CN116046989B - Lithium battery fire characteristic gas detection method and system based on array sensor - Google Patents
Lithium battery fire characteristic gas detection method and system based on array sensor Download PDFInfo
- Publication number
- CN116046989B CN116046989B CN202310033519.9A CN202310033519A CN116046989B CN 116046989 B CN116046989 B CN 116046989B CN 202310033519 A CN202310033519 A CN 202310033519A CN 116046989 B CN116046989 B CN 116046989B
- Authority
- CN
- China
- Prior art keywords
- gas
- lithium battery
- characteristic
- response
- sensor
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- WHXSMMKQMYFTQS-UHFFFAOYSA-N Lithium Chemical compound [Li] WHXSMMKQMYFTQS-UHFFFAOYSA-N 0.000 title claims abstract description 59
- 229910052744 lithium Inorganic materials 0.000 title claims abstract description 59
- 238000001514 detection method Methods 0.000 title claims abstract description 57
- 230000004044 response Effects 0.000 claims abstract description 83
- 238000011084 recovery Methods 0.000 claims abstract description 45
- 239000007789 gas Substances 0.000 claims description 256
- 238000000034 method Methods 0.000 claims description 30
- 238000012360 testing method Methods 0.000 claims description 25
- 239000004065 semiconductor Substances 0.000 claims description 22
- 230000035945 sensitivity Effects 0.000 claims description 21
- 238000002474 experimental method Methods 0.000 claims description 19
- 238000005070 sampling Methods 0.000 claims description 12
- 238000012545 processing Methods 0.000 claims description 11
- 238000012549 training Methods 0.000 claims description 11
- 238000005259 measurement Methods 0.000 claims description 9
- 229910006404 SnO 2 Inorganic materials 0.000 claims description 5
- 229910000625 lithium cobalt oxide Inorganic materials 0.000 claims description 5
- BFZPBUKRYWOWDV-UHFFFAOYSA-N lithium;oxido(oxo)cobalt Chemical compound [Li+].[O-][Co]=O BFZPBUKRYWOWDV-UHFFFAOYSA-N 0.000 claims description 5
- RTAQQCXQSZGOHL-UHFFFAOYSA-N Titanium Chemical compound [Ti] RTAQQCXQSZGOHL-UHFFFAOYSA-N 0.000 claims description 4
- GELKBWJHTRAYNV-UHFFFAOYSA-K lithium iron phosphate Chemical compound [Li+].[Fe+2].[O-]P([O-])([O-])=O GELKBWJHTRAYNV-UHFFFAOYSA-K 0.000 claims description 4
- KFDQGLPGKXUTMZ-UHFFFAOYSA-N [Mn].[Co].[Ni] Chemical compound [Mn].[Co].[Ni] KFDQGLPGKXUTMZ-UHFFFAOYSA-N 0.000 claims description 3
- 238000007599 discharging Methods 0.000 claims description 3
- 238000004140 cleaning Methods 0.000 claims description 2
- 238000012544 monitoring process Methods 0.000 abstract description 5
- 238000013145 classification model Methods 0.000 abstract description 4
- 238000004458 analytical method Methods 0.000 abstract description 3
- 238000013507 mapping Methods 0.000 abstract description 2
- 230000006872 improvement Effects 0.000 description 9
- 230000008569 process Effects 0.000 description 8
- 230000015654 memory Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 238000010438 heat treatment Methods 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 230000006641 stabilisation Effects 0.000 description 2
- 238000011105 stabilization Methods 0.000 description 2
- 238000013022 venting Methods 0.000 description 2
- SOXUFMZTHZXOGC-UHFFFAOYSA-N [Li].[Mn].[Co].[Ni] Chemical compound [Li].[Mn].[Co].[Ni] SOXUFMZTHZXOGC-UHFFFAOYSA-N 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 239000000969 carrier Substances 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000002485 combustion reaction Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004146 energy storage Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 238000002347 injection Methods 0.000 description 1
- 239000007924 injection Substances 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 229910052751 metal Inorganic materials 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000000644 propagated effect Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000001172 regenerating effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 239000000779 smoke Substances 0.000 description 1
- 230000002269 spontaneous effect Effects 0.000 description 1
- 239000000758 substrate Substances 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/0004—Gaseous mixtures, e.g. polluted air
- G01N33/0009—General constructional details of gas analysers, e.g. portable test equipment
- G01N33/0027—General constructional details of gas analysers, e.g. portable test equipment concerning the detector
- G01N33/0031—General constructional details of gas analysers, e.g. portable test equipment concerning the detector comprising two or more sensors, e.g. a sensor array
- G01N33/0034—General constructional details of gas analysers, e.g. portable test equipment concerning the detector comprising two or more sensors, e.g. a sensor array comprising neural networks or related mathematical techniques
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/0004—Gaseous mixtures, e.g. polluted air
- G01N33/0009—General constructional details of gas analysers, e.g. portable test equipment
- G01N33/0062—General constructional details of gas analysers, e.g. portable test equipment concerning the measuring method, e.g. intermittent, or the display, e.g. digital
- G01N33/0067—General constructional details of gas analysers, e.g. portable test equipment concerning the measuring method, e.g. intermittent, or the display, e.g. digital by measuring the rate of variation of the concentration
-
- G01N33/0068—
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
- Y02E60/10—Energy storage using batteries
Abstract
The invention relates to a lithium battery fire characteristic gas detection method and system based on an array sensor. According to the detection method, a gas detection module is built by designing a plurality of gas-sensitive sensors in an array mode, mixed gas samples with different concentration gradients are configured for the gas detection module, a data set associated with the mixed gas is built, a corresponding response recovery curve and fingerprint information thereof are specifically included, an LSTM model is trained, tested and improved, a gas sequence classification model is formed, a mapping relation between the response curve and classification and concentration of the mixed gas is obtained, the gas detection module is finally utilized to regenerate data in an actual environment to be detected, and the gas sequence classification model is utilized to analyze the actual data, so that characteristic gas types and concentrations in the environment to be detected are obtained. The classification and identification result of the detection method has higher analysis accuracy, and can effectively provide data support for monitoring fire of the lithium battery, so that potential safety hazards of the fire are reduced.
