the energy storage unit transferring from the health state to the warning state is 30%, and the probability of transferring from the warning state to the failure state is 10%, and the probability table is initialized according to these data to provide probability support for subsequent state transition process. In this way, the initial probability of the energy storage unit transferring between different states is obtained.

[0061] Subsequently, the state conversion rule is defined, which determines the state transition of the energy storage unit according to the health parameters. Specifically, for each health state, there is a clear conversion condition. For example, the condition for the energy storage unit to transfer from the "healthy" state to the "warning" state is that any health score is below the first level threshold for three consecutive times; the condition for transferring from the "warning" state to the "failure" state is that the key parameters such as internal resistance or temperature rise exceed the second level threshold and continue to exceed for a certain time; and the condition for transferring from the "failure" state to the "healthy" state is to undergo manual reset and pass the benchmark test. According to these conversion rules, combined with the initial probability table, a discrete state space is further constructed. The discrete state space includes three types of nodes, i.e., the health state, the warning state and the failure state, and each energy storage unit in the system can be in any one of the three states.

[0062] Subsequently, the multi-level power command set and the multi-unit state parameter set are mapped to a discrete state space for state transition. This process involves mapping the actual state of each energy storage unit to a corresponding node in the discrete state space based on the health parameters (electrical health parameter, thermal health parameter, and aging health parameter) of the energy storage unit and the state transition rules. For example, when the electrical health parameter of the energy storage unit is low and the temperature gradient is high, it will be transferred from the "healthy" state to the "warning" state; when the internal resistance increases significantly and exceeds the secondary threshold, the state of the energy storage unit will be transferred from "warning" to "failure". According to the initial probability table, the probability of state transition is calculated according to the health parameters of the current energy storage unit, and the simulated transition of the state is carried out.

[0063] After completing the state transition, the node transition frequency information is obtained according to the initial probability table. This process involves tracking and recording the state transition of the energy storage unit, and counting the transition frequency between each node (healthy state, warning state, failure state). For example, if the energy storage unit has been transferred from the "healthy" state to the "warning" state 20 times, and from the "warning" state to the "failure" state 5 times, these transition frequencies are recorded to obtain the node transition frequency information.

[0064] After obtaining the node transition frequency information, the transition path backtracking matching is performed according to the standard health parameter vector combined with the node transition frequency information to determine the state transition distribution data. Transition path backtracking matching refers to backtracking the transition path between each state according to historical state transition data and health parameters, and analyzing the frequency of different paths. For example, if the transition frequency from the "healthy" state to the "warning" state of the energy storage unit is high, the path is analyzed in detail, and the future transition mode is predicted according to the change of the health parameter. Through backtracking analysis, the rule of the health state change of the energy storage unit is found out, and the possible path of the state transition is determined.

[0065] Finally, the state transition distribution data is stored in a structured manner, and a state evaluation matrix is constructed based on these data. The state evaluation matrix is a matrix containing the transition probability and frequency between each health state. Through this matrix, the future possible state change of the energy storage unit is evaluated. The data in the matrix reflects the transition frequency and probability of the energy storage unit from one state to another, and provides data support for scheduling decisions. For example, the matrix may contain a transition probability of 30% from the "healthy" state to the "warning" state, and a transition probability of 10% from the "warning" state to the "failure" state.

[0066] Step S300: mapping the multi-level collaborative scheduling instruction to the high-voltage cascade energy storage system for predictive control, converting it into a charging and discharging control instruction set, executing the charging and discharging control instruction set for monitoring and recording, generating system response data, and performing closed-loop feedback correction on the multi-level collaborative scheduling instruction to obtain a system optimization control parameter set.

[0067] In the embodiments of the present application, the multi-level collaborative scheduling instruction is first mapped to the high-voltage cascade energy storage system for predictive control. In this process, the grid interaction parameters are obtained through grid point detection equipment for grid interaction recording, and the state constraint mapping is performed using these parameters to construct the grid constraint state space. Then, interval rolling is performed in the grid constraint state space, a rolling time window is set, and the multi-level collaborative scheduling instruction is mapped to the high-voltage cascade energy storage system for scheduling prediction within the time window to generate a collaborative scheduling prediction result. Subsequently, based on the prediction result, a multi-objective optimization is performed to generate an optimization solution set, and the solution set is compensated with charging and discharging data according to the control step length to finally generate a charging and discharging control instruction set.

