CN110716012B - An Intelligent Monitoring System for Oil and Gas Concentration Based on Fieldbus Network - Google Patents

Abstract

The invention discloses an oil and gas concentration intelligent monitoring system based on a field bus network. The system consists of a gas station oil tank area environment parameter acquisition platform based on a CAN field bus network and a multi-point oil and gas concentration leakage classifier in the gas station oil tank area environment. The system is composed of the system, which realizes the intelligent detection of the concentration of oil and gas leakage in the oil tank area of the gas station and the classification of the concentration level of the oil and gas leakage; the invention not only effectively solves the problem of the traditional oil and gas concentration detection device in the oil tank area of the gas station, the design of the oil and gas concentration detection equipment Unreasonable, outdated equipment, imperfect detection system and other reasons lead to many problems in the detection of oil and gas concentration in the gas station tank area environment, and effectively solve the existing gas station tank area environment monitoring system. The concentration of the gas station can be detected and classified, so as to greatly improve the accuracy and robustness of the detection of the environmental oil and gas concentration in the oil tank area of the gas station.

Description

An Intelligent Monitoring System for Oil and Gas Concentration Based on Fieldbus Network

technical field

The invention relates to the technical field of automatic monitoring of oil and gas environment in gas stations, in particular to an intelligent monitoring system for oil and gas concentration based on a field bus network.

Background technique

The environment of the oil tank area of the gas station mainly stores oil storage containers or pipelines. If there is a leak, flammable liquid vapor will be generated. When the vapor pressure is high, there will be a danger of combustion and explosion, and the flammable liquid has flowability. If it encounters a fire source at room temperature, it will catch fire and burn. If the storage container leaks, the flammable vapor will continue to evaporate in the process of flowing. Once it comes into contact with the fire source, even the smallest spark will cause combustion. Therefore, strengthening fire safety management, real-time detection of the concentration of oil and gas leakage caused by fire, eliminating hidden dangers of safety accidents, and avoiding fire accidents are the most important tasks of environmental fire safety management in gas station oil tank farms. When the storage tank leaks and encounters sparks, it will cause a fire. If the fire fighting measures are not timely, it will cause a series of chain reactions, resulting in greater losses, continuous explosions, and huge shock waves that can destroy equipment and workshops in an instant. , is extremely destructive. The oil tank area of the gas station is the area where the most oil products are involved in the gas station. The oil products are all flammable liquids, and the probability of fire and explosion accidents is high, and once an accident occurs, the consequences are quite serious. A fire accident in the oil tank area of a gasoline filling station not only causes harm to people and surrounding equipment and facilities, but also may cause an explosion accident when the vapor concentration increases, such as when the explosion concentration limit of gasoline is reached. In this way, the economic loss will be more serious and the social impact will be stronger. Therefore, there is still a lot of work to be done in the safety management and emergency management of the gas station tank farm environment. From the analysis of accident types in the oil tank farm environment of gas stations, leakage and fire and explosion accidents are the key points of environmental safety precautions for gas station oil tank farms. The patent of the present invention invents an intelligent monitoring system for oil and gas concentration based on a field bus network. The system is composed to realize the detection, prediction and classification of the environmental oil and gas concentration in the oil tank area of the gas station, and the detection, prediction and classification of the oil and gas leakage concentration in the oil tank area of the gas station have relatively high accuracy and robustness.

SUMMARY OF THE INVENTION

The invention provides an intelligent monitoring system for oil and gas concentration based on a field bus network, which not only effectively solves the problems of unreasonable design of oil and gas concentration detection equipment, backward equipment, imperfect detection system, etc. There are still many problems in the detection of the environmental oil and gas concentration in the oil tank area of the gas station. According to the characteristics of the non-linearity of oil and gas concentration changes, the large lag and the complex change of oil and gas concentration in the large environmental area of the gas station oil tank area, the existing oil and gas concentration is effectively solved. The station oil tank area environment monitoring system detects and classifies the environmental concentration of the gas station oil tank area, thereby greatly improving the accuracy and robustness of the gas station oil tank area environmental oil and gas concentration detection.

The present invention is achieved through the following technical solutions:

An intelligent monitoring system for oil and gas concentration based on a fieldbus network is composed of a CAN fieldbus network-based acquisition platform for gas station oil tank area environmental parameters and a multi-point oil and gas concentration leakage classification subsystem in the gas station oil tank area environment. Intelligent detection, prediction and classification of the concentration of oil and gas leakage in the environment of the station tank area.