Description
Technical Field
The invention relates to the technical field of lithium battery fire monitoring, in particular to a lithium battery fire characteristic gas detection method and system based on an array sensor.
Background
In recent years, with the development of new energy automobiles and industrial upgrading, the development of power lithium batteries is rapid due to the advantages of high energy density, high-rate charge and discharge performance, long cycle life and the like, and the power lithium batteries gradually become the main power battery type adopted by the storage carriers of electric energy and energy storage power stations of electric automobiles. However, spontaneous combustion and explosion accidents of lithium batteries are also in high-rise situations.
Aiming at early detection and early warning of lithium battery fire, the traditional method at present mainly uses a sensor to acquire characteristic parameters in the lithium battery fire process, such as heating position characteristics, smoke flame characteristics and the like based on environmental detection data, and analyzes and pre-judges according to certain expert experience. The error rate and delay of the method are relatively high, and the problem of false alarm and missing report exists in practical application, so that effective monitoring of the lithium battery is difficult, and potential safety hazard is caused.
Disclosure of Invention
Based on the above, the invention provides a method and a system for detecting fire characteristic gas of a lithium battery based on an array sensor, which are necessary to solve the technical problem of low accuracy of fire monitoring of the lithium battery in the prior art.
The invention discloses a lithium battery fire characteristic gas detection method based on an array sensor, which comprises the following steps:
s1, constructing a gas detection module by adopting a plurality of array gas sensors.
The plurality of gas sensors are provided with five types, and correspond to the following five types of lithium battery thermal runaway characteristic gases, namely H 2、CO、CH4、C2H4 and DEC steam respectively.
S2, preparing mixed gas containing five lithium battery thermal runaway characteristic gases according to a preset gas component proportion, and preparing multiple groups of samples with different volumes by using the mixed gas.
S3, placing the gas detection module into an experiment cabin with specific volume under a standard experiment environment, injecting each group of mixed gas samples into the experiment cabin, keeping the preset residence time, discharging, and simultaneously collecting and recording the response of each gas sensor according to a preset sampling period and sampling frequency to obtain a plurality of groups of data sets corresponding to a plurality of gas sensors.
Wherein each set of data comprises a response recovery curve of the corresponding gas sensor for the corresponding characteristic gas in the single concentration gradient mixed gas, and fingerprint information of the gas sensor for the corresponding characteristic gas obtained based on the response recovery curve. The fingerprint information includes response time, slope, maximum sensitivity, average sensitivity, variance, and recovery time of the response recovery curve. Each response recovery curve is made up of several response recovery data points.
S4, judging whether the mixed gas sample which is injected and exhausted currently is the last group of the plurality of groups of samples. And if yes, S5 is performed. Otherwise, returning to the step S3 until the data acquisition of a plurality of groups of different concentration gradient samples is completed.
S5, dividing the obtained data sets into a training set and a testing set, and training, testing and improving the constructed LSTM model to form a gas sequence classification and identification model.
S6, acquiring a plurality of groups of actual measurement data sets corresponding to each gas sensor in an environment to be detected by the gas detection module in a mode of referring to S3, inputting the actual measurement data sets into a gas sequence classification and identification model, and further detecting the characteristic gas types and concentrations in the environment to be detected according to each response recovery curve and fingerprint information thereof.
As a further improvement of the scheme, the gas sensor is a SnO 2 resistor type semiconductor sensor.