[0068] Next, the charging and discharging control instruction set is executed for monitoring and recording to generate system response data for closed-loop feedback correction of the multi-level collaborative scheduling instruction. In this process, the charging and discharging control instruction set is first executed, multi-dimensional monitoring is performed through the driving power conversion unit to generate multi-dimensional monitoring data, including the charging and discharging power of the energy storage unit, battery health parameters, temperature, and other key information. Then, the monitoring data is associated with a time stamp to generate system response data, which reflects the actual performance of the energy storage unit during the charging and discharging process. Based on the system response data, scheduling deviation analysis is performed to identify the differences between the scheduling target and the actual execution, and scheduling deviation parameters are generated. Then, parameter stability analysis is performed according to these deviation parameters to generate parameter stability coefficients to evaluate the stability of the scheduling result. By analyzing these stability parameters, the scheduling deviation parameters are prioritized to determine which deviations need to be corrected first. Finally, the scheduling deviation correction sequence is activated, and closed-loop feedback correction is performed through parameter correction instructions to optimize the scheduling strategy of the energy storage unit. Through this process, the system optimization control parameter set is continuously optimized to ensure that the energy storage system can operate efficiently and stably while responding to changes in grid load and the state of the energy storage unit.

[0069] Further, the method provided by the embodiments of the present application, wherein the multi-level collaborative scheduling instruction is mapped to the high-voltage cascade energy storage system for predictive control, and converted into a charging and discharging control instruction set, further comprises:

[0070] The grid interaction record is recorded by a grid point detection device, grid interaction parameters are obtained, state constraint mapping is performed according to the grid interaction parameters, and a grid constraint state space is constructed; interval rolling promotion is performed based on the grid constraint state space, and a rolling time window is set; the multi-stage collaborative scheduling instruction is mapped to the high-voltage cascade energy storage system according to the grid constraint state space, scheduling prediction is performed according to the rolling time window, and a collaborative scheduling prediction result is generated; multi-objective optimization is performed based on the collaborative scheduling prediction result, and an optimization solution set is generated; the optimization solution set is converted by injecting charge and discharge compensation data according to a control step, and the charge and discharge control instruction set is generated.

[0071] In the embodiments of the present application, first, the grid interaction record is recorded by a grid point detection device, and real-time interaction parameters of the grid, such as voltage, current, frequency, etc. are obtained. These grid interaction parameters are used to describe the interaction between the grid and the energy storage system.

[0072] Then, the grid constraint state space is constructed by state constraint mapping. The process of state constraint mapping is to convert the actual grid state (such as voltage range, current limit, etc.) into a state space containing various constraint conditions according to the operation standard and load demand of the grid. This state space describes the working boundary of the grid, ensuring that the scheduling decision of the energy storage system will not cause the grid to overload or operate unstably. The grid constraint state space includes the maximum load, minimum voltage, maximum current, etc. of the grid, and provides clear restrictions for the scheduling of the energy storage system.

[0073] Then, interval rolling promotion is performed based on the grid constraint state space, and a rolling time window is set. The rolling time window is a fixed length time period, which is used to dynamically adjust and optimize the scheduling scheme of the energy storage system. In each time window, the load demand of the grid, the charge and discharge capacity of the energy storage system and the constraint conditions of the grid will be re-evaluated, and updated scheduling instructions will be generated according to the new data. At the end of each window, the new grid interaction parameters and state space are recalculated, and a new scheduling plan is generated. Through this rolling promotion mechanism, the operation strategy of the energy storage system can be continuously adjusted to ensure that it can always respond in time according to the load fluctuation of the grid.

[0074] Based on the above power grid constraint state space, the multi-level collaborative scheduling instruction is mapped to the high-voltage cascade energy storage system, and scheduling prediction is performed according to the rolling time window. In this step, a predictive control model is used for scheduling prediction. The predictive control model is established based on historical data of the system, performance characteristics of the energy storage unit, and changes in power grid load. The model uses an LSTM neural network to model the relationship between power grid load changes, energy storage unit states, and scheduling instructions. By inputting historical power grid data and energy storage system states, the model can predict the energy storage unit charging and discharging requirements in future periods, thereby generating collaborative scheduling prediction results.

