CR2450 battery.Research on expert diagnostic system for lithium-ion battery packs

　　Research on expert diagnostic system for lithium-ion battery packs

　　1 Introduction

　　With the continuous development of the national economy, the requirements for modernization in the fields of energy, electricity, transportation, communications, environmental protection and other fields are also constantly increasing. The battery system as a backup energy source is being widely used. The battery pack is an indispensable backup power system for all power supply systems that do not allow power outages. Moreover, battery systems are increasingly used in various industries. Whether the battery is running normally or not directly affects the normal, reliable and safe operation of various equipment in the application field. In particular, backup batteries for unattended sites, e-commerce centers, and banks are particularly important.

　　There are still performance differences between the individual cells in the tested and optimally grouped battery packs, and these differences can produce new differences to varying degrees due to small differences in the environment (such as temperature differences) during the long-term operation of the battery. After long-term operation, the performance of individual batteries declines significantly, which seriously affects the performance of the battery pack and even causes accidents. It is necessary to diagnose the performance degradation and early failure of single cells early. In addition, the performance degradation and failure of single cells will reduce the SOC (state of charge) value of the battery pack, because the power of a single cell with poor performance determines the state of charge of the entire battery pack. A battery pack is generally composed of several single cells or battery modules connected in series; single cells with poor performance may cause the entire battery pack to terminate discharge early. Therefore, it is necessary to equip an expert system for early diagnosis of battery failure. The diagnostic system can achieve early prediction of battery faults and hidden dangers, thereby effectively increasing the driving range and trouble-free working time of the electric vehicle battery pack, minimizing the maintenance workload, thus ensuring the reliable operation of the electric vehicle.

　　All or part of the energy of electric vehicles comes from energy storage batteries. A major issue restricting the development of electric vehicle technology and the industrialization of electric vehicles is energy storage batteries and their applications. A prominent problem in the use of batteries is the differences and inconsistencies between batteries. Therefore, early detection of battery inconsistencies and failures has become a very critical technical problem. It is necessary to develop an early failure expert diagnosis system for battery packs. Through the expert diagnosis system, we can achieve early diagnosis of unhealthy batteries and make the remaining power estimation model more accurate, which can extend the service life of the battery and further reduce the cost of battery use, increase the driving range of electric vehicles, and improve vehicle Driving reliability. Research on this system and technology has theoretical value and is of great significance in promoting the industrialization of electric vehicles.

　　2. Development status of expert diagnostic systems at home and abroad

　　According to the requirements of the battery management system standard formulated by the International Electrotechnical Commission (IEC) in 1995, the battery management system for electric vehicles must have certain battery diagnostic functions, including early warning of unhealthy batteries and providing battery aging information. Over the past 10 years, major foreign companies have conducted vigorous research on this issue and added certain battery diagnostic functions to the battery management systems they operate. Among them, representative companies include the BADICOaCH system designed by Germany's Mentzer Electronic GmbH and Werner Retzlaff; the SmartGuard system (Long--LifeBatteryUsingIntelligentModularControlSystem) developed by the American Aerovironmevt company.

　　(1) BADICOaCH system

　　The BADICOaCH system is an improvement of the BADICHEQ system. It has battery diagnosis-related functions: it stores detailed data of the last 24 charge and discharge cycles and allows quick search of basic battery information and incorrect usage when judging the quality of the battery;

　　(2) SmartGuard system

　　The battery diagnosis-related functions of the SmartGuard system mainly provide discharge polarity reverse alarm and battery history recording and archiving.

　　There is also some research on battery fault diagnosis in China, such as Chunlan Research Institute and Tsinghua University, but most of the research is still in its infancy.