As a further improvement of the scheme, the response recovery curve of the gas sensor to the characteristic gas is drawn by collecting the ratio of the sensitive steady-state resistance value R s of the gas sensor to the reference resistance value R 0.
As a further improvement of the above-mentioned scheme, in the built gas detection module, the number of the gas sensors is ten. Wherein the number of each gas sensor is two.
As a further improvement of the above-described scheme, in S2, mixed gases of four different gas component ratios are configured according to the kind of lithium battery. Wherein, lithium battery types include: lithium iron phosphate batteries, nickel manganese cobalt batteries, lithium cobalt oxide batteries, and lithium titanate batteries.
As a further improvement of the above scheme, in S2, the volumes of the plurality of groups of samples are respectively 10mL, 20mL, 50mL and 100mL. Wherein the number of samples per volume is set to 40±2 groups.
As a further improvement of the scheme, in S3, the temperature under the standard experiment environment is 25+/-5 ℃ and the humidity is 70+/-5%.
As a further improvement of the above scheme, in S3, the sampling period is not less than 250S. The sampling frequency was 1Hz.
As a further improvement of the above scheme, in S3, after each data set of a group of mixed gas samples is obtained, air with a preset cleaning time is continuously injected into the experimental cabin, so as to exhaust the residual mixed gas in the experimental cabin.
The invention also discloses a lithium battery fire characteristic gas detection system based on the array sensor, which adopts any one of the above detection methods based on the array sensor. The detection system comprises: and the gas detection module and the data processing module.
The gas detection module is built by adopting a plurality of gas sensors in an array mode.
The data processing module is used for training, testing and improving a pre-constructed LSTM model according to corresponding gas-sensitive response data formed by the gas detection module under lithium battery thermal runaway characteristic mixed gases with different concentration gradients, so as to form a gas sequence classification and identification model. The data processing module is also used for inputting corresponding gas-sensitive response data acquired by the gas detection module in an environment to be detected into the formed gas sequence classification and identification model, and further detecting the characteristic gas types and concentrations in the environment to be detected according to each response recovery curve and fingerprint information thereof.
Compared with the prior art, the technical scheme disclosed by the invention has the following beneficial effects:
1. According to the detection method, a plurality of gas sensors in an array mode are designed to build a gas detection module, the gas detection modules respectively correspond to main characteristic gases of thermal runaway of a lithium battery, so that mixed gas atmospheres with different concentration gradients are formulated for the gas detection module through configuration of samples of the mixed gases, further, a data set which is related to the mixed gases with known concentrations is built, a corresponding response recovery curve and fingerprint information thereof are specifically contained, the built LSTM model is trained, tested and improved through the data sets, a gas sequence classification model is formed, mapping relation between the response curve and classification and concentration of the mixed gases is obtained, finally, the gas detection module can be used for regenerating data in an actual environment to be detected, and the regenerated actual data is analyzed through the formed gas sequence classification model, so that characteristic gas types and concentrations in the environment to be detected are obtained. Experiments show that the classification and identification result of the detection method has higher analysis accuracy, and can effectively provide data support for monitoring fire of the lithium battery, so that the potential safety hazard of the fire is reduced.
In addition, the invention selects the array sensor and acquires the fingerprint characteristic data of the gas response recovery curve, and based on the characteristic data set, the invention can carry out the learning training of a small sample to construct a model.
2. The beneficial effects of the detection system are the same as those of the detection method, and are not repeated here.
Drawings
FIG. 1 is a flow chart of a method for detecting fire characteristic gas of a lithium battery based on an array sensor in embodiment 1 of the present invention;
FIG. 2 is a diagram showing the power-on operation of the SnO 2 semiconductor sensor in example 1 of the present invention;
FIG. 3 is a circuit diagram of the gas sensitivity test of the semiconductor sensor in example 1 of the present invention;
FIG. 4 is a response recovery curve of the gas sensor in example 1 of the present invention;
FIG. 5 is a schematic diagram of the gas-sensitive response acquisition of the array sensor of example 1 of the present invention;
FIG. 6 is a graph showing the response recovery curve of the array sensor in the mixed gas generated by thermal runaway of the lithium iron phosphate battery in example 1 of the present invention;
FIG. 7 is a graph showing the response recovery curve of the array sensor in the mixed gas generated by thermal runaway of the nickel-manganese-cobalt-lithium battery in example 1 of the present invention;
FIG. 8 is a graph showing the response recovery curve of the array sensor in the mixed gas generated by thermal runaway of the lithium cobalt oxide battery in example 1 of the present invention;
FIG. 9 is a graph showing the response recovery curve of the array sensor in the mixed gas generated by thermal runaway of the lithium titanate battery in example 1 of the present invention;
FIG. 10 is a schematic diagram of the structure of an LSTM neural network model;
Fig. 11 is a block diagram of a gas sequence classification recognition model in embodiment 1 of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "or/and" as used herein includes any and all combinations of one or more of the associated listed items.
