Battery thermal runaway detection method for fire monitoring of electrochemical energy storage power station

Abstract

The invention discloses a battery thermal runaway detection method for fire monitoring of an electrochemical energy storage power station, which comprises the following steps: completing battery thermal runaway detection through a battery thermal runaway alarm test box, and collecting gas information generated during battery thermal runaway; the method comprises the following steps: recording gas concentration time sequence change conditions in a battery thermal runaway process, and modeling gas concentration time sequence information by adopting a long-short-term memory network to obtain a battery thermal runaway detection model; and the battery thermal runaway detection model is combined with multi-source information of carbon monoxide, smoke density and hydrogen to realize battery thermal runaway detection. The method not only can improve the precision of battery thermal runaway detection for fire monitoring of the electrochemical energy storage power station, but also has good interpretability, and can be directly applied to a battery thermal runaway detection system.

Description

Technical Field

[0001] The present invention relates to the field of monitoring, and particularly to a method for detecting battery thermal runaway for fire monitoring of an electrochemical energy storage power station. Background Art

[0002] With the implementation of the national energy strategy, electrochemical energy storage power stations have developed rapidly in recent years. As a solution for high energy density storage, lithium-ion batteries have been widely used in electrochemical energy storage power stations. Once a lithium-ion battery undergoes thermal runaway, it is extremely likely to cause chain reactions such as jet fires and explosions. The remarkable characteristics of accidents in electrochemical energy storage power stations are the rapid development of accidents, great accident hazards, ineffective post-alarm and fire-fighting interventions, and the batteries are extremely likely to reignite and cause chain explosions.

[0003] From a large number of cases at home and abroad, it can be seen that for large-scale thermal runaway fires, the fire extinguishing effect is not ideal, and generally, losses are reduced by controlling the combustion. Therefore, the prevention and control of battery thermal runaway must be advanced, and the detection of early characteristic parameters of battery thermal runaway and the emergency response of linkage control are particularly important.

[0004] Using gas detectors to detect the gases released during battery thermal runaway is an effective early detection method. Some studies have shown that the effectiveness of different detectors for warning battery thermal runaway. During the process of battery thermal runaway, characteristic gas detectors such as hydrogen and carbon monoxide alarm earlier than other sensors. Therefore, currently, battery thermal runaway warning is mainly carried out based on information such as carbon monoxide gas concentration, hydrogen concentration, and ambient temperature. However, these battery thermal runaway detection methods based on gas information threshold determination have great limitations and are easily affected by the threshold setting. Therefore, the present invention designs a battery thermal runaway alarm test chamber for thermal runaway tests and research of lithium-ion batteries, and uses a long short-term memory network to achieve intelligent detection of battery thermal runaway. Summary of the Invention

[0005] The purpose of the present invention is to provide a method for detecting battery thermal runaway for fire monitoring of an electrochemical energy storage power station.

[0006] To achieve the above object, the present invention is implemented according to the following technical solution:

[0007] The present invention includes the following steps:

[0008] Complete battery thermal runaway detection through a battery thermal runaway alarm test chamber, and collect the gas information generated during battery thermal runaway;

[0009] By recording the time series change of gas concentration during the battery thermal runaway process, use a long short-term memory network to model the gas concentration time series information to obtain a battery thermal runaway detection model;

[0010] The battery thermal runaway detection is realized by combining the battery thermal runaway detection model with multi-source information of carbon monoxide, smoke density, and hydrogen.

[0011] Further, the method for the gas concentration time series information includes:

[0012] Using the gas concentration time series information for battery thermal runaway detection;

[0013] Represent the gas concentrations of carbon monoxide, smoke density, and hydrogen as a one-dimensional vector for the time series information, and the expression is:

[0014] x = (x0, x1, …, x t , …, x N )

[0015] where the length of the sequence data is N, the gas input data at time t is x t , and the length of the sequence data can be adjusted according to the data sampling frequency and the detection sensitivity requirements;

[0016] Detect the battery thermal runaway state according to the time series information to obtain the detection sequence information, and the expression is:

[0017] y = (y0, y1, …, y t , …, y N )

[0018] where the battery thermal runaway detection result at time t is y t .