[0075] Then, based on the collaborative scheduling prediction results, multi-objective optimization is performed to generate an optimization solution set. In the multi-objective optimization process, multiple factors such as the health of the energy storage unit, power grid demand, charging and discharging efficiency, etc. are considered, and techniques such as or particle swarm optimization are used to solve the optimal scheduling scheme. Through this process, an optimization solution set is obtained.

[0076] Finally, the optimization solution set is converted into charging and discharging compensation data according to the control step size. The control step size is a fixed value set in advance, which is used to control the adjustment range of the charging and discharging power of the energy storage unit. Specifically, the control step size determines the size of the power change of the energy storage unit when adjusting the charging and discharging power each time. By applying the control step size to the optimization solution set, the charging and discharging tasks of the energy storage unit are fine-tuned according to real-time power grid load changes and the state of the energy storage unit. This adjustment ensures that the energy storage system can respond smoothly and accurately to power grid load fluctuations, and ensures that each energy storage unit completes the predetermined charging and discharging task without exceeding its capacity range. Finally, the adjusted instructions are summarized into a charging and discharging control instruction set.

[0077] Further, the method provided by the application embodiment further comprises:

[0078] The driving power conversion unit performs multi-dimensional monitoring according to the charging and discharging control instruction set, generates multi-dimensional monitoring data, associates time stamps based on the multi-dimensional monitoring data, generates system response data, performs scheduling deviation analysis on the multi-level collaborative scheduling instruction according to the system response data, generates scheduling deviation parameters, performs parameter stability analysis according to the scheduling deviation parameters, generates a parameter stability coefficient, prioritizes the scheduling deviation parameters according to the parameter stability coefficient, determines a scheduling deviation correction sequence, activates a parameter correction instruction, and performs closed-loop feedback correction according to the scheduling deviation correction sequence through the parameter correction instruction to generate the system optimization control parameter set.

[0079] In the embodiments of the present application, first, the power conversion unit is driven to execute the charge and discharge control instructions for multi-dimensional monitoring. In this process, first, the voltage of the high-voltage cascaded energy storage system is monitored to obtain transient voltage fluctuation data. Then, current monitoring is performed to obtain multi-branch current data through multi-branch current monitoring. At the same time, the temperature of the energy storage system is monitored, and a two-dimensional thermal map is generated to show the temperature distribution of the energy storage unit during the charge and discharge process. Finally, the transient voltage fluctuation data, multi-branch current data and two-dimensional thermal map are associated and integrated to generate complete multi-dimensional monitoring data.

[0080] Next, the multi-dimensional monitoring data is associated with the time stamp, and through time synchronization technology, the system ensures that each monitoring data is consistent with the corresponding time stamp, and generates system response data reflecting the actual operation of the energy storage system.

[0081] Subsequently, the multi-stage collaborative scheduling instructions are scheduled for deviation analysis. In this process, the deviation between the actual execution result and the predetermined scheduling is compared through difference analysis method (such as error calculation) to generate scheduling deviation parameters. The deviation parameters are quantitative description of the difference between the energy storage unit in executing the charge and discharge task and the predetermined target. For example, the system analyzes the error between the actual charging power of the energy storage unit and the target charging power, or the time delay in the charge and discharge process.

[0082] Then, based on the scheduling deviation parameters, parameter stability analysis is performed. Stability analysis is performed by time series analysis method. Specifically, by comparing historical scheduling deviation and current deviation, moving average method is used to smooth the scheduling deviation to reduce noise fluctuation in the data. Through this method, the parameter stability coefficient is generated. Assuming that the scheduling deviations at the past five time points are 3%, 5%, 2%, 4% and 3%, first, the moving average of these deviations is calculated (for example: the average value of the past five time points is 3.4%). Then, by calculating the standard deviation, the stability of the deviation fluctuation is reflected. Smaller standard deviation indicates that the scheduling is more stable, while larger standard deviation indicates that there is greater fluctuation in the system scheduling. Assuming that the standard deviation is 1.2%, then the parameter stability coefficient = 1 - standard deviation / average value = 1-1.2 / 3.4 = 0.65.

