Method for preventing misinformation of mine carbon monoxide sensor based on information fusion

Abstract

The invention discloses a method for preventing false alarm of a mine carbon monoxide sensor based on information fusion. A hardware system comprises a power supply, a LoRa network system, an accurate positioning system, a gas sensing system, a data processing system and an alarm system. The false alarm method comprises the following steps: step 1, establishing a relational expression VCO-0 = F (VNO) between CO concentration and NO concentration in the tail gas of the diesel engine; 2, establishing a spatial attenuation equation VCO-x = VCO-max * G (x) of CO concentration; 3, the real-time concentration # imgabs0 # of NO in tail gas of the diesel engine is detected, and the real-time concentration # imgabs1 # of CO in the tail gas is calculated according to VCO-0 = F (VNO); 4, the concentration # imgabs2 # of CO escaping from an exhaust pipe of the diesel engine to a CO sensing assembly is calculated according to a spatial attenuation equation of the CO concentration; 5, the actual mine CO concentration # imgabs3 # is calculated; 6, whether the CO concentration is about to exceed the limit or not is judged, and if yes, an alarm system gives an alarm; otherwise, repeating the steps 3 to 6. According to the invention, the false alarm phenomenon of the mine CO sensor caused by the influence of diesel engine tail gas can be prevented.

Description

Technical Field

[0001] The invention relates to a method for preventing false alarms of carbon monoxide in a mine, in particular to a method for preventing false alarms of carbon monoxide sensors in a mine based on information fusion, and belongs to the technical field of coal mine safety prediction. Background Art

[0002] CO is a colorless, odorless, flammable and explosive toxic gas, and is the enemy of coal mine safety production. When CO enters the human blood circulation system, it will combine with hemoglobin to cause poisoning, and in severe cases, it will cause death. When drilling underground, the mixed gas in the borehole may burn and a large amount of CO may gush out. The mass concentration of CO is one of the important indicators of spontaneous combustion of underground coal mine fires, so the ability to accurately and timely detect whether the underground CO gas exceeds the limit is one of the important factors directly related to the safe production of coal mines.

[0003] In recent years, underground CO monitoring technology has made significant progress, but it is very easy to cause false alarms when diesel engines are close to CO monitoring components, which to some extent hinders mine production. Summary of the invention

[0004] The purpose of the present invention is to provide a method for preventing false alarms of mine carbon monoxide sensors based on information fusion, which can prevent false alarms caused by the mine CO sensor being affected by diesel engine exhaust.

[0005] In order to achieve the above-mentioned object, the present invention provides a method for preventing false alarms of carbon monoxide sensors in mines based on information fusion, and the hardware system for implementing the method includes a power supply, a LoRa network system, a precise positioning system, a gas sensing system, a data processing system and an alarm system;

[0006] The LoRa network system consists of a ground network system and an underground network system, and the ground network system and the underground network system are connected by optical fiber;

[0007] The gas sensing system includes a NOx sensing component and a COx sensing component; the NOx sensing component is installed in the exhaust pipe of the diesel engine, and a one-way valve is installed at the outer port of the exhaust pipe of the diesel engine, so that the gas can only flow out in one direction; the NOx sensing component includes a NOx sensor, a development board and a LoRa module, and the COx sensing component is arranged in a specific area of ​​the mine, and includes a COx sensor and a development board;

[0008] The precise positioning system is installed at the outlet of the diesel engine exhaust pipe and sends the location of the diesel engine in real time. It includes a locator and a LoRa module;

[0009] The data processing system includes a single chip microcomputer and a LoRa module, and is connected to the CO sensing component;

[0010] The alarm system includes a buzzer and red and green LED lights, and is connected to the data processing system;

[0011] The false alarm method of the present invention comprises the following steps:

[0012] Step 1: Establish the relationship between CO concentration and NO concentration in diesel engine exhaust V CO-0 =F(V NO );

[0013] Step 2: Establish the spatial attenuation equation V of CO concentration CO-x =V CO-max *G(x);

[0014] Step 3: Detect the real-time concentration of NO in diesel exhaust And according to V CO-0 =F(V NO ) Calculate the real-time concentration of CO in the exhaust gas

[0015] Step 4: Calculate the CO concentration that escapes from the diesel engine exhaust pipe to the CO sensing component based on the spatial attenuation equation of CO concentration

[0016] Step 5: Calculate actual mine CO concentration

[0017] Step 6: Determine whether the CO concentration is about to exceed the limit. If so, the alarm system will sound an alarm. Otherwise, repeat steps 3 to 6.