[0019] Further, the method for the battery thermal runaway detection includes:

[0020] Using the long short-term memory network, abbreviated as LSTM, as the battery thermal runaway detection model;

[0021] LSTM enables the network to process time series information through a gating mechanism, and retains and transmits important information in the time series through the gating mechanism;

[0022] The gate structure of LSTM includes a forget gate, an input gate, and an output gate. The state of LSTM includes a hidden state and a cell state. Among them, the forget gate determines whether to delete information from the cell state, and the output result is between 0 and 1. 0 means complete forgetting, and 1 means retention; the input gate determines whether to add new information to the cell state, and the output gate determines the information output in the hidden state;

[0023] The specific expression of LSTM is:

[0024] f t = σ(W f · [h t-1 , x t+b f )

[0025] i t =σ(W i ·[h t-1 ,x t +b i )

[0026]

[0027] u t =σ(W u ·[h t-1 ,x t +b u )

[0028] h t =u t *tanh(C t )

[0029] where the forget gate is f t , the input gate is i t , the output gate is u t , the sigmoid function is σ(·), the input information at time t is x t , the hidden state at time t - 1 is h t-1 , the hidden state at time t is h t , the hyperbolic tangent function is tanh(·), the weight matrix of the forget gate is W f , the bias term of the forget gate is b f , the weight matrix of the forget gate is W i , the bias term of the forget gate is b i , the weight matrix of the forget gate is W u , the bias term of the forget gate is b u , the weight matrix of the forget gate is W c , the bias term of the forget gate is b c , the cell state at time t - 1 is C t-1 , the cell state at time t is C t , the updated value of the cell state at time t is

[0030] Furthermore, the method for detecting battery thermal runaway includes:

[0031] Detecting the battery thermal runaway situation using the output information of the hidden state of the neural network;

[0032] The battery thermal runaway detection model is divided into a three-layer structure; the first layer is the time series input information of carbon monoxide, hydrogen, and smoke density; the second layer is a neural network model based on the LSTM module; the third layer is the detection result of battery thermal runaway; among which, the battery thermal runaway detection result is sequence information.

[0033] Second, a battery thermal runaway detection device for fire monitoring of an electrochemical energy storage power station according to claim 1, comprising:

[0034] The battery thermal runaway alarm test chamber includes a single cell battery for testing, a temperature-adjustable heating plate, a screen, a test working area, a hydrogen detector, a carbon monoxide detector, a camera, an optical smoke density meter transmitting and receiving device, and an optical smoke density meter reflecting device; the temperature-adjustable heating plate is arranged under the single cell battery for testing, and the screen is placed between the temperature-adjustable heating plate and the test working area; the test working area is provided with a detection device sample, an optical smoke density meter transmitting and receiving device; the hydrogen detector and the carbon monoxide detector are arranged on both sides of the detection device sample; the optical smoke density meter transmitting and receiving device and the optical smoke density meter reflecting device are correspondingly arranged on both sides of the battery thermal runaway alarm test chamber;

[0035] The hydrogen detector, the optical smoke density meter transmitting and receiving device, and the carbon monoxide detector are respectively connected to a microcontroller, and the microcontroller is connected to a host computer.

[0036] Furthermore, a battery thermal runaway detection device for fire monitoring of an electrochemical energy storage power station, the carbon monoxide detector, the optical smoke density meter, and the hydrogen detector are used to collect the gas information generated during battery thermal runaway.

[0037] The beneficial effects of the present invention are:

[0038] The present invention is a battery thermal runaway detection method for fire monitoring of an electrochemical energy storage power station. Compared with the prior art, the present invention has the following technical effects:

[0039] The present invention provides a battery thermal runaway detection method for fire monitoring of an electrochemical energy storage power station. In view of the fact that gas release is likely to occur during the thermal runaway of lithium-ion batteries, the present invention collects gas information through detectors and processes sequence information through a neural network model to realize the intelligent judgment of the battery thermal runaway state. Solve the problem of relying on threshold judgment during battery thermal runaway detection. Description of the Drawings