The further technical improvement scheme of the present invention is:

The environmental parameter acquisition platform of gas station oil tank farm based on CAN field bus network is composed of detection nodes and on-site monitoring terminals, which are constructed into a gas station oil tank farm parameter acquisition platform through CAN field bus. The detection nodes are respectively composed of a sensor group module, a single-chip microcomputer and a communication module. The sensor group module is responsible for detecting the temperature, oil and gas concentration, wind speed, smoke and other environmental parameters of the gas station tank area in the gas station tank area. The single-chip microcomputer controls the sampling interval and communicates with The module is sent to the on-site monitoring terminal; the on-site monitoring terminal is composed of an industrial control computer and RS232/CAN communication module, which realizes the management of the parameters of the gas station oil tank area detected by the detection node and the fusion of multi-point oil and gas concentrations in the gas station oil tank area. Prediction and classification. Figure 1 shows the environmental parameter acquisition platform of the gas station tank farm based on the CAN field bus.

The further technical improvement scheme of the present invention is:

The multi-point oil and gas concentration leakage classification subsystem in the gas station tank area environment consists of multiple detection point oil and gas concentration sensors, multiple time series triangular fuzzy number neural networks, multi-point oil and gas concentration fusion model in the gas station tank area environment, and triangular fuzzy number prediction. The module and the wavelet neural network oil and gas leakage concentration classifier are composed of 5 parts. The oil and gas concentration sensors at multiple detection points sense the concentration of oil and gas leakage at the detected points, and the output of the oil and gas concentration sensor at each detection point is used as the corresponding triangular fuzzy number for each time series. The input of the neural network, the output of multiple time-series triangular fuzzy number neural networks are used as the input of the multi-point oil and gas concentration fusion model of the gas station tank area environment, and the output of the multi-point oil and gas concentration fusion model of the gas station tank area environment is used as the triangular fuzzy number The input of the prediction module and the output of the triangular fuzzy number prediction module are used as the input of the wavelet neural network oil and gas leakage concentration classifier. The wavelet neural network oil and gas leakage concentration classifier divides the detected oil and gas leakage concentration into different levels. The multi-point oil and gas concentration leakage classification subsystem in the tank area environment realizes the detection, fuzzy quantification, multi-point fusion, prediction and classification process of the oil and gas leakage concentration level of the gas station. The subsystem is shown in Figure 2.

The further technical improvement scheme of the present invention is:

Multiple time series triangular fuzzy number neural networks are composed of each time series triangular fuzzy number neural network corresponding to each detection point. The input of the radial basis neural network, the triangular fuzzy value of the radial basis neural network and the oil and gas concentration of the detected point are composed as the output of the radial basis neural network, and the triangular fuzzy value output by the radial basis neural network respectively represents the The lower limit value, the maximum possible value and the upper limit value of the oil and gas concentration; the time series triangular fuzzy number neural network converts a conventional time series value of the oil and gas concentration of the detected point into the detected oil and gas concentration according to the dynamic change characteristics of the oil and gas concentration of the detected point. The triangular fuzzy value of oil and gas concentration is represented, and this transformation is more in line with the dynamic change law of oil and gas concentration at the detected point.

The further technical improvement scheme of the present invention is:

The multi-point oil and gas concentration fusion model of the gas station tank area environment is composed of the triangular fuzzy number array of oil and gas concentration time series, the relative closeness of the predicted value of the oil and gas concentration triangular fuzzy number and the positive and negative ideal values, and the calculated triangular fuzzy number fusion value of the oil and gas concentration. It consists of 3 parts. The predicted values of the triangular fuzzy numbers of the oil and gas concentration of multiple parameter detection units for a period of time constitute a triangular fuzzy number array of oil and gas concentration time series, and the positive and negative ideal values of the triangular fuzzy number array of oil and gas concentration time series are determined, and each detection unit is calculated separately. The distance between the predicted value of the oil and gas concentration time series triangular fuzzy number and the positive and negative ideal values of the oil and gas concentration time series triangular fuzzy number array, divided by the distance of the negative ideal value of the time series triangular fuzzy number predicted value of each detection unit by each detection unit The quotient obtained by summing the distance of the negative ideal value of the predicted value of the time series triangular fuzzy number of the unit and the distance of the positive ideal value of the predicted value of the time series triangular fuzzy number of each detection unit is the time series triangular fuzzy number of each detection unit The relative closeness of the predicted value, the quotient obtained by dividing the relative closeness of the time series triangular fuzzy number predicted value of each detection unit by the sum of the relative closeness of the time series triangular fuzzy number predicted value of all detection units is each detection unit The fusion weight of the predicted value of the time series triangular fuzzy number, the sum of the product of the predicted value of the time series triangular fuzzy number of each detection unit and the fusion weight of the predicted value of the time series triangular fuzzy number of the detection unit to obtain the time of multiple detection points The fused value of sequential triangular fuzzy predictions.

