Fire detection system

Abstract

To obtain a fire detection system that can quickly detect thermal runaway in lithium-ion batteries and prevent it from escalating into a major incident. [Solution] The system includes an odor detection unit that outputs an odor detection signal composed of multiple channel odor patterns representing a physical quantity corresponding to an odor generated in the fire monitoring area; a storage unit that pre-stores an odor detection signal output from the odor detection unit when gas is released from a lithium-ion battery due to a lithium-ion battery malfunction as a specific factor odor pattern; and a control unit that, during monitoring, determines whether or not a gas causing a fire has been generated in the fire monitoring area based on a comparison result between the odor detection signal output from the odor detection unit and the specific factor odor pattern stored in the storage unit, and outputs the determination result.

Description

Technical Field

[0001] The present disclosure relates to a fire detection system specialized in detecting gases released from a lithium-ion battery due to abnormalities in the lithium-ion battery.

Background Art

[0002] Lithium-ion batteries are installed in a wide range of electronic and electrical devices, including mobile phones, notebook computers, digital cameras / video cameras, and portable music players. In addition, battery energy storage facilities using lithium-ion batteries have also been put into practical use. Thus, lithium-ion batteries have high practicality.

[0003] On the other hand, in lithium-ion batteries, there is a risk of "thermal runaway," which is a serious failure mode in which an uncontrollable overheating state occurs. The main causes of thermal runaway include internal short circuits, overcharging, exposure to excessive external heat, aging of the battery, and defects in battery design and materials. Furthermore, this thermal runaway is a self-sustaining exothermic reaction, and once it occurs, it escalates rapidly.

[0004] The signs and symptoms of thermal runaway include rapid temperature rise, expansion of the battery case, abnormal voltage fluctuations, excessive heat generation during charging or discharging, gas release, and leakage of electrolyte. There is a high risk of smoke, fire, explosion, etc., occurring in a relatively short time after thermal runaway occurs.

[0005] Therefore, it is important to detect the state of thermal runaway at an early stage and quickly carry out fire extinguishing activities before a major incident occurs. Thus, there is a conventional technique in which a temperature sensor is provided in a lithium-ion battery to monitor a rapid temperature rise to detect signs of thermal runaway (see, for example, Patent Document 1).

[0006] According to Patent Document 1, even when no fire signal is output from the smoke detector, temperature information of the lithium-ion battery is acquired, and if it is determined that the temperature exceeds a threshold, a confirmation operation is performed by a surveillance camera, after which a fire is determined to have occurred and firefighting operations are initiated. [Prior art documents] [Patent Documents]

[0007] [Patent Document 1] Japanese Patent Publication No. 2022-150911 [Overview of the project] [Problems that the invention aims to solve]

[0008] As described above, Patent Document 1 enables detection of thermal runaway in a lithium-ion battery based on temperature information and enables fire extinguishing operations.

[0009] However, in order to predict the occurrence of thermal runaway as described in Patent Document 1, a configuration for acquiring temperature information of the lithium-ion battery is essential. In other words, Patent Document 1 requires a temperature sensor to be installed in each lithium-ion battery, which makes the configuration complex and increases the cost.

[0010] Furthermore, in order to enable firefighting operations even when a fire signal cannot be obtained, as described in Patent Document 1, it is essential that administrators visually check images from surveillance cameras. In other words, in this respect as well, Patent Document 1 becomes complex in its configuration, increases in price, and may not be able to implement a quick and accurate response because it involves personal judgment by administrators.

[0011] This disclosure is made to solve the above-mentioned problems and aims to provide a fire detection system that can quickly detect a state in which thermal runaway has occurred in a lithium-ion battery and prevent it from becoming a serious problem. [Means for solving the problem]

[0012] The fire detection system described herein comprises: an odor detection unit that outputs an odor detection signal consisting of multiple channel odor patterns representing a physical quantity corresponding to an odor generated in a fire monitoring area; a storage unit that pre-stores an odor detection signal output from the odor detection unit when gas is released from a lithium-ion battery due to a lithium-ion battery malfunction as a specific factor odor pattern; and a control unit that, during monitoring, determines whether or not a gas causing a fire has been generated in the fire monitoring area based on a comparison between the odor detection signal output from the odor detection unit and the specific factor odor pattern stored in the storage unit, and outputs the determination result. [Effects of the Invention]

[0013] According to this disclosure, a fire detection system can be obtained that can quickly detect a state in which thermal runaway has occurred in a lithium-ion battery and prevent it from becoming a serious problem. [Brief explanation of the drawing]

