Intelligent monitoring method, system and equipment for liquid cooling energy storage cabinet based on self-iteration deep learning algorithm and storage medium

Abstract

The invention discloses a liquid cooling energy storage cabinet intelligent monitoring method, system and device based on a self-iteration deep learning algorithm, and a storage medium, and relates to the technical field of intelligent battery management and energy storage system control, and the method comprises the steps: collecting temperature data, and carrying out the prediction and storage; a control decision is made based on distributed liquid cooling cabinet temperature monitoring; updating and training the model through a self-iteration deep learning algorithm; and performing predictive control and early warning based on the model. According to the liquid cooling energy storage cabinet intelligent monitoring method based on the self-iteration deep learning algorithm, real-time accurate control is carried out through the mathematical model, and the timeliness is high. The cloud platform can iterate corresponding mathematical models for different distributed energy storage cabinets, and can integrate high-risk fault prediction data in different places into deep learning models in other places, so that the problem of frequent alarm or missing report caused by too high or too low threshold setting due to the problem of artificial experience is reduced; the method has better effects in the aspects of timeliness, missing report rate and cost.

Description

Technical Field

[0001] The present invention relates to the field of intelligent battery management and energy storage system control technology, and specifically to an intelligent monitoring method, system, device and storage medium for liquid-cooled energy storage cabinets based on a self-iterative deep learning algorithm. Background Art

[0002] New energy generation technologies, especially wind power and photovoltaic power generation, are growing rapidly and gradually becoming the main body of new power generation installations, and distributed photovoltaics are experiencing explosive growth. Currently, energy storage systems are valued as key technologies for balancing the supply and demand between distributed photovoltaics, power grids and electricity users. As a key thermal management solution in energy storage systems, liquid-cooled energy storage cabinet technology is particularly suitable for occasions with high battery pack energy density, fast charging and discharging speeds, and large ambient temperature changes due to its high heat dissipation efficiency, fast heat dissipation speed, and good temperature uniformity.

[0003] Existing liquid-cooled energy storage cabinets mostly use two operation and maintenance management solutions: local monitoring and data cloud. In addition, control management mostly relies on simple threshold triggering and manual monitoring, which cannot meet the growing demand for automation and intelligence. Local deployment monitoring has strong real-time performance, but it is not suitable for distributed scenarios and deployment of corresponding data analysis services; while data cloud needs to be transmitted through the collector, gateway, cloud server to the database for storage, and then through another client for data query and access. Although historical data can be analyzed, it is difficult for real-time faults on site to reach the analysis terminal in the first place, which limits the efficiency of remote monitoring. Summary of the invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is that the existing operation and maintenance management method of liquid-cooled energy storage cabinets cannot meet the growing demand for automation and intelligence, local deployment monitoring is not suitable for distributed scenarios and deployment of corresponding data analysis services, it is difficult to analyze and process real-time faults on site in the first time when data is uploaded to the cloud, and the efficiency of remote monitoring is limited.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: an intelligent monitoring method for liquid-cooled energy storage cabinets based on a self-iterative deep learning algorithm, comprising collecting temperature data for prediction and storage; making control decisions based on distributed liquid cooling cabinet temperature monitoring; updating and training models through a self-iterative deep learning algorithm; and predictive control and early warning based on the model.

[0007] As a preferred solution of the intelligent monitoring method of liquid-cooled energy storage cabinet based on self-iterative deep learning algorithm described in the present invention, wherein: the acquisition of temperature data for prediction and storage includes the BMS system acquiring the real-time voltage U of each battery according to time sampling tb (t), real-time current I b (t) and real-time temperature T b (t) The data is given to the local lightweight database, and the temperature T of each battery at the next moment is predicted based on the BMS temperature model based on the deep learning algorithm and the real-time voltage and current data. p (t+1) data and store it in the database, and then the current temperature data T b (t) and the temperature prediction T made at the last moment stored in the database p (t) The data is expressed as a difference:

[0008] T d (t) = T b (t)-T p (t)

[0009] Among them, T d (t) represents the current temperature data difference;

[0010] The difference is stored and awaits access by the host controller.