The further technical improvement scheme of the present invention is:

The triangular fuzzy number prediction module is composed of 3 NARX neural network prediction models and 3 Elman neural network prediction models of phase space reconstruction technology. The lower limit, maximum possible value and upper limit of the fuzzy number are the inputs of NARX neural network prediction model 1, NARX neural network prediction model 2 and NARX neural network prediction model 3, respectively. The difference between the lower limit, maximum possible value and upper limit of the triangular fuzzy number of the detected environmental oil and gas concentration output by the model and the output of NARX neural network prediction model 1, NARX neural network prediction model 2 and NARX neural network prediction model 3 respectively It is the input of Elman neural network prediction model 1 of phase space reconstruction technology, Elman neural network prediction model 2 of phase space reconstruction technology and Elman neural network prediction model 3 of phase space reconstruction technology, NARX neural network prediction model 1, NARX The outputs of neural network prediction model 2 and NARX neural network prediction model 3 are respectively the same as Elman neural network prediction model 1 of phase space reconstruction technology, Elman neural network prediction model 2 of phase space reconstruction technology and Elman neural network prediction model of phase space reconstruction technology. The sum of the outputs of the network prediction model 3 is used as the predicted value of the triangular fuzzy number of the oil and gas concentration in the detected environment, and the predicted value of the triangular fuzzy number is used as the output of the triangular fuzzy number prediction module.

The further technical improvement scheme of the present invention is:

The wavelet neural network oil and gas leakage concentration classifier is established to evaluate the environmental oil and gas leakage of the detected gas station oil tank area according to the engineering practice of the detected oil and gas leakage concentration in the oil tank area of the gas station and the national "Gas Station Seepage Leak Pollution Control Standard". The language variables of five concentration levels of concentration and the corresponding relationship table of five different triangular fuzzy numbers. There are 5 leakage levels: relatively high, normal oil and gas leakage concentration and very low oil and gas leakage concentration; the wavelet neural network oil and gas leakage concentration classifier classifies the detected oil and gas concentration leakage concentration levels in the gas station tank area, and the wavelet neural network oil and gas leakage concentration class The output of the concentration classifier is a triangular fuzzy value representing the concentration level of oil and gas leakage. By calculating the output of the wavelet neural network oil and gas leakage concentration classifier and the five kinds of triangles representing the five kinds of oil and gas leakage concentration levels of the detected gas station tank environment The similarity of the fuzzy numbers, where the oil and gas leakage level corresponding to the triangular fuzzy number with the largest similarity is the current concentration level of the detected oil and gas leakage in the oil tank area of the gas station.

Compared with the prior art, the present invention has the following obvious advantages:

1. The present invention aims at the uncertainty and randomness of the problems such as sensor accuracy error, interference and abnormal measured temperature value in the process of measuring the environmental parameters of the oil tank area of the gas station. The measured parameter values are converted into triangular fuzzy numbers through the time series triangular fuzzy number neural network model, which effectively deal with the fuzziness, dynamics and uncertainty of the parameters measured by the parameter sensor in the gas station tank farm environment. The objectivity and reliability of the sensor value detection parameters of the environmental parameter detection in the gas station oil tank area are analyzed.

2. The multi-point oil and gas concentration fusion model of the gas station oil tank environment environment of the present invention realizes dynamic fusion of the oil and gas concentration triangular fuzzy prediction values of multiple detection points, and determines the oil and gas concentration of the time series triangular fuzzy number prediction values of the plurality of detection points. Time series triangular fuzzy number array, determine the positive and negative ideal values of the oil and gas concentration time series triangular fuzzy number array, respectively calculate the oil and gas concentration time series triangular fuzzy number predicted value of each detection unit and the positive and negative values of the oil and gas concentration time series triangular fuzzy number array The distance of the ideal value, the relative closeness of each detection unit to the positive and negative ideal values, and the fusion weight improve the accuracy of the predicted value of the oil and gas concentration triangle fuzzy number at the detected point.