Example 1
Referring to fig. 1, the present embodiment provides a method for detecting fire characteristic gas of a lithium battery based on an array sensor, which includes steps S1 to S6.
S1, constructing a gas detection module by adopting a plurality of array gas sensors.
The plurality of gas sensors are provided with five types, and correspond to the following five types of lithium battery thermal runaway characteristic gases, namely H 2、CO、CH4、C2H4 and DEC steam respectively. In this embodiment, the gas sensor is a SnO 2 resistor type semiconductor sensor. After the resistance type semiconductor gas sensor is electrified, the resistance value of the gas sensor is suddenly reduced, the resistance value is stabilized after about 10 minutes, and after the resistance type semiconductor gas sensor enters a stable state, the gas detector can enter a gas detection working procedure. The resistance type semiconductor gas sensor is generally provided with a heating component, and the internal heating temperature can reach 200-300 ℃.
Referring to fig. 2 and 3, fig. 2 illustrates the power-on operation of the SnO 2 semiconductor sensor. According to the power-on process and principle of the resistance type semiconductor sensor, a sensor test circuit is designed, and the response of the semiconductor sensor is obtained through measuring related parameters of the circuit, as shown in fig. 3: the reference resistance value R 0 of the semiconductor device, and the gas-sensitive steady-state resistance value R s; the circuit is connected with a fixed resistor R m in series, and the R s value is obtained by measuring the change of R m. When the sensor is stably operated in a clean air environment during operation test, the R 0 value is obtained, after the sensor is stabilized, the tested gas is introduced, the conductivity of the semiconductor material is changed, and the sensitive resistance (R s) of the sensor is also changed. As shown in fig. 3, R s can be obtained according to the change of the conductivity, and the sensitivity of the semiconductor material to the gas can be converted to obtain the gas concentration characterization.
Wherein V c represents a loop voltage; v m represents the voltage of the resistor to be measured.
In the embodiment, the metal semiconductor oxide has different gas sensitivity characteristics to the thermal runaway characteristic gases (H 2、CO、CH4、C2H4 and DEC) of the lithium battery, based on the gas-sensitive working principle of the resistance type semiconductor gas sensor, a circuit measurement method is adopted to record the responses of the semiconductor sensitive material to different gases and the concentrations thereof, the response curve of the semiconductor sensor to the gases is obtained, the characteristic parameters are extracted, and the method is used for researching the identification method of the thermal runaway gases.
As shown in fig. 4, based on the response and recovery characteristic curves of the semiconductor device to the gas, basic parameters such as response time, recovery time, response sensitivity and the like can be obtained, and fingerprint information of a specific sensor to the specific gas can be obtained based on the curves, so that the method is used for identifying the thermal runaway gas of the lithium battery.
Referring to fig. 5, in the embodiment, the design of the gas detection module (i.e. the array sensor) is based on the sensor gas sensitivity testing principle, 5 kinds of sensors are selected for 1001-1002-1004-1006-1011 kinds of sensors, two sensors are selected for each kind as the sensing module of the sensor array, an AD acquisition circuit of 10 paths of sensors is designed, the AD acquisition circuit is periodically sampled and buffered by a microprocessor, and response curves of the 10 sensors to different kinds of single gas and mixed gas are recorded and stored by a serial port, so as to construct a gas measurement system of the sensor array.
As shown in FIG. 5, the sensor array is placed in the mixed gas, generates resistance change and enters the 10 AD acquisition circuits of the microprocessor in a voltage measurement mode, the microprocessor samples voltage and converts the voltage into binary data, the binary data is cached in the microprocessor, the acquired signals are subjected to low-pass filtering and average value filtering, the microprocessor outputs the acquired voltage signals through the serial port at the frequency of 1Hz, the acquired voltage signals are converted into R s/R0, and the gas-sensitive response process curve is stored.
S2, preparing mixed gas containing five lithium battery thermal runaway characteristic gases according to a preset gas component proportion, and preparing multiple groups of samples with different volumes by using the mixed gas.
S3, placing the gas detection module into an experiment cabin with specific volume under a standard experiment environment, injecting each group of mixed gas samples into the experiment cabin, keeping the preset residence time, discharging, and simultaneously collecting and recording the response of each gas sensor according to a preset sampling period and sampling frequency to obtain a plurality of groups of data sets corresponding to a plurality of gas sensors.
Wherein each set of data comprises a response recovery curve of the corresponding gas sensor for the corresponding characteristic gas in the single concentration gradient mixed gas, and fingerprint information of the gas sensor for the corresponding characteristic gas obtained based on the response recovery curve. The fingerprint information includes response time, slope, maximum sensitivity, average sensitivity, variance, and recovery time of the response recovery curve. Each response recovery curve is made up of several response recovery data points.