[0014] [Figure 1] This is a schematic diagram of the fire detection system in Embodiment 1 of the present disclosure. [Figure 2] This is an explanatory diagram illustrating an example of an odor pattern detected by the odor detection unit according to Embodiment 1 of the present disclosure. [Figure 3] This is a flowchart showing the flow of a series of processes executed by the control unit in Embodiment 1 of this disclosure. [Figure 4] This is a schematic diagram of the fire detection system in Embodiment 2 of the present disclosure. [Figure 5] This is an explanatory diagram of the overall configuration when a verification experiment was conducted using the fire detection system according to Embodiment 2 of this disclosure. [Figure 6] This figure shows a list of data summarizing the results of the verification experiment conducted in Embodiment 2 of this disclosure. [Figure 7]Based on the results of the verification experiment conducted in Embodiment 2 of the present disclosure, it is an explanatory diagram graphing the behavior of the odor detection unit and the smoke detector over time. [Figure 8] It is a flowchart showing the flow of a series of processes executed by the control unit in Embodiment 2 of the present disclosure.

Mode for Carrying Out the Invention

[0015] Hereinafter, a preferred embodiment of the fire detection system of the present disclosure will be described with reference to the drawings. Lithium-ion batteries are provided with a vent function to prevent the accumulation of flammable gas inside the battery or the explosion of the battery body when thermal runaway occurs, and the gas is released by opening the valve.

[0016] Therefore, the fire detection system according to the present disclosure focuses on the gas released during abnormal times such as when a lithium-ion battery undergoes thermal runaway, installs an odor detection unit that outputs an odor detection signal to the fire monitoring area, stores in advance the odor detection signal when gas is released from the lithium-ion battery and the ambient air is polluted as a specific factor odor pattern, and determines the presence or absence of gas generation based on the comparison between the odor detection signal obtained during monitoring and the stored specific factor odor pattern, and outputs the determination result as a technical feature.

[0017] Furthermore, when the fire detection system according to the present disclosure detects the generation of gas in the fire monitoring area based on the odor detection signal, it can change the detection sensitivity of the fire detector installed in the fire monitoring area to a high sensitivity, and realizes a configuration that can detect the occurrence of a fire caused by gas earlier, which is regarded as a further technical feature.

[0018] In the following Embodiments 1 and 2, the "gas" simply expressed as "gas is generated" or "gas is released" means the gas released from the lithium-ion battery due to an abnormality of the lithium-ion battery. Further, the gas released from the lithium-ion battery due to an abnormality of the lithium-ion battery includes electrolyte vapor (off-gas) generated at the initial stage of the failure of the lithium-ion battery.

[0019] Embodiment 1. FIG. 1 is a schematic configuration diagram of a fire detection system according to Embodiment 1 of the present disclosure. The fire detection system 100 according to this Embodiment 1 includes a smell detection unit 10, a control unit 20, and a storage unit 30.

[0020] The smell detection unit 10 monitors the smell of ambient air in the fire monitoring area. In this Embodiment 1, a crystal oscillator type sensor provided with a transducer and a plurality of types of sensitive films is adopted as an example of the smell detection unit 10. Further, in the following description, it is assumed that, as an example, the smell detection unit 10 is provided with five types of sensitive films. That is, the five types of sensitive films correspond to five types of smell sensors.

[0021] Here, the sensitive film is a sensor that senses when it adsorbs a smell component. The sensitive film is provided on the surface of a crystal substrate together with electrodes, and vibrates when a voltage is applied to the electrodes. The vibration frequency of this sensitive film changes according to the weight of the adsorbed smell component when the smell component is adsorbed.

[0022] The transducer detects a change in the vibration frequency generated in the sensitive film, and converts this change amount into data in a form that can be digitally processed.

[0023] The smell detection unit 10 generates a waveform representing the temporal change of the detection value by acquiring the detection value of the smell corresponding to each sensitive film at a prescribed period. In this Embodiment 1, the smell detection unit 10 is provided with five different types of sensitive films.

[0024] Therefore, the odor detection unit 10 can generate five-channel waveforms corresponding to the type of sensitive membrane and output them as an odor detection signal. The pattern formed by combining these five-channel waveforms is called an odor pattern or simply a pattern.

[0025] Furthermore, in systems using multiple sensitive membranes, the ability to distinguish odors can be improved by increasing the number of types of sensitive membranes and adopting odor patterns based on a larger number of channels.

[0026] Furthermore, the odor detection unit 10 applicable to the fire detection system according to this disclosure is not limited to a quartz crystal oscillator type sensor equipped with multiple types of sensitive membranes. Various odor sensors capable of identifying odors using multi-element sensors can be used as the odor detection unit 10. In this case as well, the odor identification ability can be improved by increasing the number of multi-element sensors and adopting odor patterns based on a larger number of channels.

[0027] Figure 2 is an explanatory diagram illustrating the odor patterns detected by the odor detection unit 10 according to Embodiment 1 of this disclosure. In Figure 2, the detection results for the following three types of odor patterns, Pattern A, Pattern B, and Pattern C, are illustrated.

[0028] Pattern A: Odor pattern based on 5-channel odor detection signals when smoke is emitted due to server or production equipment failure in a fire monitoring area. Pattern B: Odor pattern based on 5-channel odor detection signals when smoke is emitted due to an air conditioning malfunction in a fire monitoring area. Pattern C: Odor pattern based on 5-channel odor detection signals when gas is released from a lithium-ion battery in a fire monitoring area.

[0029] In Figure 2, the odor detection unit 10 was able to obtain a pattern in which, when smoke is emitted according to patterns A and B, channels 1 and 2 of the 5 channels react strongly, while the other channels react moderately or weakly.

[0030] Furthermore, regarding the odor detected when gas is released from the lithium-ion battery according to pattern C, the odor detection unit 10 was able to obtain a pattern in which, compared to patterns A and B, channels 1 and 2 of the 5 channels reacted less, while channel 5 reacted more.

[0031] Therefore, pattern C can be clearly distinguished from patterns A and B, and smoke emission can be distinguished from gas generation from the lithium-ion battery. In this way, the odor detection unit 10 is equipped with five different types of sensitive membranes suitable for identifying the odor of gas emitted from the lithium-ion battery, and can output an odor detection signal composed of multiple channel odor patterns so that it can identify various odors.

[0032] Furthermore, by selecting an appropriate combination of multiple sensitive membranes depending on the environment of the fire monitoring area, the odor detection unit 10 can output an odor detection signal that can easily identify desired factors, including gas, in order to improve the accuracy of distinguishing gases emitted from lithium-ion batteries from other factors.

[0033] The control unit 20 is a controller that determines whether or not the fire monitoring area is in an abnormal state based on the odor detection signal. Specifically, the control unit 20 stores in the storage unit 30 the odor detection signal output from the odor detection unit 10 when gas is generated by the lithium-ion battery as a specific factor odor pattern.

[0034] Furthermore, the control unit 20 monitors the odor detection signal output from the odor detection unit 10 during monitoring, compares it with a stored specific factor odor pattern, and if a specific factor odor pattern is detected as an odor detection signal, it determines that a gas causing a fire has been generated in the fire monitoring area and outputs the determination result.

[0035] In the specific example described above, the pattern generated by the five odor detection signals when gas is generated is stored in the storage unit 30 as a "specific factor odor pattern," and the gas generation is identified based on the odor detection signal. However, by also storing factors other than gas as "specific factor odor patterns" in the storage unit 30, the control unit 20 can identify the desired factor.

[0036] Furthermore, by storing each of the multiple factors as a "specific factor odor pattern" in the memory unit 30, it is possible to identify each of the multiple factors.

[0037] Furthermore, instead of comparing the odor detection signal obtained from the odor detection unit 10 with specific odor patterns pre-stored in the memory unit 30, the control unit 20 can also employ a method of identifying factors using machine learning with AI (Artificial Intelligence).

[0038] When using machine learning, the control unit 20 is pre-equipped with a machine learning discrimination function, which is obtained by training the control unit 20 with various odor patterns pre-generated by the odor detection unit 10 as training data, for each specific factor that is to be identified.

[0039] Furthermore, the control unit 20, using the odor pattern output from the odor detection unit 10 as an input parameter during monitoring, can identify specific factors corresponding to the odor pattern using a machine learning discrimination function.

[0040] Furthermore, if the control unit 20 detects a specific factor odor pattern as an odor detection signal received from the odor detection unit 10, it outputs information as a determination result that identifies the occurrence of the specific factor, and although not shown in Figure 1, it can trigger an alarm from the alarm unit.

[0041] The control unit 20 may also be equipped with communication equipment that enables wireless or wired communication. If communication equipment is provided, the control unit 20 can transmit the determination result to an external device (not shown) located remotely, a mobile terminal owned by a facility manager, etc.

[0042] Figure 3 is a flowchart showing the flow of a series of processes performed by the control unit 20 in Embodiment 1 of this disclosure. In step S301, the control unit 20 monitors the odor detection signal output from the odor detection unit 10 during monitoring.

[0043] Next, in step S302, the control unit 20 determines whether or not the specific factor odor pattern stored in the storage unit 30 was detected as the odor detection signal monitored during monitoring, based on the comparison result between the odor detection signal output from the odor detection unit 10 during monitoring and the specific factor odor pattern stored in the storage unit 30 in advance.

[0044] If the control unit 20 determines that a specific odor pattern has been detected as an odor detection signal, it proceeds to step S303. If it determines that a specific odor pattern has not been detected as an odor detection signal, it returns to step S301 and continues monitoring.

[0045] By pre-storing odor patterns related to multiple types of factors in the storage unit 30 as specific factor odor patterns, the control unit 20 can also identify multiple types of factors.

[0046] When specializing in detecting gases emitted from lithium-ion batteries, the accuracy of gas detection can be improved by pre-programming specific odor patterns that can identify the source of the gas, depending on the environment of the fire monitoring area.

[0047] When the process proceeds to step S303, the control unit 20 determines, based on the odor detection signal, that gas has been generated in the fire monitoring area, outputs the determination result, and terminates the series of processes.

[0048] As described above, Embodiment 1 includes a configuration that, based on the odor detection signal, stores in advance the odor pattern related to the gas for which factor analysis is to be performed as a specific factor odor pattern according to the environment of the fire monitoring area, and determines whether or not the odor detection signal during monitoring is a specific factor odor pattern.

[0049] As a result, a fire detection system can be realized that can quickly determine when the ambient air in a fire monitoring area has changed to an abnormal state due to the generation of gas. In other words, it becomes possible to quickly identify the gas released from lithium-ion batteries based on odor patterns, depending on the installation environment of various fire monitoring areas.

[0050] Regarding the odor detection unit, instead of permanently installing it in the fire monitoring area, it is also possible to adopt a handheld configuration that can be carried by administrators or other personnel.

[0051] Furthermore, in the above-described embodiment 1, a specific example was described in which a quartz crystal oscillator type sensor is used as the odor detection unit, which is equipped with multiple types of sensitive membranes and outputs the odor patterns detected by each sensitive membrane as an odor detection signal. However, the odor detection unit according to this disclosure is not limited to a sensor having such a configuration with multiple types of sensitive membranes.

[0052] For example, a semiconductor sensor like the one described below can be used as the odor detection unit. A semiconductor sensor employs a method that utilizes the fact that the resistance value of a semiconductor changes due to the adsorption of odor molecules on the semiconductor surface and the resulting surface reaction. On the upper surface of the sensor unit, an oxide semiconductor SnO2 (tin oxide), which is the gas-sensitive material, is formed on an alumina substrate. On the other hand, a heater for heating is attached to the lower surface of the sensor unit.

[0053] When placed in clean air, oxygen adsorbs onto the surface of the gas-sensing element. Because oxygen has electron affinity, it captures electrons in the gas-sensing element. At this time, the flow of electrons is obstructed, and the electrical resistance inside the gas-sensing element increases.

[0054] On the other hand, when placed in an odorous gas, an oxidation reaction occurs between the odorous gas and adsorbed oxygen on the surface of the gas-sensitive element, removing the adsorbed oxygen. As a result, electrons become more mobile, and the electrical resistance decreases.

[0055] Many odor molecules are reducing gases, and when these gases come into contact with a gas-sensing element, their electrical resistance decreases. Therefore, since the electrical resistance changes depending on the odor molecules present in the atmosphere, the change in electrical resistance can be converted into a voltage value, which can then be used as an odor detection signal to electrically identify and detect odors.

[0056] Furthermore, by configuring semiconductor sensors as multi-element sensors, it becomes possible to improve the accuracy of identifying specific factors.

[0057] In other words, the odor detection unit according to this disclosure can be any method that can output an odor detection signal composed of multiple channels of odor patterns, which is a physical quantity corresponding to the odor generated in the fire monitoring area. The desired odor detection unit can be adopted depending on the installation site, application, etc. It is also possible to use multiple odor detection units in combination.

[0058] Embodiment 2. This second embodiment describes a fire detection system that further includes a fire detector installed in a fire monitoring area, and has a configuration that dynamically changes the sensitivity of the fire detector when it is determined that gas has been generated based on an odor detection signal.

[0059] In this embodiment 2, a smoke detector is used as a specific example of a fire detector. However, other fire detectors besides smoke detectors can be used if the detection sensitivity can be changed.

[0060] Figure 4 is a schematic diagram of the fire detection system in Embodiment 2 of the present disclosure. The fire detection system 101 according to Embodiment 2 comprises an odor detection unit 10, a control unit 20, a storage unit 30, a smoke detector 40, a sampling tube 50, and a suction fan 60.

[0061] The fire detection system 101 according to this second embodiment, shown in Figure 4, differs from the fire detection system 100 according to the first embodiment, shown in Figure 1, in that it further includes a smoke detector 40, a sampling tube 50, and a suction fan 60. Therefore, this explanation will focus on these differences when describing the fire detection system 101 according to this second embodiment.

[0062] The sampling tube 50 is laid out in the indoor area of ​​the fire monitoring area S, such as the ceiling, and multiple sampling holes 51 are provided in order to draw in air from the fire monitoring area S.

[0063] The smoke detector 40 is an example of a fire detector and corresponds to a highly sensitive smoke sensor. In Figure 4, a sampling tube 50 is connected to the suction section 41 of the smoke detector 40, and a suction fan 60 is connected to the exhaust section 42 of the smoke detector 40.

[0064] A suction fan 60 located downstream of the smoke detector 40 draws ambient air from within the fire monitoring area S into the sampling tube 50 and brings it into the smoke detector 40. Through the action of the suction fan 60, the smoke detector 40 acquires ambient air from the fire monitoring area S via multiple sampling holes 51 in the sampling tube 50 and optically monitors the acquired ambient air.

[0065] Furthermore, the smoke detector 40 outputs a fire signal if, as a result of monitoring, the ambient air deviates from the acceptable range indicating a normal state and changes to an abnormal state. For example, the smoke detector 40 can determine with high accuracy whether or not smoke is present by sensing whether or not smoke particles are present in the acquired ambient air, and outputs a fire signal if it determines that smoke has been generated.

[0066] In this way, fire detectors, such as the smoke detector 40, monitor the fire monitoring area S, calculate fire-related features, and output a fire signal if the features exceed a predetermined threshold.

[0067] In Figure 4, the suction fan 60 is located in the exhaust section 42 downstream of the smoke detector 40. However, the suction fan 60 is not limited to this position, and it is also possible to install it in the suction section 41 upstream of the smoke detector 40. In other words, as long as a configuration can be achieved in which air from the fire monitoring area S is drawn in through multiple sampling holes 51 and supplied to the suction section 41 of the smoke detector 40, the suction fan 60 is not limited to the arrangement shown in Figure 4.

[0068] As described in the previous embodiment 1, the control unit 20 has a first control function that, during monitoring, determines whether or not gas has been generated in the fire monitoring area S based on a comparison between the odor detection signal output from the odor detection unit 10 and a specific factor odor pattern stored in the storage unit 30, and outputs the determination result.

[0069] Furthermore, the control unit 20 according to this second embodiment further performs the following two control functions in relation to the smoke detector 40. Second control function: When it is determined that gas has been generated based on the detection results from the odor detection unit 10, the detection threshold used in the determination process of the smoke detector 40 can be dynamically set in multiple stages, and the detection sensitivity can be changed to a higher sensitivity by lowering it from the current setting value.

[0070] Third control function: When a fire signal is received from the smoke detector 40, this function determines that a fire has occurred in the fire monitoring area S and outputs the determination result that a fire has occurred.

[0071] In particular, the control unit 20 according to this second embodiment has a second control function that can dynamically adjust the sensitivity of the smoke detector when it is determined that gas is being generated in the fire monitoring area S based on the odor detection signal. Therefore, the smoke detection sensitivity when gas is being generated can be increased compared to when no gas is being generated, and it becomes possible to detect more quickly when smoke is being generated due to gas.

[0072] Next, the results of the verification experiments regarding odor detection and smoke detection will be explained in detail using Figures 5 to 7.

[0073] Figure 5 is an explanatory diagram of the overall configuration when a verification experiment was conducted using the fire detection system according to Embodiment 2 of this disclosure. In Figure 5, the laboratory where the verification experiment was conducted is designated as the fire monitoring area S, and the diagram shows the arrangement of an odor source 1, a smoke detector 40 connected to a sampling tube 50, an air conditioner 70, and three odor detection units 10(1) to 10(3) within the fire monitoring area S.

[0074] The fire monitoring area S, where the verification experiment was conducted, is as follows: Area: 47.6m 2 Actual volume: 176.1 m³ 3

[0075] Furthermore, the odor detection unit 10(1) and the smoke detector 40 are located on a desk 2 installed on the left side of the laboratory, the odor detection unit 10(2) is located on the right side of the ceiling, and the odor detection unit 10(3) is located on the left side of the ceiling. The odor detection unit 10(1) is located on the exhaust side 42 of the smoke detector 40.

[0076] Furthermore, a sampling tube 50 having multiple sampling holes 51 is positioned on the ceiling and connected to the suction section 41 of the smoke detector 40.

[0077] Furthermore, an air conditioner 70 was installed in the laboratory, and verification experiments were conducted with the air conditioner 70 turned off and with the air conditioner turned on to circulate ambient air.

[0078] Note that Figure 5 omits the illustration of the control unit 20, memory unit 30, and suction fan 60 shown in Figure 4.

[0079] In the verification experiment, at odor source 1, a lithium-ion battery was placed on top of a heat source, and the lithium-ion battery was heated externally by the heat source, forcibly creating a state in which gas was released due to valve opening, and a state in which the lithium-ion battery ignited.

[0080] Figure 6 is a diagram showing a data list summarizing the results of the verification experiment conducted in Embodiment 2 of this disclosure. Figure 6 shows the experimental results for four cases, No. 1 to No. 4.

[0081] Furthermore, in Figure 6, for each case, (A) Status of Lithium Ion Battery (LIB) (B) Reaction start time of each odor detection unit 10(1) to 10(3) (C) Response start time of smoke detector 40 (D) Condition of air conditioner 70 These are shown as corresponding data.

[0082] This section will explain the specific details of each column (A) through (D). Column (A) shows the time from the start of heating of the lithium-ion battery until the valve opens, and the time until rupture and ignition, along with the temperature at that time.

[0083] In column (B), the time it takes for each of the odor detection units 10(1) to 10(3) to detect the odor pattern of the gas is shown as the reaction start time.

[0084] In column (C), the time it takes for the smoke detector 40 to detect smoke with a sensitivity of 0.05% / m, and the time it takes for the smoke detector 40 to output a fire signal as an alarm, are shown as the reaction start time.

[0085] Column (D) indicates whether the air conditioner 70 was ON or OFF in each case.

[0086] In cases No. 1 to No. 3, the results of three verification experiments were conducted under the same conditions with the air conditioner 70 in the OFF state, taking reproducibility into consideration. In case No. 4, the results of one verification experiment were conducted to examine the effect of turning the air conditioner 70 ON, and the results are summarized.

[0087] The following was confirmed from the verification results in Figure 6. (1) Changes in the state of lithium-ion batteries In all cases, approximately 7 minutes elapsed between the time T1 when the lithium-ion battery valve opened and the time T2 when the main unit exploded and caught fire. The lithium-ion battery valve opens when the temperature reaches approximately 158°C to 168°C, and when the temperature rises further to approximately 195°C to 202°C, rupture and ignition occur.

[0088] (2) Regarding the reaction start time of the odor detection unit 10 The odor detection unit 10(1) located in the exhaust section 42 of the smoke detector 40 can detect the odor of gas earlier than the odor detection units 10(2) and 10(3) located on the ceiling. In other words, by identifying the odor based on the air drawn in by the sampling tube, it becomes possible to determine the generation of gas more quickly.

[0089] Specifically, the odor detection unit 10(1) detects the gas odor within one minute of the lithium-ion battery valve opening. Therefore, although it takes about seven minutes from valve opening to explosion and ignition, the odor detection unit 10(1) is able to detect the gas odor at a relatively early stage, within one minute of valve opening.

[0090] In other words, by monitoring the odor of the ambient air and detecting gas generation based on the smell, it becomes possible to detect abnormal conditions about 6 minutes before ignition occurs, enabling prompt and appropriate action to be taken.

[0091] (3) Regarding the reaction start time of the smoke detector 40 The smoke detector 40 outputs a fire signal as an alarm at a time after the lithium-ion battery ruptures and ignites.

[0092] Figure 7 is an explanatory diagram illustrating the behavior of the odor detection unit 10 and the smoke detector 40 over time, based on the results of a verification experiment conducted in Embodiment 2 of this disclosure. The horizontal axis in Figure 7 represents the heating time, which corresponds to the elapsed time since the start of heating of the lithium-ion battery.

[0093] In Figure 7, the vertical axis corresponds to the output values ​​of each channel in the odor detection unit 10 and the smoke concentration detected by the smoke detector 40. Note that the output values ​​of each channel and the smoke concentration are shown after offset adjustment so that the value at time T1 when the valve opens is 0.

[0094] From the graph in Figure 7, we can see the following: (1) Output values ​​of each channel of the odor detection unit 10 The output values ​​of channels 1 (CH1) and 2 (CH2) tend to rise sharply after time T1, when the valve opens. Also, while the output values ​​of channels 1 (CH1) and 2 (CH2) do not change significantly after time T1, they tend to rise sharply after time T2, when ignition occurs. On the other hand, the output value of channel 5 (CH5) does not change as the heating time progresses and remains almost at 0.

[0095] Therefore, if the control unit 20 defines the period from time T1 to time T2 as a prediction region for predicting gas generation, it can detect the gas released after the valve was opened at time T1 early based on the state of the output values ​​of CH1 to CH5 during this period.

[0096] Figure 7 illustrates a specific case where gas generation is detected under the following three conditions. <Decision process using condition 1> By using the condition 1, which is that the output values ​​of CH3 to CH5 are almost zero, the output of CH1 is OV1 or higher, and the output of CH2 is OV2 or higher, as a criterion, the control unit 20 can estimate that gas was generated at time t1.

[0097] <Decision process using condition 2> By using the condition 2 that the output values ​​of CH3 to CH5 are almost zero, the output of CH1 rises further than OV1 to OV3 or higher, and the output of CH2 is OV2 or higher as a criterion, the control unit 20 can estimate that gas was generated at time t2.

[0098] <Decision process using condition 3> By using the condition 3 that the output values ​​of CH3 to CH5 are almost zero, the output of CH1 rises further than OV2 to OV4 or higher, and the output of CH2 is OV2 or higher as a criterion, the control unit 20 can estimate that gas was generated at time t3.

[0099] In this way, by appropriately setting the conditions based on the combination of output values ​​of each channel of the odor detection unit 10 according to the environment of the fire monitoring area S, the control unit 20 can predict the generation of gas in the prediction area after the valve is opened.

[0100] Condition 2 corresponds to a stricter judgment criterion than Condition 1, and Condition 3 corresponds to a stricter judgment criterion than Condition 2. By setting the conditions loosely, the control unit 20 can predict gas generation earlier, and by setting the conditions strictly, although the time it takes to predict gas generation will be delayed, the prediction accuracy can be improved.

[0101] (2) Regarding the smoke concentration detected by the smoke detector 40 The smoke concentration gradually increases after the valve opens at time T1, and increases more rapidly after ignition occurs at time T2. If the initial threshold for outputting a fire signal for the smoke detector 40 is TH1, then the smoke detector will output a fire signal when it reaches time t5.

[0102] In response to this, the control unit 20, when it can predict the generation of gas based on the odor detection signal in the prediction region, dynamically changes the detection sensitivity by lowering the judgment threshold from TH1, making it possible to output a fire signal earlier than time t5.

[0103] In the example shown in Figure 7, the control unit 20 can output a fire signal at time t4, earlier than time t5, by lowering the judgment threshold from TH1 to TH2 when gas generation is predicted.

[0104] The control unit 20 can change the judgment threshold to a highly sensitive setting, but only for a short period of time after gas generation is predicted and the lithium-ion battery is about to ignite. This suppresses the risk of false alarms and increases the sensitivity of the smoke detector 40, making it possible to output a fire signal earlier.

[0105] The control unit 20 can output a determination result containing information to indicate gas generation when gas generation is predicted. Furthermore, the control unit 20 can dynamically change the detection sensitivity of the smoke detector according to the gas generation prediction result, and if it receives a fire signal from the smoke detector 40, it can output a determination result containing information to indicate that a fire signal has been output.

[0106] Therefore, the fire detection system according to this second embodiment makes it possible to quickly carry out appropriate firefighting activities, evacuation guidance, information transmission, etc., based on these determination results when a lithium-ion battery experiences thermal runaway.

[0107] Furthermore, lithium-ion batteries generate more gas when they are highly charged or have a large capacity, which can lead to larger fires. Therefore, the control unit 20 can appropriately set specific odor patterns based on the monitoring results of the charge level or the capacity of the lithium-ion battery, and control the detection sensitivity of the smoke detector 40 to an appropriate value, thereby constructing an appropriate fire detection system that is suitable for the environment of the fire monitoring area S.

[0108] One application of the fire detection system described in this disclosure is in solar power charging equipment. In such charging equipment, multiple lithium-ion batteries are arranged, and if one lithium-ion battery catches fire, there is a risk that adjacent lithium-ion batteries will also catch fire. Therefore, the early detection described in this disclosure will function effectively in such applications.

[0109] Figure 8 is a flowchart showing the flow of a series of processes executed by the control unit 20 in Embodiment 2 of this disclosure. In the flowchart shown in Figure 8, the processes from step S301 to step S303 are the same as those shown in Figure 3 in Embodiment 1.

[0110] In Figure 8, steps S801 to S803, which relate to the control of the smoke detector 40, have been newly added as processing from step S303 onwards. Therefore, the following explanation will focus on the newly added processing from steps S801 to S803.

[0111] After the control unit 20 outputs a determination result indicating that gas has been generated in step S303, it executes the process in step S801.

[0112] In step S801, the control unit 20 dynamically adjusts the sensitivity of the smoke detector 40 according to the progression of the gas detection status. To illustrate with a specific example of the detection status progression shown in Figure 7, the control unit 20 determines that gas has been generated at time t1 using condition 1 as the judgment criterion, and can dynamically change the judgment threshold from TH1 to TH2.

[0113] Furthermore, the control unit 20 monitors the gas detection status by continuously reading the output values ​​of the 5 channels from the odor detection unit 10 after time t1.

[0114] Furthermore, the control unit 20 can, if necessary, change the judgment threshold to a value even lower than TH2 at time t2 when condition 2 is met, or at time t3 when condition 3 is met.

[0115] Next, in step S802, the control unit 20 determines whether or not it has received a fire signal from the smoke detector 40, which has been set to an appropriate sensitivity according to the progression of the gas detection status. If the control unit 20 determines that it has received a fire signal, it proceeds to step S803.

[0116] On the other hand, if the control unit 20 determines that it has not received a fire signal, it returns to step S801 and repeats the processing of steps S801 and S802.

[0117] When the process proceeds to step S803, the control unit 20 determines that a fire has occurred after gas has been generated in the fire monitoring area S, outputs a determination result that includes information indicating that a fire signal has been output, and terminates the series of processes.

[0118] As described above, according to Embodiment 2, the sensitivity of the smoke detector can be changed to a higher setting in accordance with the determination result that gas has been generated, and furthermore, it is possible to dynamically adjust the sensitivity of the smoke detector according to the progression of the gas detection status as needed, thereby constructing a fire detection system that can output appropriate determination results.

[0119] As a result, when a lithium-ion battery experiences thermal runaway, it becomes possible to quickly and appropriately carry out firefighting, evacuation guidance, and information dissemination. In other words, even when thermal runaway occurs in a lithium-ion battery, a fire detection system can be realized that can quickly detect the condition and prevent it from escalating into a major incident. [Explanation of symbols]

[0120] 1 Odor source, 10, 10(1), 10(2), 10(3) Odor detection unit, 20 Control unit, 30 Memory unit, 40 Smoke detector, 41 Suction unit, 42 Exhaust unit, 50 Sampling tube, 51 Sampling hole, 60 Suction fan, 70 Air conditioner, 100, 101 Fire detection system.

Claims

1. An odor detection unit outputs an odor detection signal composed of multiple channels of odor patterns, which is a physical quantity corresponding to the odor generated in the fire monitoring area. A storage unit stores in advance the odor detection signal output from the odor detection unit when gas is released from the lithium-ion battery due to an abnormality in the lithium-ion battery, as a specific factor odor pattern. A control unit that, during monitoring, determines whether or not the gas causing the fire has been generated in the fire monitoring area based on the comparison result between the odor detection signal output from the odor detection unit and the specific factor odor pattern stored in the storage unit, and outputs the determination result. A fire detection system equipped with the following features.

2. The system further includes a fire detector that monitors the aforementioned fire monitoring area, calculates fire-related feature quantities, and outputs a fire signal if the feature quantities exceed a preset threshold. If the control unit determines, based on the detection result by the odor detection unit, that the gas has been generated, it lowers the determination threshold value from the current set value and changes the detection sensitivity of the fire detector to a higher sensitivity. The fire detection system according to claim 1.

3. The aforementioned fire detector has a suction unit and an exhaust unit, and is a smoke detector that draws in air from the fire monitoring area through multiple sampling holes, takes it in as ambient air from the fire monitoring area via the suction unit, optically monitors the taken-in ambient air, and detects the generation of smoke with high sensitivity. The odor detection unit is installed in the suction or exhaust section of the smoke detector and outputs a physical quantity corresponding to the odor of the ambient air drawn in through the sampling hole as the odor detection signal. The fire detection system according to claim 2.

4. The control unit can set the judgment threshold in multiple stages based on the detection result from the odor detection unit, and dynamically changes the detection sensitivity of the fire detector according to the detection result. The fire detection system according to claim 2 or 3.

Images