3. The input of the NARX neural network prediction model used in the present invention includes the input and output history feedback of the lower limit a, the possible value b and the upper limit c of the triangular fuzzy number of the detected point for a period of time, and this part of the feedback input can be It is considered that the state history information of the detected triangular fuzzy numbers including a period of time participates in the prediction of the detected triangular fuzzy numbers. For a suitable feedback time length, the prediction has obtained a good effect. The NARX neural network prediction mode of this patent is used. An effective triangular fuzzy number detection method for environmental parameters of oil tank farms in gas stations is provided.

4. The NARX neural network prediction model used in the present invention is a nonlinear and non-stationary time series that can effectively determine the lower limit a, possible value b and upper limit c of the triangular fuzzy number of the parameters of the detected point of the gas station. The dynamic neural network model for prediction can improve the prediction accuracy of the time series of the triangular fuzzy numbers of the detected points of the gas station when the non-stationarity of the time series is reduced. Compared with the traditional forecasting model method, this method has the advantages of good effect in dealing with non-stationary time series, fast calculation speed and high accuracy. Through the actual comparison of the experimental data of the non-stationary oil and gas leakage concentration of the oil tanker, this patent verifies the feasibility of the NARX neural network prediction model to predict the triangular fuzzy number time series of the detected points of the gas station. At the same time, the experimental results also prove that the NARX neural network forecasting model performs better than the traditional model in non-stationary time series forecasting.

5. The present invention uses the NARX neural network to establish the triangular fuzzy parameter prediction model of the detected point of the gas station. Since the dynamic recursive network of the model is established by introducing the delay module and the output feedback, it introduces the input and output vector delay feedback into the network training. , forming a new input vector, which has good nonlinear mapping ability. The input of the network model includes not only the original input data, but also the output data after training. The generalization ability of the network is improved, making it suitable for nonlinear oil tankers. Compared with the traditional static neural network, the time series prediction of oil and gas leakage concentration has better prediction accuracy and adaptive ability.

6. The Elman neural network prediction model of the phase space reconstruction technology adopted in the present invention realizes the prediction of the residual error of the parametric triangular fuzzy number of the detected point, and the predicted value is used as the compensation value of the triangular fuzzy number of the detected point, which improves the The accuracy of the triangular fuzzy number detection of the detected points, the Elman neural network prediction model is generally divided into 4 layers: input layer, middle layer (hidden layer), successor layer and output layer, its input layer, hidden layer and output layer The connection of the layers is similar to the feedforward network, the units of the input layer only play the role of signal transmission, and the units of the output layer play the role of linear weighting. The transfer function of the hidden layer unit can be a linear or nonlinear function. The successor layer is also called the context layer or the state layer. It is used to memorize the output value of the hidden layer unit at the previous moment, which can be considered as a one-time delay operator. . The characteristic of Elman neural network prediction model is that the output of the hidden layer is self-connected to the input of the hidden layer through the delay and storage of the receiving layer. This self-connection method makes it sensitive to the data of historical state, and the internal feedback network Joining increases the ability of the network itself to deal with dynamic information, so as to achieve the purpose of dynamic modeling. Elman neural network prediction model regression neuron network is characterized in that the output of the hidden layer is self-connected to the input of the hidden layer through the delay of the structural unit, and the self-connection method makes it sensitive to the data of the historical state, and the internal feedback network The addition of the network increases the ability of the network itself to process dynamic information, which is beneficial to the modeling of the dynamic process; the model uses the feedback connection of the dynamic neurons in the association layer to fuse the information of the future prediction network and the past prediction network, so that the network can understand the time series. The memory of feature information is strengthened, thereby improving the prediction accuracy of triangular fuzzy numbers of detected points.

7. The present invention combines phase space reconstruction and Elman neural network to construct a phase space reconstructed Elman neural network prediction model to compensate and predict the parameters of oil and gas concentration, temperature and smoke that affect fire risk. The phase configuration space can estimate the evolution information of the one-dimensional time series of the predicted parameters and expand the one-dimensional time series to a multi-dimensional sequence containing polymorphic information, so that the results of the predicted parameters are more consistent with the actual values; The output vector of the spatial reconstruction is used as the input value of the Elman neural network to avoid the randomness when selecting the input parameters of the Elman neural network, and the Elman neural network prediction model of the phase space reconstruction is used to improve the accuracy and reliability of predicting the parameters affecting the gas station fire. , which is of great value in accurately predicting the fire hazards of gas stations.