[0040] Figure 1 It is a step flow chart of a battery thermal runaway detection method for an electrochemical energy storage power station according to the present invention;

[0041] Figure 2Schematic diagram of the structure of a battery thermal runaway alarm test chamber according to the present invention;

[0042] In the figure: 1. Battery thermal runaway alarm test chamber; 2. Single cell for testing; 3. Temperature-adjustable heating plate; 4. Sieve; 5. Test working area; 6. Probe device sample; 7. Hydrogen detector; 8. Carbon monoxide detector; 9. Camera; 10. Optical smoke density meter transmitting and receiving device; 11. Optical smoke density meter reflecting device. Specific embodiments

[0043] The present invention will be further described below through specific embodiments. The illustrative embodiments and descriptions of the present invention are used to explain the present invention, but do not limit the present invention.

[0044] A battery thermal runaway detection method for electrochemical energy storage power station fire monitoring according to the present invention includes the following steps:

[0045] As Figure 1 shown, in this embodiment, it includes the following steps:

[0046] Complete battery thermal runaway detection through a battery thermal runaway alarm test chamber, and collect the gas information generated during battery thermal runaway;

[0047] By recording the time series change of gas concentration during battery thermal runaway, use a long short-term memory network to model the time series information of gas concentration, and obtain a battery thermal runaway detection model;

[0048] Use the battery thermal runaway detection model to combine multi-source information of carbon monoxide, smoke density, and hydrogen to achieve battery thermal runaway detection;

[0049] In actual evaluation, use a test chamber to detect the thermal runaway of lithium-ion batteries. Before the test, the battery should be fully charged, and then the battery under test is placed in the test chamber for testing.

[0050] In this embodiment, the method for the time series information of gas concentration includes:

[0051] Use the time series information of gas concentration for battery thermal runaway detection;

[0052] Represent the gas concentrations of carbon monoxide, smoke density, and hydrogen by one-dimensional vectors to represent the time series information. The expression is:

[0053] x = (x0, x1, …, x t , …, x N )

[0054] where the length of the sequence data is N, and the gas input data at time t is x t, the length of the sequence data can be adjusted according to the data sampling frequency and the requirements of detection sensitivity;

[0055] Detect the thermal runaway state of the battery according to the time series information to obtain the detection sequence information, and the expression is:

[0056] y = (y0, y1, …, y t , …, y N )

[0057] Among them, the thermal runaway detection result of the battery at time t is y t .

[0058] In this embodiment, the method for detecting the thermal runaway of the battery includes:

[0059] Use the long short-term memory network, abbreviated as LSTM, as the detection model for the thermal runaway of the battery;

[0060] LSTM enables the network to process time series information through a gating mechanism, and retains and transmits important information in the time series through the gating mechanism;

[0061] The gate structure of LSTM includes a forget gate, an input gate, and an output gate. The state of LSTM includes a hidden state and a cell state. Among them, the forget gate determines whether to delete information from the cell state, and the output result is between 0 and 1. 0 means complete forgetting, and 1 means retention; the input gate determines whether to add new information to the cell state, and the output gate determines the output of the information in the hidden state;

[0062] The specific expression of LSTM is:

[0063] f t = σ(W f · [h t-1 , x t + b f )

[0064] i t = σ(W i · [h t-1 , x t + b i )

[0065]

[0066] u t = σ(W u · [h t-1 , x t + b u )

[0067] h t = u t * tanh(Ct )

[0068] where the forgetting gate is f t , the input gate is i t , the output gate is u t , the sigmoid function is σ(·), the input information at time t is x t , the hidden state at time t-1 is h t-1 , the hidden state at time t is h t , the hyperbolic tangent function is tanh(·), the weight matrix of the forgetting gate is W f , the bias term of the forgetting gate is b f , the weight matrix of the forgetting gate is W i , the bias term of the forgetting gate is b i , the weight matrix of the forgetting gate is W u , the bias term of the forgetting gate is b u , the weight matrix of the forgetting gate is W c , the bias term of the forgetting gate is b c , the cell state at time t-1 is C t-1 , the cell state at time t is C t , the updated value of the cell state at time t is

[0069] In this embodiment, the method for detecting battery thermal runaway includes:

[0070] Detecting the battery thermal runaway situation using the output information of the hidden state of the neural network;

[0071] The battery thermal runaway detection model is divided into three-layer structure; the first layer is the time series input information of carbon monoxide, hydrogen, and smoke density; the second layer is a neural network model based on the LSTM module; the third layer is the detection result of battery thermal runaway; where the battery thermal runaway detection result is sequence information.

[0072] In the actual evaluation, a battery thermal runaway alarm test chamber is used to collect test data, and a fire detection alarm signal is collected as the test result; a single-layer LSTM structure is used, and the dimension is converted through a fully connected layer for hidden state output. The trained battery thermal runaway detection model is obtained and saved.

[0073] Such as Figure 2As shown in the figure, the battery thermal runaway alarm test chamber includes a single battery for testing, a temperature-adjustable heating plate, a sieve, a test working area, a hydrogen detector, a carbon monoxide detector, a camera, an optical smoke density meter transmitting and receiving device, and an optical smoke density meter reflecting device; a temperature-adjustable heating plate is arranged under the single battery for testing, and the sieve is placed between the temperature-adjustable heating plate and the test working area; the test working area is provided with a detection device sample, an optical smoke density meter transmitting and receiving device; the hydrogen detector and the carbon monoxide detector are arranged on both sides of the detection device sample; the optical smoke density meter transmitting and receiving device and the optical smoke density meter reflecting device are correspondingly arranged on both sides of the battery thermal runaway alarm test chamber;

[0074] The hydrogen detector, the optical smoke density meter transmitting and receiving device, and the carbon monoxide detector are respectively connected to a microcontroller, and the microcontroller is connected to a host computer;

[0075] In actual evaluation, the battery is placed on the temperature-adjustable heating plate, and the temperature of the heating plate is increased according to the test requirements; the gas detectors are installed at designated positions in the battery thermal runaway alarm test chamber, mainly including a carbon monoxide detector, an optical smoke density meter, and a hydrogen detector;

[0076] Adjust the direction of the camera so that the test personnel can conveniently observe the area where the battery is located. At the start of the test, keep the sample in a normal monitoring state for 5 minutes;

[0077] After the start of the test, start the temperature-adjustable heating plate and control the heating plate to rise to 380 °C at a heating rate of (20 ± 3) °C / min.

[0078] In the present invention, the values of the carbon monoxide detector, the hydrogen detector, and the optical smoke density meter are observed, and timing starts when one of the following conditions is met, and the time when the sample emits a fire detection alarm signal is recorded:

[0079] (1) The carbon monoxide concentration reaches 150×10 -6 volume fraction; (2) The hydrogen concentration reaches 150×10 -6 volume fraction; (3) The m value of the optical smoke density meter reaches 0.3 dB / m.

[0080] In this embodiment, a battery thermal runaway detection device for fire monitoring of an electrochemical energy storage power station, the carbon monoxide detector, the optical smoke density meter, and the hydrogen detector are used to collect the gas information generated during battery thermal runaway.

[0081] In the actual evaluation, the carbon monoxide, hydrogen, and smoke density gas information collected by the battery thermal runaway alarm test chamber; the collected gas information is input into the battery thermal runaway detection model for testing to obtain the detection result of battery thermal runaway; the detection result is compared and verified with the fire detection and alarm signal, and the accuracy of the battery thermal runaway detection method is confirmed according to the comparison and verification result.

[0082] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A battery thermal runaway detection method for fire monitoring in electrochemical energy storage power stations, characterized in that: include: The battery thermal runaway detection is completed through the battery thermal runaway alarm test box, and the gas information generated when the battery thermal runaway occurs is collected; By recording the time series changes of gas concentration during battery thermal runaway, the gas concentration time series information is modeled using a long short-term memory network to obtain a battery thermal runaway detection model. The battery thermal runaway detection model is used in combination with multi-source information of carbon monoxide, smoke density and hydrogen to realize battery thermal runaway detection.