[0011] As a preferred solution of the method for intelligent monitoring of liquid cooling energy storage cabinets based on self-iterative deep learning algorithm described in the present invention, wherein: the control decision based on temperature monitoring of distributed liquid cooling cabinets includes the distributed liquid cooling energy storage cabinet collecting the real-time temperature data of the current liquid cooling cabinet through a collector and transmitting it to the main controller;

[0012] The main controller queries the local database to see whether the difference between the temperature prediction and the real-time data of each battery pack is within the difference range;

[0013] The difference range is the range where the difference between the temperature prediction and the real-time data is between -5% and 5%, while the ambient temperature for normal operation of the battery pack is 5 to 45°C;

[0014] If it is not within the difference range, an update message will be sent to the cloud platform;

[0015] If the data difference is still within the difference range, the main controller predicts the next temperature data of the entire liquid-cooled energy storage cabinet based on the deep learning algorithm.

[0016] As a preferred solution of the method for intelligent monitoring of liquid cooling energy storage cabinet based on self-iterative deep learning algorithm described in the present invention, wherein: the control decision based on distributed liquid cooling cabinet temperature monitoring includes: p (t) Comprehensively consider whether it is necessary to increase or decrease the power of the liquid cooler P p (t), power depends on temperature data T p (t);

[0017] If T p(t) When the difference range is 5 to 45°C, the current liquid cooler power is not adjusted;

[0018] When the predicted temperature data T p When (t)>45, the simplified heat conduction model is used for adaptive adjustment, which is expressed as:

[0019] P d (t) = k(T p (t)-45)

[0020] Where k is a proportional coefficient, which depends on the properties of the coolant, flow rate, thermal capacity of the battery, and heat exchange efficiency. d (t) represents the current liquid cooler power.

[0021] As a preferred solution of the intelligent monitoring method of liquid-cooled energy storage cabinet based on self-iterative deep learning algorithm described in the present invention, wherein: the updating and training model through the self-iterative deep learning algorithm includes that after the cloud platform receives the data of each distributed energy storage cabinet, the cloud platform queries the latest data in the database and updates the prediction model based on the self-iterative deep learning algorithm.

[0022] As a preferred solution of the method for intelligent monitoring of liquid-cooled energy storage cabinets based on self-iterative deep learning algorithm described in the present invention, wherein: the updating and training model through the self-iterative deep learning algorithm includes backing up the latest data to the cloud database, which is divided into a training set and a test set. The BMS deep learning model uses the voltage and current data of each battery as variables, and the temperature 0-40 of each battery as the dependent variable, and sets the network layer and activation function of the deep learning algorithm respectively for training. After the trained data model reaches 95% accuracy through the test set test, it is transmitted back to each distributed main controller by the cloud platform. At the same time, if the faulty data is identified, it will be returned to all distributed nodes to avoid it.

[0023] As a preferred solution of the method for intelligent monitoring of liquid-cooled energy storage cabinets based on self-iterative deep learning algorithm described in the present invention, wherein: the model-based predictive control and early warning include the main control receiving the updated mathematical model and applying it to the next collection, prediction, comparison and control of BMS and liquid cooling system;

[0024] Control the battery temperature in advance through prediction, and perform temperature control for scenarios that require alarms according to predictions;

[0025] If the battery temperature cannot be controlled in advance, the power will be cut off according to the threshold.

[0026] Another object of the present invention is to provide an intelligent monitoring system for liquid-cooled energy storage cabinets based on a self-iterative deep learning algorithm, which can make control decisions through distributed liquid-cooled cabinet temperature monitoring, solving the problem that the current operation and maintenance management method of liquid-cooled energy storage cabinets contains data on the cloud, making it difficult to analyze and process real-time faults on site in the first time.

[0027] As a preferred solution of the intelligent monitoring system for liquid-cooled energy storage cabinets based on the self-iterative deep learning algorithm described in the present invention, it includes: an acquisition module, a BMS module, a data processing module, a control module, and an output module; the acquisition module is used to collect the key temperature and humidity parameters of the liquid-cooled energy storage cabinet in real time; the BMS module is used to collect the voltage, current and temperature parameters of each battery pack in real time and manage the battery pack; the data processing module is used to clean the collected data, perform edge computing and store it using the trained deep learning model; the control module is used to analyze the real-time data, predict the operating status of the energy storage cabinet, and automatically adjust the operating parameters of the liquid cooling system according to the status prediction results; the output module is used to train the deep neural network model based on historical data obtained from the distributed deployment database, and according to the latest difference, it automatically iterates and optimizes the deep learning prediction mathematical model corresponding to each distributed energy storage cabinet.

[0028] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement a step of an intelligent monitoring method for a liquid-cooled energy storage cabinet based on a self-iterative deep learning algorithm.