S4, judging whether the mixed gas sample which is injected and exhausted currently is the last group of the plurality of groups of samples. And if yes, S5 is performed. Otherwise, returning to the step S3 until the data acquisition of a plurality of groups of different concentration gradient samples is completed.
In this embodiment, S2 to S4 belong to a construction process of a dataset, and a mixed gas generated by thermal runaway of lithium iron phosphate (LFP), nickel Manganese Cobalt (NMC), lithium Cobalt Oxide (LCO), lithium titanate (NMC/LTO) batteries mainly consists of CO 2、CO、H2、CH4、C2H4 and the like, and based on a thermal runaway characteristic gas research result, an array sensor gas sensitivity experiment is performed to construct a response recovery dataset by using the mixed gas generated by the above 4 lithium battery experiments as an experimental gas. The main process needs to pay attention to the following points:
1. and (5) experimental gas configuration. The mixed experimental gas was prepared according to the gas composition ratio indicated by the subscript.
TABLE 4-1 ratio of Mixed gas components
2. And (5) testing system design. 10 sensors of 5 semiconductor gas sensors are selected by using array sensor acquisition equipment, namely 1001A, 1001B, 1002A, 1002B, 1004A, 1004B, 1006A, 1006B, 1011A and 1011B. Performing a gas sensitivity experiment; and (3) performing experiments by using an array sensor gas sensitivity test system, and collecting response and recovery curve data of the gas.
3. Different concentration gradient profile data set designs. The gas volumes of the experiments are respectively 10mL, 20mL, 50mL and 100mL, the gas volumes are injected into a gas sensitivity test system with the volume of the experiment cabin being 1L, the gas sensitivity experiment is carried out, and a data set is acquired.
4. And (5) effectively designing. And placing the array sensor into a gas sensitivity measurement system, electrifying for half an hour, and starting experiments by injecting gas after the response of the sensor is stable, so that the obtained data set is ensured to be reliable and effective.
5. Sensor response and data acquisition. The data acquisition process includes gas injection, sensor response to stabilization, gas venting (air venting), sensor recovery to stabilization. The response of the sensor is acquired and recorded by adopting the sampling frequency of 1Hz, and the sampling period is more than 250s; that is, one data set includes a response curve, a stable response curve, and a recovery curve, for a total of no less than 250 time domain data points.
6. The experimental data set was designed as follows: 4 sets of mixed gas, 10 sensors, 4 concentration gradients, 250 per sensor sample dataset, single sample data 250 x 10.
TABLE 4-2 Mixed gas samples and parameter settings
According to the 4 mixed gas test parameter settings listed in Table 4-2, each gas has a concentration of not less than 40 samples, and not less than 643 samples in total can be obtained; and performing response test on characteristic gases of the lithium battery by using 10 sensors, and constructing a thermal runaway characteristic gas sample library of the power lithium battery.
Referring to fig. 6-9, response recovery patterns of gas (20 mL) were measured by an array sensor.
In this embodiment, a response model of the sensor array is obtained using the array sensor, and the response model is described as follows:
An array of M gas sensors is used to analyze N mixed gases, the response of the mth sensor in the sensor array to the nth gas is a random process, and the mathematical description is as follows:
Fm,n(ω,t)=fm,n(ω,t)+δm,n(ω,t),t∈T,ω∈Ω
Where m=1, 2, …, M. n=1, 2, …, N. T is the time space and Ω is the sample space. f m,n (ω, t) is the theoretical curve of the sensor, and δ m,n (ω, t) is the result of the combined influence of the environment and noise.
The response of the sensor array to the nth gas is a random process in the M dimension, the number of which is described as:
Fm(ω,t)=[F1,n(ω,t),F2,n(ω,t),...,Fm,n(ω,t)],t∈T,ω∈Ω
likewise, the combined response of the mth sensor in the array to the N gas mixtures is mathematically described as:
Fm(ω,t)=[A1Fm,1(ω,t)+A2Fm,2(ω,t)+...+ANFm,N(ω,t)+δm,0]
Where δ m,0 is the intrinsic response of the mth sensor and a N is the concentration of the nth gas.
From the above, the mathematical expression of the response of the array to the N mixed gases, the concentration, is:
The analysis by the above mathematical expression can be obtained: the array sensor has unique response to each target gas, and theoretically 5 sensors can be used for detecting and identifying 5 gases constrained by the invention, so that 10 different substrate sensors are used for achieving a better detection and identification effect.
Based on the response curve of the semiconductor gas sensor to the gas, parameters such as response time, slope, maximum sensitivity, average sensitivity, variance, recovery time and the like of the characteristic gas response curve are obtained and used as fingerprint information of the sensor to the characteristic gas, after the information is normalized, the detected gas is classified by using methods such as pattern recognition, deep learning and the like, and effective classification and recognition of the characteristic gas can be realized.