　　This article's research on battery fault diagnosis is mainly to find the relationship between battery performance faults and one or N parameters of the battery. It is synthesized by real-time monitoring and comparison of parameter changes between different single cells of the same battery pack and considering some other factors. diagnosis. After long-term research, a preliminary feasible algorithm has been summarized and used for battery diagnosis in electric vehicle sports car experiments. The idea is that under the same charge and discharge current of each single cell in the same battery pack, most of the performance of each single cell is similar, but there are also inconsistent performances of some cells. By comprehensively considering the deviation of each battery to the average voltage during this period of time and the voltage change of each battery during this period of time, the performance of the battery can be estimated. Battery performance is relatively good if the voltage deviation of a single cell is small and the voltage change is small. Because there is no battery failure that manifests itself as a slow voltage rise during charging and a slow voltage drop during discharge.

　　3 Design of expert fault diagnosis system

　　3.1 Introduction to expert systems

　　An expert system is a program system with a large amount of specialized knowledge. It can reason based on the provided knowledge in a special field, simulate the decision-making process of human experts in a certain field, and solve complex problems that require experts to solve.

　　The expert system mainly consists of five parts: knowledge base, inference engine, working storage area, knowledge acquisition subsystem and interpretation interface. Among them, the knowledge base and reasoning engine are the core of the expert system. The knowledge base is mainly used to store the expertise provided by domain experts. It includes a fact base and a rule base. The function of the inference engine is to select relevant knowledge from the knowledge base according to a certain reasoning strategy, and to reason on the evidence provided by the user until the corresponding conclusion is drawn [1].

　　In addition, some expert systems also have the function of automatic knowledge acquisition. On the one hand, knowledge is obtained from the outside through questions and answers with experts. On the other hand, the system can continuously summarize its own experience during operation and obtain knowledge from the inside.

　　3.2 Overall design

　　This article provides an in-depth understanding of the working principle of lithium iron phosphate batteries currently being developed and used in electric vehicles and their use in electric vehicles, analyzes the battery diagnosis experience of battery experts, and summarizes the battery diagnosis rules. At the same time, applying the knowledge of fuzzy mathematics, a reasonable and practical battery pack fault diagnosis model is initially proposed.

　　According to the battery pack fault diagnosis model's demand for battery usage status data, a battery pack data acquisition system and the corresponding host computer human-computer interaction interface are developed. Then a battery fault expert diagnosis system was constructed according to the expert system architecture. include:

　　1Create a global database

　　2 Establish battery history archive database

　　3. Establish a rule base

　　4. Develop inference scheduler module

　　5. Develop human-computer interaction program modules.

　　Finally, the battery fault expert diagnosis system is used to conduct diagnostic tests on the batteries used in electric vehicles. During the test, the parameters of the battery diagnostic model are adjusted to continuously improve it. The overall design structure diagram is shown in Figure 1.

　　Figure 1 Structure of battery pack diagnosis expert system Figure 3.3 Rules used in battery diagnosis fuzzy expert system

　　We analyze and organize the battery fault diagnosis rules, battery diagnosis and maintenance data provided by battery experts and then write them into the expert system. Then through experimental verification, trade-offs and increases are realized [2]. Taking lithium-ion batteries as an example, the main rules in the system are:

　　(1) If the discharge voltage drops quickly and the voltage is low, and the charging voltage rises quickly and the voltage is high, the battery capacity will become smaller;

　　(2) The battery terminal voltage drops quickly when left standing, and if the voltage is low for a long time, the self-discharge will be too large;

　　(3) The battery terminal voltage drops rapidly during discharge, and the voltage is about 1 volt lower than the average voltage, indicating that the unit cell is damaged;

　　(4) The open-circuit voltage of the battery is very low and cannot be loaded, so the battery is damaged or the connection is abnormal;

　　(5) If the voltage is high during charging and low during discharging, the internal resistance of the single battery is too large;

　　(6) If the voltage is extremely high during charging, the battery will have an internal open circuit;

　　(7) Since the battery started to discharge, its voltage has been slightly lower than other batteries. If the discharge platform performance is normal, the battery may be undercharged;

　　(8) During the battery discharge process, if the temperature of a certain single cell is more than 3°C higher than that of other single cells, the internal resistance of the battery is too large.