[0029] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a method for intelligently monitoring a liquid-cooled energy storage cabinet based on a self-iterative deep learning algorithm.

[0030] Beneficial effects of the present invention: The intelligent monitoring method for liquid-cooled energy storage cabinets based on self-iterative deep learning algorithms provided by the present invention, compared with a completely local deployment solution, can centrally manage distributed scenarios on the one hand, can realize remote monitoring on the other hand, and can reduce the cost of local deployment equipment in multiple distributed scenarios on the other hand. Compared with a complete data cloud solution, it can perform edge computing locally, and execute the mathematical model obtained by deep learning for real-time and precise control, with high timeliness. The cloud platform can not only iterate the corresponding mathematical models for different distributed energy storage cabinets, but also integrate the high-risk fault prediction data occurring in different places into the deep learning models of other places, which is perfectly suitable for distributed control scenarios, and reduces the frequent alarms or missed reports caused by the excessively high or low threshold settings due to problems with manual experience. Under the premise of more accurate prediction, the alarm frequency of small and medium-sized faults can be greatly reduced, and major faults can also be prevented in advance. The present invention achieves better results in terms of timeliness, missed report rate and cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0032] Figure 1 A multi-layer feedforward DNN fully connected training model diagram of a liquid-cooled energy storage cabinet intelligent monitoring method based on a self-iterative deep learning algorithm is provided in the first embodiment of the present invention.

[0033] Figure 2 A graph of model prediction accuracy after iterative optimization for an intelligent monitoring system for a liquid-cooled energy storage cabinet based on a self-iterative deep learning algorithm provided in the second embodiment of the present invention.

[0034] Figure 3 A comparison chart of a liquid-cooled energy storage cabinet intelligent monitoring system based on a self-iterative deep learning algorithm provided in the second embodiment of the present invention and the mainstream regression prediction model currently on the market.

[0035] Figure 4 An overall flow chart of an intelligent monitoring system for a liquid-cooled energy storage cabinet based on a self-iterative deep learning algorithm provided in the third embodiment of the present invention. DETAILED DESCRIPTION

[0036] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0037] Example 1, reference Figure 1 , is an embodiment of the present invention, and provides a method for intelligent monitoring of a liquid-cooled energy storage cabinet based on a self-iterative deep learning algorithm, comprising:

[0038] S1: Collect temperature data for prediction and storage.

[0039] Furthermore, the BMS system collects the real-time voltage U of each battery at time sampling t b (t), real-time current I b (t) and real-time temperature T b(t) The data is given to the local lightweight database, and the temperature T of each battery at the next moment is predicted based on the BMS temperature model based on the deep learning algorithm and the real-time voltage and current data. p (t+1) data and store it in the database, and then the current temperature data T b (t) and the temperature prediction T made at the last moment stored in the database p (t) The data is expressed as a difference:

[0040] T d (t) = T b (t)-T p (t)

[0041] Among them, T d (t) represents the current temperature data difference;

[0042] The difference is stored and awaits access by the host controller.

[0043] S2: Make control decisions based on distributed liquid cooling cabinet temperature monitoring.

[0044] Furthermore, the distributed liquid-cooled energy storage cabinet collects the real-time temperature data of the current liquid-cooled energy storage cabinet through the collector and transmits it to the main controller. The main controller then queries the local database to see whether the difference between the temperature prediction and real-time data of each battery pack is within a reasonable range. The difference between the temperature prediction and real-time data needs to be in the range of -5% to 5%. The difference is considered to be within a reasonable range, and the normal working environment temperature of the battery pack is 5 to 45°C. If not, an update message will be sent to the cloud platform; if the data difference is still within a reasonable range, the main controller predicts the temperature data of the entire liquid-cooled energy storage cabinet at the next moment based on the deep learning algorithm, and comprehensively considers whether it is necessary to increase or decrease the power of the liquid cooler based on the currently measured temperature data.

[0045] It should be noted that according to the current predicted temperature data T p (t) Comprehensively consider whether it is necessary to increase or decrease the power of the liquid cooler P p (t), where the power depends on the temperature data T p (t), if T p (t) Within the normal operating range (5-45°C), the current liquid cooler power is not adjusted. p When (t)>45, the simplified heat conduction model is used for adaptive adjustment, which is expressed as:

[0046] P d (t) = k(T p (t)-45)

[0047] Where k is a proportional coefficient, which depends on the properties of the coolant, flow rate, thermal capacity of the battery, and heat exchange efficiency. d (t) represents the current liquid cooler power.