8. The scientificity and reliability of the wavelet neural network oil and gas leakage concentration classifier of the present invention, the wavelet neural network oil and gas leakage concentration classifier of the present invention is based on the engineering practice of the detected oil and gas leakage concentration in the oil tank area of the gas station and the national regulations on "Gas Oil and Gas Leakage". "Standards for Pollution Control of Station Seepage and Leakage", establishes a correspondence table between five concentration levels of linguistic variables and five different triangular fuzzy numbers to evaluate the concentration of oil and gas leakage in the oil tank area of the gas station to be detected. The environmental oil and gas leakage concentration is divided into 5 leakage levels: high oil and gas leakage concentration, high oil and gas leakage concentration, relatively high oil and gas leakage concentration, normal oil and gas leakage concentration and very low oil and gas leakage concentration; wavelet neural network oil and gas leakage concentration classifier The detected oil and gas leakage concentration levels in the oil tank area of the gas station are classified. The output of the wavelet neural network oil and gas leakage concentration classifier is a triangular fuzzy value representing the oil and gas leakage concentration level. The output of the wavelet neural network oil and gas leakage concentration classifier is calculated separately. Similarity with 5 kinds of triangular fuzzy numbers representing the 5 kinds of oil and gas leakage concentration levels of the detected gas station oil tank environment, among which the oil and gas leakage level corresponding to the triangular fuzzy number with the largest similarity is the detected gas station oil tank The current concentration level of oil and gas leakage in the regional environment can realize the dynamic performance and scientific classification of the fire hazard classification of gas stations.

Description of drawings

Fig. 1 The acquisition platform of environmental parameters of gas station oil tank farm based on CAN field bus network;

Figure 2. The multi-point oil and gas concentration leakage classification subsystem in the oil tank area of the gas station;

Figure 3 Detection node function diagram;

Figure 4 On-site monitoring terminal software function diagram;

Figure 5. Time series triangular fuzzy number neural network model;

Figure 6. The floor plan of the environmental parameter collection platform in the oil tank area of the gas station.

Detailed ways

In conjunction with accompanying drawing 1-6, the technical scheme of the present invention is further described:

1. Design of the overall function of the system

The invention realizes the detection of the environmental factor parameters of the oil tank area of the gas station and the detection, prediction and classification of the multi-point oil and gas leakage concentration in the oil tank area of the gas station. The system is collected by the environmental parameters of the gas station oil tank area based on the CAN field bus network The multi-point oil and gas concentration leakage classification subsystem in the platform and the gas station tank area environment consists of two parts. It is composed of multiple detection nodes 1 and on-site monitoring terminal 2 of the environmental parameter acquisition platform of the gas station oil tank area based on the CAN fieldbus network. The monitoring and control network is constructed through the CAN fieldbus to realize the connection between the detection node 1 and the on-site monitoring terminal 2. The detection node 1 sends the detected environmental parameters of the gas station tank farm to the site monitoring terminal 2 and performs preliminary processing of the sensor data; the site monitoring terminal 2 transmits the control information to the detection node 1. The entire system structure is shown in Figure 1.

2. Design of detection nodes

The detection node 1 based on the CAN fieldbus is used as the environmental parameter perception terminal of the oil tank area of the gas station. The detection node 1 includes sensors for collecting the ambient temperature, oil and gas concentration, wind speed and smoke parameters of the gas station oil tank area, the corresponding signal conditioning circuit, and STC89C52RC microprocessor; the software of the detection node mainly realizes the field bus communication and the gas station oil tank area environment. Parameter acquisition and preprocessing. The software adopts C language programming with high compatibility, which greatly improves the work efficiency of software design and development, and enhances the reliability, readability and portability of the program code. The structure of the detection node is shown in Figure 3.