2. A battery thermal runaway detection method for fire monitoring in electrochemical energy storage power stations according to claim 1, characterized in that: The method for obtaining gas concentration time series information comprises: Battery thermal runaway detection using gas concentration time series information; The gas concentrations of carbon monoxide, smoke density, and hydrogen are expressed as one-dimensional vectors to represent time series information. The expression is: x=(x0,x1,…,x t ,…,x N ) The length of the sequence data is N, and the gas input data at time t is x t ,The length of sequence data can be adjusted according to the data sampling frequency and ,detection sensitivity requirements; According to the time series information, the battery thermal runaway state is detected and the detection sequence information is obtained. The expression is: y=(y0,y1,…,y t ,…,y N ) The battery thermal runaway detection result at time t is y t .

3. A battery thermal runaway detection method for fire monitoring in an electrochemical energy storage power station according to claim 1, characterized in that: The method for detecting thermal runaway of a battery comprises: Use long short-term memory network, LSTM for short, as a detection model for battery thermal runaway; LSTM uses a gating mechanism to enable the network to process time series information and retain and transmit important information in the time series; The gate structure of LSTM includes forget gate, input gate and output gate. The state of LSTM includes hidden state and cell state. The forget gate determines whether to delete information from the cell state. The output result is between 0 and 1, where 0 means complete forgetting and 1 means keeping. The input gate determines whether to add new information to the cell state, and the output gate determines the output of information in the hidden state. The specific expression of LSTM is: f t =σ(W f ·[h t-1 ,x t ]+b f ) i t =σ(W i ·[h t-1 ,x t ]+b i ) you t =σ(W u ·[h t-1 ,x t ]+b u ) h t =u t *tanh(C t ) The forget gate is f t , the input gate is i t , the output gate is u t , the sigmoid function is σ(·), and the input information at time t is x t , the hidden state at time t-1 is h t-1 , the hidden state at time t is h t , the hyperbolic tangent function is tanh(·), and the weight matrix of the forget gate is W f , the bias term of the forget gate is b f , the weight matrix of the forget gate is W i , the bias term of the forget gate is b i , the weight matrix of the forget gate is W u , the bias term of the forget gate is b u , the weight matrix of the forget gate is W c , the bias term of the forget gate is b c , the cell state at time t-1 is C t-1 , the cell state at time t is C t , the updated value of the cell state at time t is 4. A battery thermal runaway detection method for fire monitoring in electrochemical energy storage power stations according to claim 1, characterized in that: The method for detecting thermal runaway of a battery comprises: Detect battery thermal runaway using the hidden state output information of the neural network; The battery thermal runaway detection model is divided into a three-layer structure; the first layer is the time series input information of carbon monoxide, hydrogen, and smoke density; the second layer is a neural network model based on the LSTM module; the third layer is the detection result of the battery thermal runaway; among which the battery thermal runaway detection result is sequence information.

5. A battery thermal runaway detection device for fire monitoring in an electrochemical energy storage power station according to claim 1, used to execute the method according to any one of claims 1 to 4, characterized in that: include: The battery thermal runaway alarm test box includes a test single battery, an adjustable temperature heating plate, a screen, a test work area, a hydrogen detector, a carbon monoxide detector, a camera, an optical smoke density meter transmitting and receiving device, and an optical smoke density meter reflecting device; an adjustable temperature heating plate is arranged under the test single battery, and the screen is placed between the adjustable temperature heating plate and the test work area; the test work area is provided with a detection device sample, an optical smoke density meter transmitting and receiving device; the hydrogen detector and the carbon monoxide detector are arranged on both sides of the detection device sample; the optical smoke density meter transmitting and receiving device and the optical smoke density meter reflecting device are correspondingly arranged on both sides of the battery thermal runaway alarm test box; The hydrogen detector, the optical smoke density meter transmitting and receiving device and the carbon monoxide detector are respectively connected to a microcontroller, and the microcontroller is connected to a host computer.

6. A battery thermal runaway detection device for fire monitoring in electrochemical energy storage power stations according to claim 5, characterized in that: The carbon monoxide detector, the optical smoke density meter and the hydrogen detector are used to collect the gas information generated when the battery is in thermal runaway.

Images