[0048] S3: Update and train the model through a self-iterative deep learning algorithm.

[0049] Furthermore, after receiving the data from each distributed energy storage cabinet, the cloud platform queries the latest data in the database and updates the prediction model based on the self-iterative deep learning algorithm. The latest data is backed up to the cloud database and then divided into training and test sets. The BMS deep learning model uses the voltage and current data of each battery as variables and the temperature (0-40) of each battery as the dependent variable. The network layer and activation function of the deep learning algorithm are set separately, and then training is performed. After the trained data model reaches 95% accuracy through the test set test, it is transmitted back to each distributed main controller by the cloud platform. At the same time, if abnormal high-risk fault data is identified, it will be returned to all distributed nodes to avoid it.

[0050] It should be noted that, first of all, the overall framework of the DNN algorithm is built, such as Figure 2 As shown, the current DNN network layer includes three layers: input layer, hidden layer and output layer. The hidden layer uses Q neurons, the input layer contains 4 neurons, and the linear rectification function is used as the activation function of the neuron.

[0051] The real-time current I b (t), real-time voltage U b (t), real-time temperature T p (t) and the current power of the liquid cooler P d (t) is normalized and used as the input to the first layer χ i The output of the hidden layer is obtained as:

[0052]

[0053] Among them, ω ij represents the input layer weight, B j Represents the offset of each neuron in the hidden layer. After the initial values ​​of the two variables j=1…Q are set, the difference between the predicted output and the actual output in the training process will be output as a loss function and then fed back;

[0054] The output of the output layer is:

[0055]

[0056] Among them, ω j represents the hidden layer weight, and B is the offset of each neuron in the output layer;

[0057] At the same time, the error between the predicted value and the actual value of the system is e=TO

[0058] Furthermore, using the principle of gradient descent, the weights and offsets of each layer of neurons are updated iteratively and expressed as:

[0059]

[0060] Among them, β is the step size.

[0061] S4: Model-based predictive control and early warning.

[0062] Furthermore, after receiving the updated mathematical model, the main control will apply it to the next collection, prediction, comparison and control of the BMS and liquid cooling system. On the one hand, it saves the loss of the liquid cooler and realizes more precise control; on the other hand, it can also control the battery temperature in advance through such predictions, increase or decrease the cooling for the scenarios predicted to require alarms, and if it fails to meet the requirements, cut off the power according to the threshold, effectively reducing the manual participation and triggering times of low-level faults, and reducing missed reports and false alarms.

[0063] It should be noted that the normal working environment of the battery pack is 5-45, which can be set as the comparison condition for prediction adjustment, while -20 and 60 are considered to be damaged battery packs and need to be set as the alarm threshold T h When the prediction model training is completed and loaded into the local deployment, the main controller will first calculate the difference T between the prediction value at the last moment and the current actual temperature. d (t) Is it within a reasonable range (-5% to 5%)? If it is within a reasonable range, the current model is considered credible and the next temperature value T is predicted based on the prediction model. p (t+1). If T p (t+1) is within the normal working environment temperature range of 5 to 45, no operation is performed; if it is greater than 45 or less than 5, the liquid cooler power is adjusted according to the heat conduction model in reply 4. If the predicted value exceeds the alarm threshold T h Emergency measures such as alarm and power outage are directly triggered. If the predicted value is not within a reasonable range, the main controller sends interactive information to the cloud server, allowing it to read the latest data from the database to update the deep learning prediction model and return it to the main controller.

[0064] Example 2, reference Figure 2-Figure 3 , which is an embodiment of the present invention, provides an intelligent monitoring method for liquid-cooled energy storage cabinets based on a self-iterative deep learning algorithm. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0065] First of all, this experiment simulates the real-time temperature monitoring and control of different battery packs in a distributed liquid-cooled energy storage system. The voltage, current and temperature data of each battery pack are collected through the BMS system and stored in a local lightweight database. First, the system establishes a temperature prediction model for the battery pack. Based on the deep learning algorithm, it is trained through the real-time collected voltage and current data to accurately predict the temperature change of the battery at the next moment. The main controller accesses the database in real time, compares the current temperature data with the temperature data stored at the previous moment, and calculates the difference. If the difference is too large, the system sends a data update request to the cloud platform; if it is within a reasonable range, the main controller dynamically adjusts the power of the liquid cooling system according to the next moment temperature predicted by the BMS and the current temperature, thereby achieving the purpose of precise temperature control.