　　3.4 Historical archive data content and its establishment

　　Taking the lithium iron phosphate battery used in this experiment as an example, the data stored in the historical archives mainly include:

　　(1) Key data of the battery when it leaves the factory (such as date of manufacture, nominal capacity, open circuit voltage, etc.);

　　(2) Total safety hours used;

　　(3) Records of maximum voltage, current, and temperature during overcharge and overdischarge;

　　(4) In the last 10 charge-discharge cycles, the number of charging cycles with the highest voltage and the number of discharge cycles with the lowest voltage;

　　(5) Temperature rise and fall data and charging efficiency during charging in the last 10 cycles;

　　(6) The voltage difference during low current charging in the last 10 cycles;

　　(7) Self-discharge time interval;

　　(8) The degree of health (SOH) result of the last diagnosis.

　　The first time the system runs, the historical files are initialized. The principle of initialization is that except for some known basic parameters, other parts are set to the best state. During subsequent operations, the system will automatically record major battery-related events and modify the historical files. If a battery in the battery pack is removed, the history file of the battery that was just replaced should be initialized. Long-term memory and accumulation are used for records in historical files that affect battery health and service life, such as total ampere hours of use, total number of charge and discharge cycles, overcharge, over-discharge and undercharge; for other historical data that expresses performance Then use the method of regular refresh.

　　The specific implementation plan of historical archives is to use a long-term memory chip EEpROM in the system to save historical data, and at the same time add a clock circuit and a power supply battery to the system to provide time information for historical data.

　　3.5 Fault definition and handling process

　　This system uses four-level fault alarm definitions, which are first-level temperature difference faults, extremely high temperature faults, and extremely high cell voltage faults; second, third, and fourth-level temperature difference faults, pressure difference faults, excessive temperature faults, and extremely high cell voltage faults. , cell voltage is too low fault.

　　When the system is powered on, the battery pack data collection system will collect the voltage, temperature and other information of the battery pack cells at regular intervals during the charging and discharging process of the battery pack. When a fault occurs, record and calibrate the fault unit serial number. When the number of failures of the calibrated serial number cell reaches a certain number, the rule library is called to evaluate the battery performance, and the evaluation results are recorded in the history file of the cell.

　　4Experimental results and analysis

　　The lithium-ion battery pack module used in this experiment is a battery pack module for electric vehicles, using a lithium iron phosphate battery pack. The stand-alone module system is composed of 12 30Ah single cells connected in series. The battery pack module is equipped with a battery management data acquisition module, which transmits the collected data through the CAN bus. After the CAN232 interface conversion, the data is converted from the CAN data frame into a format that can be recognized by the PC machine, and is passed to the host computer human-computer interaction interface for display through the RS232 interface. The system connection is shown in Figure 3. Figure 4 shows the data collected after actually performing several charge and discharge cycles on a battery pack. Due to the differences between battery cells, the terminal voltage of No. 7 cell is significantly different from other cells, and a secondary voltage difference fault and low cell voltage fault alarm have occurred.

　　5 Conclusion

　　(1) Analyzed the connection between battery external characteristic data changes and battery faults, and combined the experience and knowledge of battery experts to summarize diagnostic rules for common battery faults.

　　(2) A battery pack data collection system platform was built, including underlying hardware data collection, data communication, and the writing of host computer human-computer interaction interface programs. The communication between the new data collection program and the battery management system is safer and more reliable, and is not affected by changes in the battery type and number of batteries monitored by the management system.

　　(3) A battery fault fuzzy diagnosis expert system was constructed. A bridge is built to find the fuzzy relationship between battery failure and battery external symptoms. At the same time, with continuous experiments, the expert diagnosis system will continue to be improved to provide more reliable guarantee for early failure warning.

　　Since the current number of experiments is relatively small, the dynamic characteristics and faults of the battery pack cannot be verified, and the fault diagnosis system cannot be verified either. As the experiment continues to deepen, battery fault information will slowly accumulate, and the rule base and inference engine will also be continuously corrected.

Read recommendations:

R20

What are the specific product characteristics of lithium iron phosphate batteries?lithium ion batter

What are the technical requirements for lithium iron phosphate battery packs?Nickel Hydride No. 5 ba

551521 lipo battery company

18650 battery 4800mah