3. On-site monitoring software

The on-site monitoring terminal 2 is an industrial control computer. The on-site monitoring terminal 2 mainly realizes the collection of environmental parameters of the gas station tank area and the fusion, prediction and classification of multi-point oil and gas concentrations, and realizes and detects the information of node 1 and control node 2. The main functions of the on-site monitoring terminal 2 are communication parameter setting, data analysis and data management, and multi-point temperature fusion of the gas station tank farm environment. The management software selects Microsoft Visual++6.0 as the development tool, and calls the Mscomm communication control of the system to design the communication program. The software functions of the on-site monitoring terminal are shown in Figure 4. The multi-point oil and gas concentration leakage classification subsystem in the gas station tank area environment consists of multiple detection point oil and gas concentration sensors, multiple time series triangular fuzzy number neural networks, multi-point oil and gas concentration fusion model in the gas station tank area environment, and triangular fuzzy number prediction. The module and the wavelet neural network oil and gas leakage concentration classifier are composed of 5 parts. The oil and gas concentration sensors at multiple detection points sense the concentration of oil and gas leakage at the detected points, and the output of the oil and gas concentration sensor at each detection point is used as the corresponding triangular fuzzy number for each time series. The input of the neural network, the output of multiple time-series triangular fuzzy number neural networks are used as the input of the multi-point oil and gas concentration fusion model of the gas station tank area environment, and the output of the multi-point oil and gas concentration fusion model of the gas station tank area environment is used as the triangular fuzzy number The input of the prediction module and the output of the triangular fuzzy number prediction module are used as the input of the wavelet neural network oil and gas leakage concentration classifier. The wavelet neural network oil and gas leakage concentration classifier divides the detected oil and gas leakage concentration into different levels. The multi-point oil and gas concentration leakage classification subsystem in the tank area environment realizes the detection, fuzzy quantification, multi-point fusion, prediction and classification process of the oil and gas leakage concentration level of the gas station. The subsystem is shown in Figure 2. The algorithm of the multi-point oil and gas concentration leakage classification subsystem in the tank area of the gas station is as follows:

1. Time series triangular fuzzy number neural network model

The time series of oil and gas concentration values at the detected points of the gas station are x(t), x(t-1),...,x(t-d+1), x(td). According to the oil and gas concentration of the detected points of the gas station A regular time series value of the parameter is used as the input of the radial basis neural network, and the output of the radial basis neural network is the triangular fuzzy value of the oil and gas concentration parameter at the detected point of the gas station at time t+1, and the triangular fuzzy number of S is expressed as [a ,b,c] is equal to [s ₁ , s ₂ , s ₃ ], a represents the lower limit of the oil and gas concentration at the detected point, b represents the maximum possible value of the oil and gas concentration at the detected point, c represents the upper limit of the oil and gas concentration at the detected point, The value of the triangular fuzzy value at time t+1 of the detected parameter depends on the value state value of the conventional time series at the first d moments of the detected parameter, where d is the time window, according to the parameter S and the oil and gas concentration value of the detected point at the first d moments The time-series values have the characteristic of functional dependence. Through the time-series triangular fuzzy number neural network of the parameters of the oil and gas concentration values at the detected points, a time-series conventional sequence value of the parameters of the oil and gas concentration values of the detected points is established to predict the oil and gas concentration values of the detected points. The relationship between the triangular fuzzy values of the oil and gas concentration value parameters at the detected point at time t+1 of the parameter, and the time series triangular fuzzy number neural network model 1 structure diagram of the oil and gas concentration value parameter of the detected point is shown in Figure 5. The radial basis vector of the neural network is H=[h ₁ , h ₂ ,...; h _p ] ^T , where h _p is the basis function. The commonly used radial basis function in radial basis neural network is the Gaussian function, and its expression is:

In the formula, X is the time series output of the sensor of the detected parameter, C is the coordinate vector of the center point of the Gaussian base function of the hidden layer neuron, δj is the width of the Gaussian base function of the _jth neuron in the hidden layer; the output connection of the network The weight vector is w _ij , and the output expression of the time series triangular fuzzy number neural network model is:

The key point of the time series triangular fuzzy number neural network model 1 of the oil and gas concentration parameters of the detected points is to use the oil and gas concentration value data of the detected points at d times of the oil and gas concentration value parameters of the detected points in the past period of time and the data at the time t+1. The triangular fuzzy data of the oil and gas concentration value parameters of the detected point is used to fit the mapping relationship f, and then the triangular fuzzy value S of the fitting function of the oil and gas concentration value of the detected point is obtained through the forward propagation of the radial basis neural network. The mathematical model of the time series triangular fuzzy number neural network of the oil and gas concentration value parameters at the detected points can be expressed as:

S=f(x(t),x(t-1),...,x(t-d+1),x(t-d)) (3)

The design method of the time series triangular fuzzy number neural network model 2 corresponding to the temperature sensor at the detection point and the time series triangular fuzzy number neural network model 3 corresponding to the detection point smoke sensor is similar to the time series triangular fuzzy number neural network model 1.