[0018] Step 1: Establishing the relationship between CO concentration and NO concentration in diesel engine exhaust gas includes the following steps:

[0019] Step 1.1, placing the diesel engine in different power working modes, and detecting the NO concentration and CO concentration in the exhaust gas multiple times in the same working mode;

[0020] Step 1.2: Use the Isolation Forest algorithm to remove abnormal detection values;

[0021] Step 1.3, calculate the average values ​​of NO concentration and CO concentration under different working modes;

[0022] Step 1.4: Use polynomial fitting to establish the relationship between CO concentration and NO concentration in diesel engine exhaust V CO-0 =F(V NO ).

[0023] Step 2: Establishing the spatial attenuation equation of CO concentration includes the following steps:

[0024] Step 2.1, establish a simulated mine tunnel. In the simulated mine, except that the CO component is replaced by N2, the other gas components and environmental parameters are consistent with the actual mine;

[0025] Step 2.2: Set the diesel engine to the highest power working mode, take the end of the diesel engine exhaust pipe as the starting point, measure the CO concentration along the roadway, and record the CO concentration value V at each point. CO-x and the distance x between this point and the diesel engine exhaust pipe until the CO concentration is 0;

[0026] Step 2.3: Perform polynomial fitting on the data at each point to obtain the spatial attenuation equation of CO concentration:

[0027] V CO-x =V CO-max *G(x)

[0028] Where: V CO-x is the CO concentration at a distance x from the diesel engine exhaust pipe;

[0029] V co-max It is the CO concentration discharged from the tail pipe when the diesel engine is working at the highest power;

[0030] x is the distance between the measuring point and the end of the diesel engine exhaust pipe;

[0031] G(x) is the fitting polynomial.

[0032] Step 4: Calculating the CO concentration escaping from the diesel engine exhaust pipe to the CO sensing component includes the following steps:

[0033] Step 4.1: The real-time concentration of CO in the diesel exhaust obtained in step 3 Substitute into the CO concentration spatial attenuation equation, that is, the CO concentration spatial attenuation equation is transformed into V CO-0 =V CO-max *G(x), solve for the corresponding horizontal coordinate x0;

[0034] Step 4.2: The precise positioning system transmits the position of the diesel engine to the data processing system through the LoRa module, and the data processing system calculates the distance x between the diesel engine and the CO sensing component. t ;

[0035] Step 4.3: Calculate the CO concentration that escapes from the diesel engine exhaust pipe to the CO sensing component

[0036] Step 5: Calculate the actual mine CO concentration The following steps are involved:

[0037] Step 5.1: Calculate the CO concentration detected by the CO component

[0038] Step 5.2: CO concentration emitted from diesel engines To make a difference,

[0039] Step 6 determines whether the CO concentration is about to exceed the limit by using a RBF-BP hybrid neural network that mixes the RBF neural network and the BP neural network, including the following steps:

[0040] Step 6.1: The actual mine CO concentration V1 N , the first-order derivative of actual mine CO concentration with respect to time And the first derivative of NO concentration in diesel exhaust with respect to time Load it into the RBF subnet through the input layer nodes, and set the output of each neuron in the first hidden layer to be but

[0041]

[0042] Where: is the connection weight between the input layer and the hidden layer of the RBF neural network;

[0043] σ is the width of the radial basis function;

[0044] j is the number of nodes in the RBF hidden layer;

[0045] Step 6.2: The output of each neuron in the hidden layer of the RBF subnet is transferred to each neuron in the hidden layer of the BP subnet through the corresponding connection weights.

[0046]

[0047] Where: and Respectively represent the input and output of each neuron in the hidden layer of the BP subnet;

[0048] is the connection weight between the hidden layer of the RBF neural network and the hidden layer of the BP neural network;

[0049] k is the number of hidden layer nodes of BP neural network;

[0050] Step 6.3: Assume that the input of each neuron in the output layer is

[0051]

[0052] Where: is the connection weight between the hidden layer and the output layer of the BP neural network;

[0053] m is the number of nodes in the output layer;

[0054] Step 6.4: Calculate the probability of CO concentration exceeding the limit

[0055]

[0056] like The alarm system buzzer will sound intermittently and the red LED light will flash intermittently. Then the green LED light is always on, is the CO concentration exceeding limit probability threshold.