S5, dividing the obtained data sets into a training set and a testing set, and training, testing and improving the constructed LSTM model to form a gas sequence classification and identification model.
The LSTM algorithm, which is fully called Long short-term memory, was proposed by Sepp Hochreiter and ju rgen Schmidhuber in 1997 at the earliest, as a specific form of RNN (Recurrent neural network ), which is a generic term for a series of neural networks capable of processing sequence data; LSTM can store more memories (hundreds of time steps), has more parameters than RNNs that maintain only a single hidden state, can better control which memories are saved at a particular time step and which memories are discarded for extensive recognition classification processing of time-series data, and can be used for recognition classification problems of time-domain curves.
The core idea of LSTM is memory block (memory block) mainly comprising three gates (for gate, input gate, output gate) and one memory cell (cell), as shown in the square frame in fig. 10. The upper horizontal line in the box, called CELL STATE (cell state), is like a conveyor belt and can control the information transfer to the next moment. In fig. 10, t is time, C t is memory cell, h t is state, x t is input, f t is forget gate, i t is refresh gate, and o t is output gate.
The input gate i t receives the current input x t and the final hidden state h t-1 as inputs, and calculates i t according to the following equation:
it=σ(Wixxt+Wihht-1+bi)
Wherein W is a weight matrix and b is a bias.
After calculation, a value of 0 indicates that no information currently entered will enter a unit state, and a value of 1 indicates that all information currently entered will enter a unit state. Then, the following formula will calculate another value, called candidate value. It is used to calculate the current cell state.
The forget gate will perform the following operations: forgetting a threshold of 0 indicates that no information is passed to the computation of C t, a value of 1 means that all the information of C t-1 is propagated to C t.
ft=σ(Wfxxt+Wfhht-1+bf)
Last state h t of LSTM cell is calculated:
ot=σ(Woxxt+Wohht-1+bo)
ht=ottanh(Ct)
The above is a principle explanation of the LSTM network, and the present embodiment classifies and identifies the time domain response curve through LSTM, and analyzes the applicability of the method to the gas classification technology.
Referring to fig. 11, in this embodiment, 10×250=2500 data value samples are constructed based on 250 response recovery data points collected by 10 gas sensors for each test sample. 1 LSTM network model with input layer, 3 hidden layers and 1 output layer is designed, and the final algorithm model is formed through repeated test and improvement.
The LSTM model tested the 642 test samples, the results of which are shown in the following table:
TABLE 4-3 Classification recognition accuracy based on LSTM algorithm
Based on training of the one-dimensional sample and parameter adjustment of the model, an identifiable model is obtained, and identification of characteristic gas can be achieved in terminal equipment.
S6, acquiring a plurality of groups of actual measurement data sets corresponding to each gas sensor in an environment to be detected by the gas detection module in a mode of referring to S3, inputting the actual measurement data sets into a gas sequence classification and identification model, and further detecting the characteristic gas types and concentrations in the environment to be detected according to each response recovery curve and fingerprint information thereof.
The environment to be detected is a sealed space with controllable air inlet and air outlet, the space volume is within 5L, and the gas detection module can be assembled.
The invention discloses a lithium battery fire characteristic gas detection method based on an array sensor. The method mainly comprises the steps of building a lithium battery fire characteristic gas detection system based on the sensitivity characteristic of a semiconductor gas sensor to gas, obtaining the sensitivity, response time, recovery time, response recovery curve and the like of the gas sensor to characteristic gas through single gas testing of the sensor to the lithium battery fire characteristic gas, including gas response testing, gas concentration gradient response testing, gas repeatability testing and the like, obtaining the response characteristic of the gas sensor to target gas, building a gas identification method through a response model, and detecting and identifying the lithium battery fire characteristic gas in a terminal processing platform.
Example 2
The embodiment provides a lithium battery fire characteristic gas detection system based on an array sensor, which adopts the lithium battery fire characteristic gas detection method based on the array sensor in embodiment 1. The detection system comprises: and the gas detection module and the data processing module.
The gas detection module is built by adopting a plurality of gas sensors in an array mode.
The data processing module is used for training, testing and improving a pre-constructed LSTM model according to corresponding gas-sensitive response data formed by the gas detection module under lithium battery thermal runaway characteristic mixed gases with different concentration gradients, so as to form a gas sequence classification and identification model. The data processing module is also used for inputting corresponding gas-sensitive response data acquired by the gas detection module in an environment to be detected into the formed gas sequence classification and identification model, and further detecting the characteristic gas types and concentrations in the environment to be detected according to each response recovery curve and fingerprint information thereof.
The technical features of the above-described embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above-described embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The above examples merely represent a few embodiments of the present invention, which are described in more detail and are not to be construed as limiting the scope of the invention. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the invention, which are all within the scope of the invention. Accordingly, the scope of the invention should be assessed as that of the appended claims.