[0083] Subsequently, the scheduling deviation parameters are prioritized according to the parameter stability coefficients. In this process, a priority sorting algorithm is used to sort the scheduling deviations according to the size of the parameter stability coefficients. The smaller the parameter stability coefficient, the more unstable the scheduling process and the larger the deviation, so these deviations are prioritized for correction. Through priority identification, it is determined which scheduling deviations should be prioritized for correction. After priority identification, a scheduling deviation correction sequence is generated, and parameter correction instructions are generated according to the sequence. The scheduling deviation correction sequence determines the deviation items that need to be corrected and the order of correction according to the results of the priority identification. By adjusting the scheduling deviation one by one according to this order, it is ensured that the deviation that has the greatest impact on system stability is corrected first. Through the parameter correction instructions, the charging and discharging tasks of the energy storage unit are adjusted in detail. For example, if there is a large deviation in the charging power of the battery, the deviation is eliminated by adjusting the charging power or extending the charging time. At this time, the correction instructions will refine the charging or discharging operation of the energy storage unit according to the analysis results described above.

[0084] Finally, through closed-loop feedback correction, further optimization is performed according to the corrected scheduling parameters. Closed-loop feedback correction is achieved through a real-time feedback mechanism. The corrected instructions are applied to the energy storage system, and the scheduling execution is re-evaluated according to the new scheduling. If there is still a deviation after correction, the task is adjusted again through the feedback mechanism until the charging and discharging tasks of the energy storage unit completely match the predetermined target. After this series of correction and feedback processes, the final system optimization control parameter set is generated. This parameter set provides optimized control instructions for the charging and discharging tasks of the energy storage unit, ensuring that the energy storage system can respond to grid demand stably and efficiently when the grid load changes, and maximizing the operating efficiency of the system.

[0085] Further, the method provided by the application embodiment further comprises:

[0086] driving the power conversion unit to execute the charging and discharging control instructions to perform voltage monitoring on the high-voltage cascaded energy storage system to obtain transient voltage fluctuation data; driving the power conversion unit to execute the charging and discharging control instructions to perform current monitoring on the high-voltage cascaded energy storage system to obtain multi-branch current data; driving the power conversion unit to execute the charging and discharging control instructions to perform temperature monitoring on the high-voltage cascaded energy storage system to draw a two-dimensional thermal diagram; and correlating and integrating the transient voltage fluctuation data, the multi-branch current data, and the two-dimensional thermal diagram to generate the multi-dimensional monitoring data.

[0087] In the embodiments of the present application, first, the power conversion unit is driven to execute the charge and discharge control instruction, the voltage of the high-voltage cascaded energy storage system is monitored, the voltage data of the energy storage unit is collected in real time through the voltage sensor, and the transient voltage fluctuation data is obtained. In this process, the voltage sensor measures the voltage change in real time during the charging and discharging process, and captures the voltage fluctuation caused by the change of the power grid load.

[0088] Then, the power conversion unit is continuously driven to execute the charge and discharge control instruction, the current of the high-voltage cascaded energy storage system is monitored, the current data of the energy storage unit is collected through the current sensor, and the multi-branch current data is obtained. Through the current sensor, the current change of each branch is recorded in real time, reflecting the current load condition of the energy storage unit under different charge and discharge paths.

[0089] Then, the charge and discharge control instruction is continuously executed, the temperature of the high-voltage cascaded energy storage system is monitored, the temperature data of the energy storage unit is obtained through the temperature sensor, and a two-dimensional thermal map is generated. The thermal map shows the temperature distribution of each region of the energy storage unit.

[0090] Finally, all the collected transient voltage fluctuation data, multi-branch current data and two-dimensional thermal map are associated and integrated. In this process, the voltage, current and temperature data are synchronized with the time stamp through data fusion technology, ensuring the time consistency of the data, and finally generating complete multi-dimensional monitoring data.

[0091] In the embodiments of the present application, as described above, the embodiments of the present application have at least the following technical effects:

[0092] The present application receives the charge and discharge instruction of the power grid dispatching center, decomposes the charge and discharge instruction according to the load fluctuation characteristics, generates a multi-level power instruction set, monitors the high-voltage cascaded energy storage system in real time, obtains a multi-unit state parameter set, performs dynamic balance analysis according to the multi-level power instruction set and the multi-unit state parameter set, generates a multi-level collaborative scheduling instruction, maps the multi-level collaborative scheduling instruction to the high-voltage cascaded energy storage system for predictive control, converts it into a charge and discharge control instruction set, executes the charge and discharge control instruction set for monitoring and recording, generates system response data for closed-loop feedback correction of the multi-level collaborative scheduling instruction, and obtains a system optimization control parameter set. The present application solves the technical problem that the charge and discharge control instruction in the prior art cannot accurately adapt to the load fluctuation of the power grid. Through decomposition of the charge and discharge instruction and dynamic balance analysis combined with the multi-unit state parameter, the effect of optimizing the scheduling instruction is achieved, thereby improving the system response speed and stability.

[0093] Embodiment two, based on the same inventive concept as the control method of the high-voltage cascaded energy storage system in the foregoing embodiments, such as Figure 2As shown, the application provides a control device for a high-voltage cascade energy storage system, and the device and method embodiments in the application are based on the same inventive concept. The device comprises:

[0094] The instruction decomposition module 11 is configured to receive the charge and discharge instruction from the power grid dispatching center, decompose the charge and discharge instruction based on the load fluctuation characteristics, and generate a multi-stage power instruction set. The real-time monitoring module 12 is configured to monitor the high-voltage cascade energy storage system in real time, obtain a multi-unit state parameter set, perform dynamic balancing analysis based on the multi-stage power instruction set and the multi-unit state parameter set, and generate a multi-stage collaborative scheduling instruction. The feedback correction module 13 is configured to map the multi-stage collaborative scheduling instruction to the high-voltage cascade energy storage system for predictive control, convert it into a charge and discharge control instruction set, execute the charge and discharge control instruction set for monitoring and recording, generate system response data, perform closed-loop feedback correction on the multi-stage collaborative scheduling instruction, and obtain a system optimization control parameter set.

[0095] Further, the device is also configured to implement the following functions:

[0096] The load fluctuation characteristics are subjected to frequency spectrum identification to determine high-frequency fluctuation components and low-frequency fluctuation components. The charge and discharge instruction is subjected to response analysis on the high-voltage cascade energy storage system to obtain a first response speed parameter and a second response speed parameter. A first-stage energy storage unit cluster is determined according to the first response speed parameter, and a second-stage energy storage unit cluster is determined according to the second response speed parameter. The high-frequency fluctuation components are distributed to the first-stage energy storage unit cluster, and the low-frequency fluctuation components are distributed to the second-stage energy storage unit cluster, thereby generating the multi-stage power instruction set.

[0097] Further, the device is also configured to implement the following functions:

[0098] The historical load data of the high-voltage cascade energy storage system is collected for change calculation to obtain a load change rate. Fast Fourier transform is performed according to the load change rate to extract a frequency spectrum energy distribution proportion value. Wavelet analysis is performed according to the fluctuation period and the frequency spectrum energy distribution proportion value to set a decomposition layer number. The load fluctuation characteristics are subjected to frequency spectrum decomposition and reconstruction according to the decomposition layer number to identify the high-frequency fluctuation components and the low-frequency fluctuation components.

[0099] Further, the device is also configured to implement the following functions:

[0100] loading the multi-stage power instruction set for real-time reading to determine real-time spectrum components; constructing a health degree analysis unit, synchronizing the multi-stage power instruction set and the multi-unit state parameter set to the health degree analysis unit to generate a plurality of health degree parameters; performing state transition evaluation on the multi-stage power instruction set and the multi-unit state parameter set according to the plurality of health degree parameters to generate a state evaluation matrix; performing dynamic balance analysis according to the state evaluation matrix to generate a state balance coefficient; performing multi-stage scheduling identification on the high-voltage cascade energy storage system based on the state balance coefficient to generate a multi-stage scheduling data tag; performing scheduling conflict analysis based on the multi-stage scheduling data tag, updating the multi-stage scheduling data tag according to the analysis result, and generating the multi-stage collaborative scheduling instruction.

[0101] Further, the device is also used to realize the following functions:

[0102] Based on the main controller of the high-voltage cascade energy storage system, a health degree analysis thread is created, and a health degree analysis unit is obtained, and the multi-stage power instruction set and the multi-unit state parameter set are synchronized to the health degree analysis unit: S1: Set up a triple data buffering mechanism, which includes a first buffer layer, a second buffer layer and a third buffer layer; S2: Temporarily store the multi-stage power instruction set and the multi-unit state parameter set through the first buffer layer to generate a first buffer result; S3: Align the first buffer result with the time stamp through the second buffer layer to generate a second buffer result; S4: Perform unit matching identification on the second buffer result through the third buffer layer to generate a third buffer result; S5: Perform priority parameter update compensation based on the third buffer result to generate the plurality of health degree parameters.

[0103] Further, the device is also used to realize the following functions:

[0104] Perform dynamic normalization processing on the plurality of health degree parameters to determine a standard health degree parameter vector; retrieve the state transition probability initialization without an event to obtain an initial probability table, define a state transition rule, and construct a discrete state space combined with the initial probability table; map the multi-stage power instruction set and the multi-unit state parameter set to the discrete state space for state transition, perform node transition statistics according to the initial probability table to obtain node transition frequency information; perform transition path backtracking matching according to the standard health degree parameter vector combined with the node transition frequency information to determine state transition distribution data; perform structured storage on the state transition distribution data to construct the state evaluation matrix.

[0105] Further, the device is also used to realize the following functions:

[0106] The grid interaction parameters are obtained through grid interaction record of a grid point detection device, state constraint mapping is performed according to the grid interaction parameters, and a grid constraint state space is constructed; interval rolling is promoted based on the grid constraint state space, a rolling time window is set; the multi-stage collaborative scheduling instruction is mapped to the high-voltage cascade energy storage system according to the grid constraint state space, scheduling prediction is performed according to the rolling time window, and a collaborative scheduling prediction result is generated; multi-objective optimization is performed based on the collaborative scheduling prediction result, and an optimization solution set is generated; the optimization solution set is converted by injecting charge and discharge compensation data according to a control step, and the charge and discharge control instruction set is generated.

[0107] Further, the device is also used to implement the following functions:

[0108] The power conversion unit is driven to execute the charge and discharge control instruction to generate multi-dimensional monitoring data; the multi-dimensional monitoring data is associated with a time stamp to generate system response data; the multi-stage collaborative scheduling instruction is analyzed for scheduling deviation according to the system response data to generate scheduling deviation parameters; the scheduling deviation parameters are analyzed for parameter stability to generate a parameter stability coefficient, the scheduling deviation parameters are identified for priority according to the parameter stability coefficient to determine a scheduling deviation correction sequence; a parameter correction instruction is activated, and the system optimization control parameter set is generated through closed-loop feedback correction of the parameter correction instruction according to the scheduling deviation correction sequence.

[0109] Further, the device is also used to implement the following functions:

[0110] The power conversion unit is driven to execute the charge and discharge control instruction to perform voltage monitoring on the high-voltage cascade energy storage system to obtain transient voltage fluctuation data; the power conversion unit is driven to execute the charge and discharge control instruction to perform current monitoring on the high-voltage cascade energy storage system to obtain multi-branch current data; the power conversion unit is driven to execute the charge and discharge control instruction to perform temperature monitoring on the high-voltage cascade energy storage system to draw a two-dimensional thermal diagram; the transient voltage fluctuation data, the multi-branch current data, and the two-dimensional thermal diagram are associated and integrated to generate the multi-dimensional monitoring data.

[0111] It should be noted that the above sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. The above describes a specific embodiment of the present application. The processes depicted in the drawings do not necessarily require the specific order and continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or can be advantageous.

[0112] The above description is only the preferred embodiment of the present application, and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0113] The specification and drawings are only exemplary and illustrative of the present application and are considered to cover any and all modifications, variations, combinations or equivalents that are within the scope of the present application. Obviously, various modifications and changes can be made to the present application by those skilled in the art without departing from the scope of the present application. Thus, it is intended that the present application cover the modifications and changes as they come within the scope of the application, and that the scope of the application be limited only by the claims.

Claims

1. A control method of a high-voltage cascade energy storage system, characterized by, The method comprises: receiving the charge-discharge instruction of the power grid dispatching center, decomposing the charge-discharge instruction according to the load fluctuation characteristics, and generating a multi-level power instruction set; real-time monitoring of the high-voltage cascaded energy storage system is performed to obtain a multi-unit state parameter set, dynamic balance analysis is performed according to the multi-level power instruction set and the multi-unit state parameter set, and a multi-level collaborative scheduling instruction is generated; the multi-level collaborative scheduling instruction is mapped to the high-voltage cascaded energy storage system for predictive control, converted into a charge-discharge control instruction set, and the charge-discharge control instruction set is executed for monitoring and recording to generate system response data for closed-loop feedback correction of the multi-level collaborative scheduling instruction, and a system optimization control parameter set is obtained; the method for generating a multi-level collaborative scheduling instruction according to the multi-level power instruction set and the multi-unit state parameter set comprises: loading the multi-level power instruction set for real-time reading to determine a real-time frequency spectrum component; a health degree analysis unit is constructed, the multi-level power instruction set and the multi-unit state parameter set are synchronized to the health degree analysis unit, and a plurality of health degree parameters are generated; state transition evaluation is performed on the multi-level power instruction set and the multi-unit state parameter set according to the plurality of health degree parameters, and a state evaluation matrix is generated; dynamic balance analysis is performed according to the state evaluation matrix, and a state balance coefficient is generated; based on the state balance coefficient, multi-level scheduling identification is performed on the high-voltage cascaded energy storage system to generate a multi-level scheduling data tag; scheduling conflict analysis is performed based on the multi-level scheduling data tag, the multi-level scheduling data tag is updated according to the analysis result, and the multi-level collaborative scheduling instruction is generated; the method for generating a state evaluation matrix by performing state transition evaluation on the multi-level power instruction set and the multi-unit state parameter set according to the plurality of health degree parameters comprises: dynamic normalization processing is performed on the plurality of health degree parameters to determine a standard health degree parameter vector; an initial probability table is obtained by initializing the state transition probability without an event occurring, and a discrete state space is constructed by defining a state transition rule in combination with the initial probability table; the multi-level power instruction set and the multi-unit state parameter set are mapped to the discrete state space for state transition, node transition statistics are performed according to the initial probability table, and node transition frequency information is obtained; state transition distribution data is determined according to the standard health degree parameter vector in combination with the node transition frequency information; the state transition distribution data is stored in a structured manner to construct the state evaluation matrix.

2. The control method of a high voltage cascade energy storage system according to claim 1, wherein, the method for decomposing the charge-discharge instruction according to the load fluctuation characteristics to generate a multi-level power instruction set comprises: spectrum recognition is performed on the load fluctuation characteristics to determine high-frequency fluctuation components and low-frequency fluctuation components; response analysis is performed on the high-voltage cascaded energy storage system according to the charge-discharge instruction to obtain a first response speed parameter and a second response speed parameter; a first-level energy storage unit cluster is determined according to the first response speed parameter, and a second-level energy storage unit cluster is determined according to the second response speed parameter; The high-frequency fluctuation component is allocated to the first-stage energy storage unit cluster, and the low-frequency fluctuation component is allocated to the second-stage energy storage unit cluster, to generate the multi-stage power instruction set.

3. The control method of a high voltage cascade energy storage system according to claim 2, wherein, The method comprises: Collecting historical load data of the high-voltage cascade energy storage system to calculate the load change rate; Performing fast Fourier transform according to the load change rate to extract the frequency spectrum energy distribution proportion value; According to the fluctuation period, the frequency spectrum energy distribution proportion value is subjected to wavelet analysis, and the decomposition layer number is set; The load fluctuation characteristics are subjected to frequency spectrum decomposition reconstruction according to the decomposition layer number, and the high-frequency fluctuation component and the low-frequency fluctuation component are identified.

4. The control method of a high voltage cascaded energy storage system according to claim 1, characterized by, A health degree analysis unit is constructed, and the multi-stage power instruction set and the multi-unit state parameter set are synchronized to the health degree analysis unit to generate a plurality of health degree parameters, the method comprising: Based on the main controller of the high-voltage cascade energy storage system, a health degree analysis thread is created to obtain a health degree analysis unit, and the multi-stage power instruction set and the multi-unit state parameter set are synchronized to the health degree analysis unit: S1: Set a triple data buffering mechanism, which includes a first buffering layer, a second buffering layer, and a third buffering layer; S2: Temporarily store the multi-stage power instruction set and the multi-unit state parameter set through the first buffering layer to generate a first buffering result; S3: Align the time stamp of the first buffering result through the second buffering layer to generate a second buffering result; S4: Perform unit matching identification on the second buffering result through the third buffering layer to generate a third buffering result; S5: Update and compensate the priority parameters based on the third buffering result to generate the plurality of health degree parameters.

5. The control method of a high voltage cascaded energy storage system according to claim 1, wherein, The multi-stage collaborative scheduling instruction is mapped to the high-voltage cascade energy storage system for predictive control, and is converted into a charge and discharge control instruction set, the method comprising: Through the grid interaction recording device, the grid interaction parameters are obtained, and the state constraint mapping is performed according to the grid interaction parameters to construct the grid constraint state space; Based on the grid constraint state space, interval rolling is performed, and a rolling time window is set; According to the grid constraint state space, the multi-stage collaborative scheduling instruction is mapped to the high-voltage cascade energy storage system, and the scheduling prediction is performed according to the rolling time window to generate a collaborative scheduling prediction result; Based on the collaborative scheduling prediction result, multi-objective optimization is performed to generate an optimization solution set; According to the control step, the optimization solution set is injected with charge and discharge compensation data for conversion to generate the charge and discharge control instruction set.

6. The control method of a high voltage cascade energy storage system according to claim 1, wherein, The charge and discharge control instruction set is executed for monitoring and recording to generate system response data, the multi-stage collaborative scheduling instruction is closed-loop feedback corrected, and a system optimization control parameter set is obtained, the method comprising: The power conversion unit is driven to execute the charge and discharge control instruction for multi-dimensional monitoring to generate multi-dimensional monitoring data; Based on the multi-dimensional monitoring data, the associated time stamp is generated to generate system response data; According to the system response data, the multi-stage collaborative scheduling instruction is subjected to scheduling deviation analysis to generate a scheduling deviation parameter; According to the scheduling deviation parameter, parameter stability analysis is performed to generate a parameter stability coefficient, the scheduling deviation parameter is marked with priority according to the parameter stability coefficient, and a scheduling deviation correction sequence is determined; An activation parameter correction instruction is generated, and the parameter correction instruction is used to perform closed-loop feedback correction according to the scheduling deviation correction sequence to generate the system optimization control parameter set.

7. The control method of a high voltage cascade energy storage system according to claim 6, characterized by, The method comprises the following steps: The power conversion unit is driven to execute the charge-discharge control instruction to perform multi-dimensional monitoring to generate multi-dimensional monitoring data. The power conversion unit is driven to execute the charge-discharge control instruction to perform voltage monitoring on the high-voltage cascade energy storage system to obtain transient voltage fluctuation data. The power conversion unit is driven to execute the charge-discharge control instruction to perform current monitoring on the high-voltage cascade energy storage system to obtain multi-branch current data. The power conversion unit is driven to execute the charge-discharge control instruction to perform temperature monitoring on the high-voltage cascade energy storage system to draw a two-dimensional thermal diagram.

8. A control device of a high-voltage cascade energy storage system, characterized by, The transient voltage fluctuation data, the multi-branch current data, and the two-dimensional thermal diagram are associated and integrated to generate the multi-dimensional monitoring data. The device is used to execute the control method of the high-voltage cascade energy storage system according to any one of claims 1-7, and the device comprises: An instruction decomposition module is configured to receive a charge-discharge instruction from a power grid dispatching center, decompose the charge-discharge instruction based on load fluctuation characteristics, and generate a multi-level power instruction set. A real-time monitoring module is configured to perform real-time monitoring on the high-voltage cascade energy storage system, obtain a multi-unit state parameter set, perform dynamic balancing analysis based on the multi-level power instruction set and the multi-unit state parameter set, and generate a multi-level collaborative scheduling instruction. A feedback correction module is configured to map the multi-level collaborative scheduling instruction to the high-voltage cascade energy storage system for predictive control, convert the multi-level collaborative scheduling instruction into a charge-discharge control instruction set, execute the charge-discharge control instruction set to perform monitoring and recording, generate system response data, perform closed-loop feedback correction on the multi-level collaborative scheduling instruction, and obtain a system optimization control parameter set.

Images