[0066] In the experimental design, after receiving data from the distributed energy storage cabinets, the cloud platform updates the data set and optimizes the temperature prediction model to ensure the accuracy of the model in different scenarios. After training, the prediction model has an accuracy of more than 95% and is applied to different distributed nodes, such as Figure 3 Shown is the prediction accuracy of the model after iterative optimization.

[0067] In addition, in order to reduce the inefficiency of on-site human intervention and avoid equipment loss, the control system performs real-time processing in edge computing mode, thereby reducing upload delays and enabling each distributed energy storage cabinet to perform edge computing and adjustment independently. The cloud platform ensures the consistency and accuracy of control of each node through centralized data iterative updates.

[0068] Table 1 Experimental data table

[0069]

[0070] It can be seen from the table data that the voltage and current parameters of each battery pack are recorded in real time during the implementation process, and the system can complete efficient temperature prediction on the edge device based on different parameters. Judging from the difference between the temperature prediction model and the real-time temperature data, the invention has a high prediction accuracy and avoids false alarms caused by excessive temperature differences, such as Figure 4 As shown in the figure, compared with the mainstream regression prediction model currently on the market, the model in this paper has higher prediction accuracy.

[0071] In addition, the power adjustment of the liquid cooling system is based on the predicted temperature changes, which can flexibly respond to the temperature status of each battery pack. This distributed edge control method ensures the accuracy and timeliness of temperature control.

[0072] Compared with completely localized deployment, the present invention can centrally manage the information of each distributed energy storage cabinet. The cloud platform improves the accuracy of system temperature prediction and reduces the local deployment cost of the equipment through iterative deep learning models. Compared with the complete cloud solution, the local edge computing mode has more advantages in timeliness and computing cost. Real-time processing reduces latency, allowing each distributed energy storage cabinet to independently achieve temperature control in different scenarios. In addition, the iteratively updated model of the cloud platform realizes the abnormal warning function of cross-regional distributed energy storage cabinets through the integration and processing of abnormal fault data, effectively reducing the risk of missed reports and false alarms, making the entire system more accurate in predicting high-risk faults, and further reducing the alarm frequency and equipment loss of small and medium-sized faults.

[0073] Example 3, reference Figure 4 , is an embodiment of the present invention, and provides an intelligent monitoring system for liquid-cooled energy storage cabinets based on a self-iterative deep learning algorithm, including an acquisition module, a BMS module, a data processing module, a control module, and an output module.

[0074] The acquisition module is used to collect the key parameters of temperature and humidity of the liquid-cooled energy storage cabinet in real time. The BMS module is used to collect the voltage, current and temperature parameters of each battery pack in real time and manage the battery pack. The data processing module is used to clean the collected data, use the trained deep learning model for edge computing and storage. The control module is used to analyze the real-time data, predict the operating status of the energy storage cabinet, and automatically adjust the operating parameters of the liquid cooling system according to the status prediction results. The output module is used to train the deep neural network model based on the historical data obtained from the distributed deployment database, and iterate and optimize the deep learning prediction mathematical model corresponding to each distributed energy storage cabinet based on the latest difference.

[0075] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0076] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0077] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0078] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc. It should be noted that the above embodiments are only used to illustrate the technical solution of the present invention and are not limited. Although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

[0079] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An intelligent monitoring method for liquid-cooled energy storage cabinet based on self-iterative deep learning algorithm, characterized in that: include: Collect temperature data for prediction and storage; Make control decisions based on distributed liquid cooling cabinet temperature monitoring; Update and train models through self-iterative deep learning algorithms; Model-based predictive control and early warning.

2. The intelligent monitoring method for liquid-cooled energy storage cabinet based on self-iterative deep learning algorithm according to claim 1 is characterized in that: The temperature data collection for prediction and storage includes the BMS system collecting the real-time voltage U of each battery at time sampling t. b (t), real-time current I b (t) and real-time temperature T b (t) The data is given to the local lightweight database, and the temperature T of each battery at the next moment is predicted based on the BMS temperature model based on the deep learning algorithm and the real-time voltage and current data. p (t+1) data and store it in the database, and then the current temperature data T b (t) and the temperature prediction T made at the last moment stored in the database p (t) The data is expressed as a difference: T d (t)=T b (t)-T p (t) Among them, T d (t) represents the current temperature data difference; The difference is stored and awaits access by the host controller.

3. The intelligent monitoring method for liquid-cooled energy storage cabinet based on self-iterative deep learning algorithm according to claim 1 or 2, characterized in that: The control decision based on the temperature monitoring of the distributed liquid cooling cabinet includes the distributed liquid cooling energy storage cabinet collecting the real-time temperature data of the current liquid cooling cabinet through the collector and transmitting it to the main controller; The main controller queries the local database to see whether the difference between the temperature prediction and the real-time data of each battery pack is within the difference range; The difference range is the range where the difference between the temperature prediction and the real-time data is between -5% and 5%, while the ambient temperature for normal operation of the battery pack is 5 to 45°C; If it is not within the difference range, an update message will be sent to the cloud platform; If the data difference is still within the difference range, the main controller predicts the next temperature data of the entire liquid-cooled energy storage cabinet based on the deep learning algorithm.

4. The intelligent monitoring method for liquid-cooled energy storage cabinet based on self-iterative deep learning algorithm according to claim 1 or 3, characterized in that: The control decision based on the distributed liquid cooling cabinet temperature monitoring includes the following steps: p (t) Comprehensively consider whether it is necessary to increase or decrease the power of the liquid cooler P p (t), power depends on temperature data T p (t); If T p (t) When the difference range is 5 to 45°C, the current liquid cooler power is not adjusted; When the predicted temperature data T p When (t)>45, the simplified heat conduction model is used for adaptive adjustment, which is expressed as: P d (t)=k(T p (t)-45) Where k is a proportional coefficient, which depends on the properties of the coolant, flow rate, thermal capacity of the battery, and heat exchange efficiency. d (t) represents the current liquid cooler power.

5. The intelligent monitoring method for liquid-cooled energy storage cabinet based on self-iterative deep learning algorithm according to claim 4 is characterized in that: The updating and training of the model through the self-iterative deep learning algorithm includes that after the cloud platform receives the data of each distributed energy storage cabinet, the cloud platform queries the latest data in the database and updates the prediction model based on the self-iterative deep learning algorithm.

6. The intelligent monitoring method for liquid-cooled energy storage cabinet based on self-iterative deep learning algorithm according to claim 1, 2, 3 or 4, characterized in that: The self-iterative deep learning algorithm is used to update and train the model, including backing up the latest data to the cloud database, which is divided into a training set and a test set. The BMS deep learning model uses the voltage and current data of each battery as variables, and the temperature of each battery 0-40 as the dependent variable, and sets the network layer and activation function of the deep learning algorithm for training. After the trained data model reaches 95% accuracy through the test set test, it is transmitted back to each distributed main controller by the cloud platform. At the same time, if a fault is identified, the data will be returned to all distributed nodes to avoid it.

7. The intelligent monitoring method for liquid-cooled energy storage cabinet based on self-iterative deep learning algorithm according to claim 6 is characterized in that: The model-based predictive control and early warning include the main control receiving the updated mathematical model, which will be applied to the next collection, prediction, comparison and control of the BMS and liquid cooling system; Control the battery temperature in advance through prediction, and perform temperature control for scenarios that require alarms according to predictions; If the battery temperature cannot be controlled in advance, the power will be cut off according to the threshold.

8. An intelligent monitoring system for liquid-cooled energy storage cabinets based on a self-iterative deep learning algorithm, characterized in that: Including acquisition module, BMS module, data processing module, control module and output module; The acquisition module is used to collect key parameters of temperature and humidity of the liquid-cooled energy storage cabinet in real time; The BMS module is used to collect voltage, current and temperature parameters of each battery pack in real time and manage the battery pack; The data processing module is used to clean the collected data, perform edge computing using the trained deep learning model, and store the data; The control module is used to analyze real-time data, predict the operating status of the energy storage cabinet, and automatically adjust the operating parameters of the liquid cooling system according to the status prediction results; The output module is used to train a deep neural network model based on historical data obtained from a distributed deployment database, and iteratively optimize the deep learning prediction mathematical model corresponding to each distributed energy storage cabinet based on the latest difference.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for intelligent monitoring of a liquid-cooled energy storage cabinet based on a self-iterative deep learning algorithm are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: The steps of the intelligent monitoring method of liquid-cooled energy storage cabinet based on self-iterative deep learning algorithm.

Images