Claims (10)
1. The lithium battery fire characteristic gas detection method based on the array sensor is characterized by comprising the following steps of:
s1, constructing a gas detection module by adopting a plurality of gas sensors in an array mode;
The plurality of gas sensors are provided with five types, and correspond to the following five types of lithium battery thermal runaway characteristic gases respectively, namely H 2、CO、CH4、C2H4 and DEC steam;
S2, preparing mixed gas containing five kinds of lithium battery thermal runaway characteristic gases according to a preset gas component proportion, and preparing multiple groups of samples with different volumes by using the mixed gas;
S3, placing the gas detection module into an experiment cabin with specific volume under a standard experiment environment, injecting each group of mixed gas samples into the experiment cabin, keeping the preset residence time, and discharging, and simultaneously collecting and recording the response of each gas sensor according to a preset sampling period and a preset sampling frequency to obtain a plurality of groups of data sets corresponding to a plurality of gas sensors;
Wherein each group of data sets comprises a response recovery curve of the corresponding gas sensor to the corresponding characteristic gas in the single concentration gradient mixed gas, and fingerprint information of the gas sensor to the corresponding characteristic gas, which is obtained based on the response recovery curve; the fingerprint information comprises response time, slope, maximum sensitivity, average sensitivity, variance and recovery time of the response recovery curve; each response recovery curve is composed of a plurality of response recovery data points;
s4, judging whether the mixed gas sample which is injected and exhausted at present is the last group of a plurality of groups of samples; if yes, executing S5; otherwise, returning to the step S3 until the data acquisition of a plurality of groups of different concentration gradient samples is completed;
S5, dividing the obtained data sets into a training set and a testing set, and training, testing and improving a constructed LSTM model to form a gas sequence classification and identification model;
S6, acquiring a plurality of groups of actual measurement data sets corresponding to each gas sensor in one environment to be detected by the gas detection module in a reference S3 mode, inputting the actual measurement data sets into the gas sequence classification and identification model, and further detecting the characteristic gas types and concentrations in the environment to be detected according to each response recovery curve and fingerprint information thereof.
2. The method for detecting fire characteristic gas of a lithium battery based on an array sensor according to claim 1, wherein the gas sensor is a SnO 2 resistor type semiconductor sensor.
3. The method for detecting the fire disaster characteristic gas of the lithium battery based on the array sensor according to claim 2, wherein a response recovery curve of the gas sensor to the characteristic gas is drawn by collecting the ratio of a sensitive steady-state resistance value R s of the gas sensor to the gas to a reference resistance value R 0.
4. The method for detecting fire characteristic gas of a lithium battery based on an array sensor according to claim 1, wherein in the built gas detection modules, the number of the gas sensors is ten; wherein the number of each gas sensor is two.
5. The method for detecting fire characteristic gas of a lithium battery based on an array sensor according to claim 1, wherein in S2, mixed gas of four different gas component ratios is configured according to the type of the lithium battery; wherein, the lithium battery type includes: lithium iron phosphate batteries, nickel manganese cobalt batteries, lithium cobalt oxide batteries, and lithium titanate batteries.
6. The method for detecting fire characteristic gas of a lithium battery based on an array sensor according to claim 1, wherein in S2, the volumes of the plurality of groups of samples are 10mL, 20mL, 50mL and 100mL, respectively; wherein the number of samples per volume is set to 40±2 groups.
7. The method for detecting fire characteristic gas of a lithium battery based on an array sensor according to claim 1, wherein in S3, the temperature in the standard experimental environment is 25±5 ℃ and the humidity is 70±5%.
8. The method for detecting fire characteristic gas of a lithium battery based on an array sensor according to claim 7, wherein in S3, the sampling period is not less than 250S; the sampling frequency is 1Hz.
9. The method for detecting fire disaster feature gas of lithium battery based on array sensor according to claim 1, wherein in S3, after each data set of the mixed gas sample is obtained, air with preset cleaning time is continuously injected into the experimental cabin, so as to exhaust residual mixed gas in the experimental cabin.