2. Multi-point oil and gas concentration fusion model of gas station tank area environment

The multi-point oil and gas concentration fusion model of the gas station tank area environment consists of three parts: the triangular fuzzy number array of oil and gas concentration time series, the calculation of the relative closeness between the predicted value and the ideal value of the oil and gas concentration triangular fuzzy number, and the calculation of the oil and gas concentration triangular fuzzy number fusion value. The triangular fuzzy number predicted value of oil and gas concentration of multiple parameter detection units for a period of time constitutes a triangular fuzzy number array of oil and gas concentration time series, and the time series triangular fuzzy number predicted value of each detection unit and the oil and gas concentration time series triangular fuzzy number array are calculated respectively. The distance between the positive ideal value of , and the distance between the time series triangular fuzzy number predicted value of each detection unit and the negative ideal value of the oil and gas concentration time series triangular fuzzy number array, respectively calculate the time series triangular fuzzy number prediction value of each detection unit. The quotient obtained by dividing the distance of the negative ideal value by the distance of the negative ideal value of the predicted value of the time series triangular fuzzy number of each detection unit and the distance of the positive ideal value of the predicted value of the time series triangular fuzzy number of each detection unit is The relative closeness of the time series triangular fuzzy values of each detection unit, the quotient obtained by dividing the pair closeness of the time series triangular fuzzy values of each detection unit by the sum of the pair closeness of the time series triangular fuzzy values of all detection units is The fusion weight of the time-series triangular fuzzy numbers of each detection unit, the sum of the products of the time-series triangular fuzzy numbers of each detection unit and the fusion weight of the time-series triangular fuzzy numbers of the detection unit obtains the time-series triangular fuzzy numbers of multiple detection points Fuzzy fusion value; the algorithm of the multi-point oil and gas concentration fusion model of the gas station tank area environment is as follows:

(1) Constructing the triangular fuzzy number array of oil and gas concentration time series

The triangular fuzzy number predicted values of the oil and gas concentration of multiple parameter detection units for a period of time constitute a triangular fuzzy number array of oil and gas concentration time series, and there are n detection points and m time The triangular fuzzy number predicted values of the parameter detection units form n rows. and the oil and gas concentration time series triangular fuzzy number array in column m, set the fuzzy triangular number prediction value of oil and gas concentration of different parameter detection units at different times as X _ij (t),X _ij (t+1),…,X _ij (d) , then the triangular fuzzy number array of oil and gas concentration time series is:

(2) Calculate the relative closeness between the predicted value of the triangular fuzzy number of the oil and gas concentration and the ideal value

The average value of the triangular fuzzy number prediction values of the oil and gas concentration of all detection units at the same time constitutes the positive ideal value of the triangular fuzzy number array of the oil and gas concentration time series, and the positive ideal value of the time series triangular fuzzy number is:

The predicted value of the triangle fuzzy number with the largest distance between the predicted value of the triangular fuzzy number and the positive ideal value of the oil and gas concentration of the detection unit at the same time constitutes the negative ideal value of the triangular fuzzy number array of the time series of oil and gas concentration, and the negative ideal value of the triangular fuzzy number of the time series is:

The distance between the predicted value of the time series triangular fuzzy number of each detection unit and the positive ideal value of the time series triangular fuzzy number array of oil and gas concentration is:

The distance between the time series triangular fuzzy number prediction value of each detection unit and the negative ideal value of the oil and gas concentration time series triangular fuzzy number array is:

The distance from the negative ideal value of the time-series triangular fuzzy number prediction value for each detection unit divided by the distance between the negative ideal value of the time-series triangular fuzzy number prediction value for each detection unit and the time-series triangular fuzzy number prediction for each detection unit The quotient obtained by the sum of the distances of the positive ideal values of the values is the relative closeness of the time series triangular fuzzy values of each detection unit:

(3) Calculate the triangular fuzzy number fusion value of oil and gas concentration