[0057] Compared with the prior art, the present invention establishes a relationship between the CO concentration and the NO concentration in the diesel engine exhaust gas V CO-0 =F(V NO ) and the spatial attenuation equation of CO concentration V CO-x =V CO-max *G(x); then detect the real-time concentration of NO in diesel engine exhaust According to the relationship between CO concentration and NO concentration V CO-0 =F(V NO ) to calculate the real-time concentration of CO in the exhaust gas The CO concentration escaping from the diesel engine exhaust pipe to the CO sensing component is calculated based on the spatial attenuation equation of CO concentration. Finally, the detection value of the CO sensing component is subtracted from the CO concentration that escapes from the diesel engine exhaust pipe to the CO sensing component. The actual mine CO concentration can be obtained Based on this, it is judged whether the CO concentration is about to exceed the limit; the present invention can monitor in real time whether the CO concentration is about to exceed the limit, thereby preventing the false alarm phenomenon of the CO sensor in the mine caused by diesel engine exhaust. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a schematic diagram of the process of the present invention;

[0059] Figure 2 is the spatial attenuation curve of CO concentration;

[0060] Figure 3 This is a network structure diagram based on the RBF-BP hybrid network. DETAILED DESCRIPTION

[0061] The present invention will be further described below in conjunction with the accompanying drawings.

[0062] like Figures 1 to 3 As shown, a method for preventing false alarms of carbon monoxide sensors in mines based on information fusion, the hardware system for implementing the method includes a power supply, a LoRa network system, a precise positioning system, a gas sensing system, a data processing system and an alarm system;

[0063] The LoRa network system consists of a ground network system and an underground network system, and the ground network system and the underground network system are connected by optical fiber;

[0064] The gas sensing system includes a NOx sensing component and a COx sensing component; the NOx sensing component is installed in the exhaust pipe of the diesel engine, and a one-way valve is installed at the outer port of the exhaust pipe of the diesel engine, so that the gas can only flow out in one direction; the NOx sensing component includes a NOx sensor, a development board and a LoRa module, and the COx sensing component is arranged in a specific area of ​​the mine, and includes a COx sensor and a development board;

[0065] The precise positioning system is installed at the outlet of the diesel engine exhaust pipe and sends the location of the diesel engine in real time. It includes a locator and a LoRa module;

[0066] The data processing system includes a single chip microcomputer and a LoRa module, and is connected to the CO sensing component;

[0067] The alarm system includes a buzzer and red and green LED lights, and is connected to the data processing system;

[0068] The false alarm method of the present invention comprises the following steps:

[0069] Step 1: Establish the relationship between CO concentration and NO concentration in diesel engine exhaust V CO-0 =F(V NO );

[0070] Specifically:

[0071] Step 1.1, placing the diesel engine in different power working modes, and detecting the NO concentration and CO concentration in the exhaust gas multiple times in the same working mode;

[0072] Step 1.2: Use the Isolation Forest algorithm to remove abnormal detection values;

[0073] Step 1.3, calculate the average values ​​of NO concentration and CO concentration under different working modes;

[0074] Step 1.4: Use polynomial fitting to establish the relationship between CO concentration and NO concentration in diesel engine exhaust V CO-0 =F(V NO ).

[0075] Step 2: Establish the spatial attenuation equation V of CO concentration CO-x =V CO-max *G(x);

[0076] Specifically:

[0077] Step 2.1, establish a simulated mine tunnel. In the simulated mine, except that the CO component is replaced by N2, the other gas components and environmental parameters are consistent with the actual mine;

[0078] Step 2.2: Set the diesel engine to the highest power working mode, take the end of the diesel engine exhaust pipe as the starting point, measure the CO concentration along the roadway, and record the CO concentration value V at each point. CO-x and the distance x between this point and the diesel engine exhaust pipe until the CO concentration is 0;

[0079] Step 2.3: Perform polynomial fitting on the data at each point to obtain the spatial attenuation equation of CO concentration:

[0080] V CO-x =V CO-max *G(x)

[0081] Where: V CO-x is the CO concentration at a distance x from the diesel engine exhaust pipe;

[0082] V co-max It is the CO concentration discharged from the tail pipe when the diesel engine is working at the highest power;

[0083] x is the distance between the measuring point and the end of the diesel engine exhaust pipe;

[0084] G(x) is the fitting polynomial.

[0085] Step 3: Detect the real-time concentration of NO in diesel exhaust And according to V CO-0 =F(V NO ) Calculate the real-time concentration of CO in the exhaust gas

[0086] Step 4: Calculate the CO concentration that escapes from the diesel engine exhaust pipe to the CO sensing component based on the spatial attenuation equation of CO concentration

[0087] Specifically:

[0088] Step 4.1: The real-time concentration of CO in the diesel exhaust obtained in step 3 Substitute into the CO concentration spatial attenuation equation, that is, the CO concentration spatial attenuation equation is transformed into V CO-0 =V CO-max *G(x), solve for the corresponding horizontal coordinate x0;

[0089] Step 4.2: The precise positioning system transmits the position of the diesel engine to the data processing system through the LoRa module, and the data processing system calculates the distance x between the diesel engine and the CO sensing component. t ;

[0090] Step 4.3: Calculate the CO concentration that escapes from the diesel engine exhaust pipe to the CO sensing component

[0091] Step 5: Calculate actual mine CO concentration

[0092] Specifically: Step 5 Calculate the actual mine CO concentration The following steps are involved:

[0093] Step 5.1: Calculate the CO concentration detected by the CO component

[0094] Step 5.2: CO concentration emitted from diesel engines To make a difference,

[0095] Step 6: Determine whether the CO concentration is about to exceed the limit. If so, the alarm system will sound an alarm. Otherwise, repeat steps 3 to 6.

[0096] Specifically:

[0097] Step 6 determines whether the CO concentration is about to exceed the limit by using a RBF-BP hybrid neural network that mixes the RBF neural network and the BP neural network, including the following steps:

[0098] Step 6.1: The actual mine CO concentration V1 N , the first derivative value of actual mine CO concentration with respect to time And the first derivative of NO concentration in diesel exhaust with respect to time Load it into the RBF subnet through the input layer nodes, and set the output of each neuron in the first hidden layer to be but

[0099]

[0100] Where: is the connection weight between the input layer and the hidden layer of the RBF neural network;

[0101] σ is the width of the radial basis function;

[0102] j is the number of nodes in the RBF hidden layer;

[0103] Step 6.2: The output of each neuron in the hidden layer of the RBF subnet is transferred to each neuron in the hidden layer of the BP subnet through the corresponding connection weights.

[0104]

[0105]

[0106] Where: and Respectively represent the input and output of each neuron in the hidden layer of the BP subnet;

[0107] is the connection weight between the hidden layer of the RBF neural network and the hidden layer of the BP neural network;

[0108] k is the number of hidden layer nodes of BP neural network;

[0109] Step 6.3: Assume that the input of each neuron in the output layer is

[0110]

[0111] Where: is the connection weight between the hidden layer and the output layer of the BP neural network;

[0112] m is the number of nodes in the output layer;

[0113] Step 6.4: Calculate the probability of CO concentration exceeding the limit

[0114]

[0115] like The alarm system buzzer will sound intermittently and the red LED light will flash intermittently. Then the green LED light is always on, is the CO concentration exceeding limit probability threshold.

[0116] The following is an example of the present invention for preventing false alarms of carbon monoxide in a mine:

[0117] Real-time monitoring value of NOx sensor at the end of diesel engine exhaust pipe As shown in Table 1 below:

[0118] Table 1 Monitoring values ​​of NO sensing components during a certain period of time

[0119]

[0120] According to the relationship between CO concentration and NO concentration in diesel engine exhaust V CO-0 =F(V NO ), the real-time concentration of CO in the diesel engine exhaust pipe can be obtained The data shown in Table 2:

[0121] Table 2 Real-time CO concentration in diesel engine exhaust pipe during a period of time

[0122]

[0123] The real-time concentration of CO in the exhaust pipe of the diesel engine is Substitute the value of into the CO concentration spatial attenuation equation V CO-x =V CO-max *G(x), the equation is transformed into V CO-0 =V CO-max *G(x), solve the corresponding horizontal coordinate x0 as shown in Table 3:

[0124] Table 3 The horizontal coordinate of the CO concentration spatial attenuation equation corresponding to the CO concentration in the diesel engine exhaust pipe during a certain period of time

[0125]

[0126]

[0127] The distance x between the diesel engine and the CO sensing component at a certain time period calculated using the precise positioning system t As shown in Table 4:

[0128] Table 4 The distance between the diesel engine and the CO sensing component during a period of time

[0129]

[0130] The CO concentration that escapes from the diesel engine exhaust pipe to the CO sensing component is As shown in Table 5:

[0131] Table 5 CO concentration escaping from the diesel engine exhaust pipe to the CO sensing component during a certain period of time

[0132]

[0133] The CO concentration detected by the CO sensor component during a certain period of time is As shown in Table 6:

[0134] Table 6 CO concentration detected by CO sensing components during a certain period of time

[0135]

[0136]

[0137] The actual CO concentration in the mine during a certain period of time is As shown in Table 7:

[0138] Table 7 Actual CO concentration in the mine itself during a certain period of time

[0139]

[0140] Finally, the CO concentration V1 during this period N=[8, 11, 11, 9, 10], the first-order derivative value of actual mine CO concentration with respect to time V2 N =[1, 3, 0, -2, 1], the first-order derivative value of NO concentration in diesel engine exhaust with respect to time V3 N =[13, -11, 23, -20, 28] is loaded into the RBF subnet through the input layer node, and the probability of CO concentration exceeding the limit in the mine is calculated by the RBF neural network and the BP neural network. The calculation results are shown in Table 8.

[0141] Table 8 Probability of CO concentration exceeding the limit in a mine during a period of time

[0142]

[0143] The CO concentration in the mine exceeds the limit probability threshold Table 8 shows the probability of CO concentration exceeding the limit All less than Therefore, the CO concentration did not exceed the limit during this period.

Claims

1. A method for preventing false alarms of carbon monoxide sensors in mines based on information fusion. The hardware system for implementing the method includes a power supply, a LoRa network system, a precise positioning system, a gas sensing system, a data processing system and an alarm system; The LoRa network system includes a ground network system and an underground network system, and the ground network system and the underground network system are connected via optical fiber; The gas sensing system includes a NOx sensing component and a COx sensing component; the NOx sensing component is installed in the exhaust pipe of the diesel engine, and a one-way valve is installed at the outer port of the exhaust pipe of the diesel engine, so that the gas can only flow out in one direction; the NOx sensing component includes a NOx sensor, a development board and a LoRa module, and the COx sensing component is arranged in a specific area of ​​the mine, and includes a COx sensor and a development board; The precise positioning system is installed at the outlet of the diesel engine exhaust pipe and can send the location of the diesel engine in real time. It includes a locator and a LoRa module; The data processing system includes a single chip microcomputer and a LoRa module, and is connected to the CO sensing component; The alarm system includes a buzzer and red and green LED lights, and is connected to the data processing system; The power supplies are for the LoRa network system, precise positioning system, gas sensing system, data processing system and alarm system; It is characterized in that The false positive method includes the following steps: Step 1: Establish the relationship between CO concentration and NO concentration in diesel engine exhaust V CO-0 =F(V NO ); Step 2: Establish the spatial attenuation equation V of CO concentration CO-x =V CO-max *G(x); Step 3: Detect the real-time concentration of NO in diesel exhaust And according to V CO-0 =F(V NO ) Calculate the real-time concentration of CO in the exhaust gas Step 4: Calculate the CO concentration that escapes from the diesel engine exhaust pipe to the CO sensing component based on the spatial attenuation equation of CO concentration Step 5: Calculate actual mine CO concentration Step 6: Determine whether the CO concentration is about to exceed the limit. If so, the alarm system will sound an alarm. Otherwise, repeat steps 3 to 6.

2. A method for preventing false alarms of carbon monoxide sensors in mines based on information fusion according to claim 1, characterized in that: Step 1: Establishing the relationship between CO concentration and NO concentration in diesel engine exhaust includes the following steps: Step 1.1, placing the diesel engine in different power working modes, and detecting the NO concentration and CO concentration in the exhaust gas multiple times in the same working mode; Step 1.2: Use the Isolation Forest algorithm to remove abnormal detection values; Step 1.3, calculate the average values ​​of NO concentration and CO concentration under different working modes; Step 1.4: Use polynomial fitting to establish the relationship between CO concentration and NO concentration in diesel engine exhaust V CO-0 =F(V NO ).

3. A method for preventing false alarms of carbon monoxide sensors in mines based on information fusion according to claim 1, characterized in that: Step 2: Establishing the spatial attenuation equation of CO concentration includes the following steps: Step 2.1, establish a simulated mine tunnel. In the simulated mine, except that the CO component is replaced by N2, the other gas components and environmental parameters are consistent with the actual mine; Step 2.2: Set the diesel engine to the highest power working mode, take the end of the diesel engine exhaust pipe as the starting point, measure the CO concentration along the roadway, and record the CO concentration value V at each point. CO-x and the distance x between this point and the diesel engine exhaust pipe until the CO concentration is 0; Step 2.3: Perform polynomial fitting on the data at each point to obtain the spatial attenuation equation of CO concentration: V CO-x =V CO-max *G(x) Where: V CO-x is the CO concentration at a distance x from the diesel engine exhaust pipe; V co-max It is the CO concentration discharged from the tail pipe when the diesel engine is working at the highest power; x is the distance between the measuring point and the end of the diesel engine exhaust pipe; G(x) is the fitting polynomial.

4. A method for preventing false alarms of carbon monoxide sensors in mines based on information fusion according to claim 1, characterized in that: Step 4: Calculating the CO concentration escaping from the diesel engine exhaust pipe to the CO sensing component includes the following steps: Step 4.1: The real-time concentration of CO in the diesel exhaust obtained in step 3 Substitute into the CO concentration spatial attenuation equation, that is, the CO concentration spatial attenuation equation is transformed into V CO-0 =V CO-max *G(x), solve for the corresponding horizontal coordinate x0; Step 4.2: The precise positioning system transmits the position of the diesel engine to the data processing system through the LoRa module, and the data processing system calculates the distance x between the diesel engine and the CO sensing component. t ; Step 4.3: Calculate the CO concentration that escapes from the diesel engine exhaust pipe to the CO sensing component 5. The method for preventing false alarms of carbon monoxide sensors in mines based on information fusion according to claim 3 is characterized in that: Step 5: Calculate the actual mine CO concentration The following steps are involved: Step 5.1: Calculate the CO concentration detected by the CO component Step 5.2: CO concentration emitted from diesel engines To make a difference, 6. A method for preventing false alarms of carbon monoxide sensors in mines based on information fusion according to claim 3, characterized in that: Step 6 determines whether the CO concentration is about to exceed the limit by using a RBF-BP hybrid neural network that mixes the RBF neural network and the BP neural network, including the following steps: Step 6.1: The actual mine CO concentration V1 N , the first-order derivative of actual mine CO concentration with respect to time And the first-order derivative value V3 of NO concentration in diesel exhaust with respect to time N Load it into the RBF subnet through the input layer nodes, and set the output of each neuron in the first hidden layer to be but Where: is the connection weight between the input layer and the hidden layer of the RBF neural network; σ is the width of the radial basis function; j is the number of nodes in the RBF hidden layer; Step 6.2: The output of each neuron in the hidden layer of the RBF subnet is transferred to each neuron in the hidden layer of the BP subnet through the corresponding connection weights. Where: and Respectively represent the input and output of each neuron in the hidden layer of the BP subnet; is the connection weight between the hidden layer of the RBF neural network and the hidden layer of the BP neural network; k is the number of hidden layer nodes of BP neural network; Step 6.3: Assume that the input of each neuron in the output layer is Where: is the connection weight between the hidden layer and the output layer of the BP neural network; m is the number of nodes in the output layer; Step 6.4: Calculate the probability of CO concentration exceeding the limit like The alarm system buzzer will sound intermittently and the red LED light will flash intermittently. Then the green LED light is always on, is the CO concentration exceeding limit probability threshold.

Images