10. An array sensor-based lithium battery fire characteristic gas detection system, which is characterized in that the method for detecting the fire characteristic gas of the lithium battery based on the array sensor is adopted according to any one of claims 1 to 9; the detection system comprises:
The gas detection module is formed by constructing a plurality of gas sensors in an array mode; and
The data processing module is used for training, testing and improving a pre-constructed LSTM model according to corresponding gas-sensitive response data formed by the gas detection module under lithium battery thermal runaway characteristic mixed gases with different concentration gradients to form a gas sequence classification and identification model; the data processing module is also used for inputting corresponding gas-sensitive response data acquired by the gas detection module in an environment to be detected into the formed gas sequence classification and identification model, and further detecting the characteristic gas types and concentrations in the environment to be detected according to each response recovery curve and fingerprint information thereof.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310033519.9A CN116046989B (en) | 2023-01-10 | 2023-01-10 | Lithium battery fire characteristic gas detection method and system based on array sensor |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310033519.9A CN116046989B (en) | 2023-01-10 | 2023-01-10 | Lithium battery fire characteristic gas detection method and system based on array sensor |
Publications (2)
Publication Number | Publication Date |
---|---|
CN116046989A CN116046989A (en) | 2023-05-02 |
CN116046989B true CN116046989B (en) | 2024-05-03 |
Family
ID=86132824
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310033519.9A Active CN116046989B (en) | 2023-01-10 | 2023-01-10 | Lithium battery fire characteristic gas detection method and system based on array sensor |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116046989B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116718648B (en) * | 2023-08-11 | 2023-11-10 | 合肥中科国探智能科技有限公司 | Method for detecting and identifying thermal runaway gas of battery and alarm device thereof |
CN116759682B (en) * | 2023-08-24 | 2024-02-27 | 宁德时代新能源科技股份有限公司 | Battery, power utilization device and gas detection method of battery |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111595461A (en) * | 2020-06-23 | 2020-08-28 | 中国民用航空飞行学院 | Lithium ion battery thermal runaway fire detection and early warning comprehensive experiment system and method |
KR20220100114A (en) * | 2021-01-06 | 2022-07-15 | 호남대학교 산학협력단 | Environmental monitoring system for batteries loaded on vehicles for storage and transportation |
-
2023
- 2023-01-10 CN CN202310033519.9A patent/CN116046989B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111595461A (en) * | 2020-06-23 | 2020-08-28 | 中国民用航空飞行学院 | Lithium ion battery thermal runaway fire detection and early warning comprehensive experiment system and method |
KR20220100114A (en) * | 2021-01-06 | 2022-07-15 | 호남대학교 산학협력단 | Environmental monitoring system for batteries loaded on vehicles for storage and transportation |
Non-Patent Citations (2)
Title |
---|
基于气体传感器阵列的混合气体定量分析;太惠玲;谢光忠;蒋亚东;;仪器仪表学报;20060728(第07期);全文 * |
空运锂离子电池热解气体探测效果对比分析;张青松;秦帅星;罗星娜;宋广韬;;消防科学与技术;20170315(第03期);全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN116046989A (en) | 2023-05-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN116046989B (en) | Lithium battery fire characteristic gas detection method and system based on array sensor | |
CN112763929B (en) | Method and device for predicting health of battery monomer of energy storage power station system | |
CN107577910B (en) | Vehicle exhaust concentration inversion method based on deep neural network | |
CN109061537A (en) | Electric vehicle lithium ion battery sensor fault diagnosis method based on observer | |
CN113866642A (en) | Lithium ion battery fault diagnosis method based on gradient lifting tree | |
CN108897354B (en) | Aluminum smelting process hearth temperature prediction method based on deep belief network | |
Shang et al. | A hybrid method for traffic incident detection using random forest-recursive feature elimination and long short-term memory network with Bayesian optimization algorithm | |
Jiang et al. | A fault diagnosis method for electric vehicle power lithium battery based on wavelet packet decomposition | |
CN112462001A (en) | Gas sensor array model calibration method for data amplification based on condition generation countermeasure network | |
CN115324698B (en) | Diesel engine tail gas analysis sensing method based on neural network | |
CN114729967A (en) | Method for estimating the state of an energy store | |
CN113255764A (en) | Method, system and device for detecting electrochemical energy storage system fault by using machine learning | |
CN110988263B (en) | Vehicle exhaust concentration estimation method based on improved Stacking model | |
CN114267150A (en) | Early fire detection and early warning method for thermal runaway of lithium battery | |
CN114049525A (en) | Fusion neural network system, device and method for identifying gas types and concentrations | |
Xie et al. | Data-driven diagnosis of multiple faults in series battery packs based on cross-cell voltage correlation and feature principal components | |
CN116400241A (en) | Battery health assessment method and device based on full life cycle operation data | |
CN111505064A (en) | Catalytic combustion type methane sensor service state evaluation method | |
CN116540108B (en) | Method, device, storage medium and equipment for early warning of capacity attenuation of battery cell | |
CN117517963A (en) | Multi-scale lithium ion battery state of charge assessment method based on expansion force | |
CN116859259A (en) | Battery health state estimation method based on integrated learning and real vehicle big data | |
CN116819332A (en) | Battery fault diagnosis method, device, medium and BMS system | |
CN116311739A (en) | Multi-sensor fire detection method based on long-short-term memory network and environment information fusion | |
Wang et al. | A Research of Neural Network Optimization Technology for Apple Freshness Recognition Based on Gas Sensor Array | |
CN116718648B (en) | Method for detecting and identifying thermal runaway gas of battery and alarm device thereof |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant |