Security risk sensing algorithm in charging process of vehicle direct current charging equipment

Abstract

The invention belongs to the technical field of new energy vehicle charging safety, and particularly relates to a safety risk perception algorithm in the charging process of vehicle direct current charging equipment, which comprises the following specific steps: S1, constructing a two-dimensional battery attribute risk mode vector xs by taking the highest battery pack monomer voltage xs1 and the maximum battery pack voltage difference xs2 as descriptors; s2, after the charging equipment interacts with the BMS to obtain the battery type and the SOC value, xs samples corresponding to the battery type and the SOC value are selected from the information base, and the average vector msj of the mode type samples is calculated; and S3, the charging equipment interacts with the BMS to obtain real-time battery pack information of the vehicle, xs corresponding information in the battery pack information is substituted into a minimum distance classification decision function dj (xs), if d23 (xS) is larger than 0 and d12 (xS) is larger than 0, the vehicle is identified as a safety mode class, and otherwise, the vehicle is identified as a safety risk mode. According to the method, safety risk sensing early warning can be sent out in real time, and potential safety hazards are effectively avoided.

Description

Technical Field

[0001] The present invention belongs to the technical field of new energy vehicle charging safety, and in particular relates to a safety risk perception algorithm during the charging process of a vehicle DC charging device. Background Art

[0002] At present, my country has built the world's largest and most extensive vehicle charging infrastructure system. The types and number of new energy electric vehicle charging equipment are growing rapidly. With the rapid development of new energy electric vehicle battery management system (BMS) technology, vehicle charging power continues to rise. Vehicle accidents frequently occur during the charging process due to overcharging, overheating or aging of vehicle batteries. The safety risk perception technology during the charging process of new energy vehicle charging equipment is very important to ensure the safety of vehicle charging.

[0003] Common safety risk management methods for new energy vehicle charging equipment include multiple threshold assessment method, electronic and electrical assessment method, and artificial neural network perception method. The multiple threshold assessment method is to set multiple safety thresholds in the software. When the detection parameters reach the threshold, the charging process is protected. It is often triggered after a safety accident occurs, and it is impossible to provide early warning protection in real time. It is often used as the last soft protection for charging process safety protection; the electronic and electrical assessment method uses the inherent properties of protection circuits and electrical components to trigger protection during the charging process. Due to the fixed hardware attribute values, the protection range is limited and lacks flexibility; the artificial neural network perception method has a large amount of calculation, the embedded computing power of the charging equipment is limited, and the real-time performance is poor. Existing safety risk management methods cannot provide real-time warnings during the charging process, and there are safety hazards. Summary of the invention

[0004] The purpose of the present invention is to provide a safety risk perception algorithm during the charging process of vehicle DC charging equipment to solve the problem that existing safety risk management and control methods cannot provide real-time warnings and have safety hazards.

[0005] A safety risk perception algorithm for the charging process of vehicle DC charging equipment, the specific steps are as follows:

[0006] S1: Build with the highest battery cell voltage x s 1 and the maximum voltage difference of the battery pack x s 2 is the two-dimensional battery attribute risk pattern vector x describing the sub s ,

[0007]

[0008] S2: The charging process safety information sample library includes three categories: safety mode, safety risk warning mode and safety risk mode. During the charging process of new energy vehicles, the charging equipment interacts with the BMS to obtain the battery type and SOC value, and then selects the x corresponding to the battery type and SOC in the information library. s Samples, calculate the average vector m of the pattern class samples respectively sj ,

[0009]

[0010] Where: s is the state of charge (SOC) value of the vehicle battery, ranging from 0-100, w j There are three types of sample libraries, N sj It is class w j The number of pattern vectors;

[0011] S3: During the vehicle charging process, the charging equipment interacts with the BMS to obtain the vehicle's real-time battery pack information and convert the battery pack information into s The corresponding information is brought into the minimum distance classification decision function d j (x s ), d 23 (x s )>0 and d 12 (x s )>0, it is identified as a safe mode, otherwise it is identified as a security risk mode and enters the alarm state.

[0012]

[0013] d ij (x s ) = d i (x s )-d j (x s ) j=1,2,3

[0014] S4: After the inherent safety risk of the battery is determined, safety management and control is carried out based on the dynamic safety risks during charging.

[0015] Preferably, the dynamic safety risk perception algorithm during charging in step S4 comprises the following specific steps:

[0016] S1: Construct a three-dimensional battery charging process dynamic risk pattern vector x with the battery pack unit temperature maximum increment x1, battery pack charging energy ratio x2 and battery pack single cell voltage boost ratio x3 as descriptors,

[0017]

[0018] Where: ΔT is the battery pack temperature increment per minute, ΔQ% is the ratio of the total charge increment to the total battery capacity of the vehicle, ΔSOC is the SOC increment during the charging process, max(ΔV s ) is the maximum battery pack single cell voltage increment during the charging process, It is the mean value of the battery pack single cell voltage increment during the charging process;

[0019] S2: Select the dynamic risk pattern vector x samples corresponding to the battery class in the information database, and calculate the average vector m of the pattern class samples respectively j and the covariance matrix C j ,

[0020]

[0021] S3: The corresponding information of x in the battery pack information is brought into the Bayesian classification decision function b j (x), b 23 (x)>0 and b 12 (x)>0, it is identified as a safe mode, otherwise it is identified as a security risk mode and enters the alarm state.

[0022]

[0023] b ij (x) = b i (x)-b j (x) j=1,2,3

[0024] Where: P(w j ) is the pattern class w j Probability of occurrence.

[0025] Compared with the prior art, the present invention has the following beneficial effects:

[0026] During the charging process of the vehicle DC charging equipment, the algorithm uses the minimum distance decision-making judgment to perceive the inherent safety risk of the battery property corresponding to the SOC of the vehicle battery during charging, judges the risk perception in real time and issues a warning, and combines the Bayesian decision-making judgment to perceive the dynamic safety risk of the vehicle battery during charging, and again judges the risk perception in real time and issues a warning. The double risk warning effectively avoids safety hazards, especially for the overcurrent, overcharging and overtemperature problems of the charging vehicle battery, reduces the probability of high-power DC charging accidents of the vehicle, and provides effective data support for the health judgment of the charging vehicle battery. Through a large number of actual verifications, the recognition accuracy of risk warning is ≥99.3%; the algorithm is designed in the embedded microprocessor chip STM32F413, and the algorithm lag time is ≤180ms, which meets the interaction cycle between the charging equipment and the BMS. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1This is a flow chart of the safety risk perception algorithm during the charging process in an embodiment of the present invention. DETAILED DESCRIPTION

[0028] The technical solution of the present invention is described clearly and completely below in conjunction with the accompanying drawings and embodiments.

[0029] This example is applied to the charging process of new energy vehicles and 120KW DC charging equipment. The vehicle battery type is a ternary material battery. In the startup phase, the charging equipment interacts with the BMS to obtain the vehicle battery information: the rated capacity of the battery system is 241Ah, the total battery energy is 84.8kWh, the rated total voltage is 362V, there are 83 battery packs, the startup SOC is 14%, the current battery voltage is 339.5V, the highest single cell voltage is 4.107V, the lowest single cell voltage is 3.602V, the highest temperature of the battery pack is 23 degrees, and the lowest temperature of the battery pack is 23 degrees. During the charging process, the SOC is 53%, the current battery voltage is 344.2V, the highest single cell voltage is 4.184V, the lowest single cell voltage is 3.809V, the highest temperature of the battery pack is 56 degrees, the lowest temperature of the battery pack is 38 degrees, and the total charging power is 24.16kWh.

[0030] like Figure 1 As shown, a safety risk perception algorithm for the charging process of a vehicle DC charging device is provided, and the specific steps are as follows:

[0031] S1: First build the battery pack with the highest cell voltage x s 1 and the maximum voltage difference of the battery pack x s 2 is the two-dimensional battery attribute risk pattern vector x describing the sub s ;

[0032]

[0033] S2: The charging process safety information sample library includes three categories: safety mode class, safety risk warning mode class and safety risk mode class. There are 9 types of battery models in the model class sample information library. The model class sample information is stored with a scale of 1% SOC value. Each scale sample information stores 2000 groups of sample values. During the charging process, the battery type of the sample information library selects ternary material batteries and x corresponding to the current SOC. s Samples, calculate the average vector m of the pattern class samples respectively sj ;

[0034]

[0035] At startup, SOC = 14%, When charging, SOC = 53%,

[0036] S3: During the vehicle charging process, the charging equipment interacts with the BMS to obtain the vehicle's real-time battery pack information and convert the battery pack information into s The corresponding information is brought into the minimum distance classification decision function d j (x s ), d 23 (x s )>0 and d 12 (x s )>0, it is identified as a safe mode, otherwise it is identified as a security risk mode and enters the alarm state;

[0037] At startup, SOC = 14%, The decision function is

[0038] d1(x s )=4.152x s1 +0.593x s2 -8.795

[0039] d2(x s )=4.192s1+0.695s2-9.028

[0040] d3(x s )=4.254x s1 +0.743x s2 -9.324

[0041] Get 12 (x s )>0,d 23 (x s )>0, the charging process safety static risk perception is identified as a safe mode;

[0042] When charging, SOC = 53%, The decision function is

[0043] d1(x s )=4.193x s1 +0.416x s2 -8.877

[0044] d2(x s )=4.221x s1 +0.492x s2 -9.029

[0045] d3(x s )=4.289x s1 +0.681x s2 -9.43

[0046] Get 12 (x s )>0,d23 (x s )>0, the static risk perception of the charging process is judged as safe mode;

[0047] S4: Under the static risk perception and identification of charging process safety as safety mode, a three-dimensional battery charging process dynamic risk pattern vector x=(x1,x2,x3) is constructed with the maximum increment of battery pack unit temperature x1, battery pack charging energy ratio x2 and battery pack single cell voltage boost ratio x3 as descriptors. T ;

[0048]

[0049] When charging, SOC = 53%. According to the parameter definition, 83 battery pack parameter values ​​and vehicle charging information values ​​are input and calculated in real time, and x = (3.2, 0.731, 1.355) is obtained. T ;

[0050] S5: Select the dynamic risk pattern vector x samples of ternary material batteries in the information database, and calculate the average vector m of the pattern class samples respectively j and the covariance matrix C j ;

[0051]

[0052] When charging, SOC = 53%, and m1 = (0.62, 0.945, 1.316) T ,m2=(0.94,0.89,1.421) T ,m3=(1.24,0.817,1.82) T ,

[0053] S6: When the SOC reaches 53% during the charging process, the corresponding information of x in the battery pack information is brought into the Bayesian classification decision function b j (x) and b ij (x);

[0054]

[0055] b ij (x) = b i (x)-b j (x) j = 1, 2, 3;

[0056] P(w j ) are Get 12 (x s )<0,d 23 (x s)<0, the dynamic risk perception of the charging process is judged as a safety risk mode and enters a risk warning state. The charging equipment warning management system determines that the warning reason is that the No. 41 battery pack heats up too quickly and the battery aging degree is 63.1%. When the charging parameters are within the safety threshold range, early warning of the vehicle charging battery risk is achieved, effectively preventing damage to the vehicle battery pack or even fire in overcharging and overheating scenarios.

[0057] The above embodiments are only preferred technical solutions of the present invention. Those skilled in the art may make simple substitutions for the contents of the present invention specification, which shall fall within the scope of patent protection of the present invention.

Claims

1. A safety risk perception algorithm for the charging process of vehicle DC charging equipment, characterized in that: The specific steps are as follows: S1: Build with the highest battery cell voltage x s 1 and the maximum voltage difference of the battery pack x s 2 is the two-dimensional battery attribute risk pattern vector x describing the sub s , S2: The charging process safety information sample library includes three categories: safety mode, safety risk warning mode and safety risk mode. During the charging process of new energy vehicles, the charging equipment interacts with the BMS to obtain the battery type and SOC value, and then selects the x corresponding to the battery type and SOC in the information library. s Samples, calculate the average vector m of the pattern class samples respectively sj , Where: s is the state of charge (SOC) value of the vehicle battery, ranging from 0-100, w j There are three types of sample libraries, N sj It is class w j The number of pattern vectors; S3: During the vehicle charging process, the charging equipment interacts with the BMS to obtain the vehicle's real-time battery pack information and convert the battery pack information into s The corresponding information is brought into the minimum distance classification decision function d j (x s ), d 23 (x s )>0 and d 12 (x s )>0, it is identified as a safe mode, otherwise it is identified as a security risk mode and enters the alarm state. d ij (x s )=d i (x s )-d j (xs) j=1,2,3 S4: After the inherent safety risk of the battery is determined, the dynamic safety risk during charging is determined.

2. According to claim 1, a safety risk perception algorithm for vehicle DC charging equipment during charging, characterized in that: The specific steps of the dynamic safety risk perception algorithm during charging in step S4 are as follows: S1: Construct a three-dimensional battery charging process dynamic risk pattern vector x with the battery pack unit temperature maximum increment x1, battery pack charging energy ratio x2 and battery pack single cell voltage boost ratio x3 as descriptors, Where: ΔT is the battery pack temperature increment per minute, ΔQ% is the ratio of the total charge increment to the total battery capacity of the vehicle, ΔSOC is the SOC increment during the charging process, max(ΔV s ) is the maximum battery pack single cell voltage increment during the charging process, It is the mean value of the battery pack single cell voltage increment during the charging process; S2: Select the dynamic risk pattern vector x samples corresponding to the battery class in the information database, and calculate the average vector m of the pattern class samples respectively j and the covariance matrix C j , S3: The corresponding information of x in the battery pack information is brought into the Bayesian classification decision function b j (x), b 23 (x)>0 and b 12 (x)>0, it is identified as a safe mode, otherwise it is identified as a security risk mode and enters the alarm state. b ij (x)=b i (x)-b j (x) j=1,2,3 Where: P(w j ) is the pattern class w j Probability of occurrence.

Images