It can be known from the calculation of formula (9) that the greater the relative closeness between the time series triangular fuzzy value of each detection unit and the positive and negative ideal values of the oil and gas concentration time series triangular fuzzy number array, the greater the relative closeness of the time series triangular fuzzy value of the detection unit. The closer it is to the positive ideal value, otherwise the time series triangular fuzzy value of the detection point is relatively far from the positive ideal value. According to this principle, the pair closeness of the time series triangular fuzzy value of each detection unit is divided by all The quotient obtained by the sum of the closeness of the time series triangular fuzzy numbers of the detection units is the fusion weight of the time series triangular fuzzy numbers of each detection unit:

According to the sum of the product of the time series triangular fuzzy value of each detection unit and the fusion weight of the time series triangular fuzzy number of the detection unit, the time series triangular fuzzy fusion value of multiple detection points is obtained:

3. Triangular fuzzy number prediction module

The triangular fuzzy number prediction module includes 3 NARX neural network prediction models and 3 Elman neural network prediction models of phase space reconstruction technology, 3 NARX neural network prediction models 1, NARX neural network prediction models 2 and NARX neural network prediction models 3 The time series triangular fuzzy number neural network model 1 outputs the lower limit value of the detected point parameter a, the maximum possible value b of the detected point parameter and the upper limit value c of the detected point parameter of the output S triangular fuzzy number respectively. Elman neural network prediction model of space reconstruction technology 1, Elman neural network prediction model of phase space reconstruction technology 2 and Elman neural network prediction model of phase space reconstruction technology 3 The lower limit of the triangular fuzzy number S of the detected point respectively The value a and the residual output of the NARX neural network prediction model 1, the maximum possible value of the triangular fuzzy number S of the detected point and the detected point b and the residual error output by the NARX neural network prediction model 2, and the difference between the detected point and the detected point. The upper limit value c of the triangular fuzzy number S and the residual error of the output of the NARX neural network prediction model 3 are predicted; the outputs of the three NARX neural network prediction model 1, NARX neural network prediction model 2 and NARX neural network prediction model 3 are respectively and 3 The outputs of the Elman neural network prediction model 1 of the phase space reconstruction technology, the Elman neural network prediction model 2 of the phase space reconstruction technology, and the Elman neural network prediction model 3 of the phase space reconstruction technology are added together to obtain the sum of the The triangular fuzzy number S of the detection point is the predicted value of the lower limit value of oil and gas concentration a, the triangular fuzzy number S of the detected point of the detected point is the predicted value of the maximum possible value of oil and gas concentration b, and the predicted value of the upper limit value of oil and gas concentration c of the detected point, And constitute the predicted value of triangular fuzzy number, that is, s' is [a', b', c']. 3 Elman neural network prediction models of phase space reconstruction technology 1, Elman neural network prediction model of phase space reconstruction technology 2 and Elman neural network prediction model of phase space reconstruction technology 3 pairs of 3 NARX neural network prediction models 1, NARX neural network prediction model 2 and NARX neural network prediction model 3 respectively predict time series triangular fuzzy numbers. Neural network model 1 outputs S triangular fuzzy numbers a, b and c for further residual prediction compensation, which improves the accuracy of prediction a, b and c sex.

⑴, NARX neural network prediction model design

The three NARX neural network prediction models of the patent of the present invention respectively output the lower limit value a of the detected point parameter of the S triangular fuzzy number, the maximum possible value b of the detected point parameter, and the The upper limit value c of the detection point parameter is predicted. The NARX neural network (Nonlinear Auto-Regression with External input neural network) is a dynamic feedforward neural network, and the NARX neural network is a nonlinear auto-regression network with predicted input parameters. It has the dynamic characteristics of a multi-step delay, and connects several layers of a closed network with input parameters through feedback. The NARX regression neural network is the most widely used dynamic neural network in nonlinear dynamic systems, and its performance is generally better than Fully Recurrent Neural Network. The NARX neural network prediction model of this patent is composed of input layer, hidden layer, output layer and input and output delay. The current output of the neural network prediction model depends not only on the past output y(t-n), but also on the current input vector X(t) and the delay order of the input vector. The structure of the NARX neural network prediction model includes an input layer, an output layer, a hidden layer and a delay layer. The predicted input parameters are transmitted to the hidden layer through the delay layer, and the hidden layer processes the input signal and then transmits it to the output layer. The layer will linearly weight the output signal of the hidden layer to obtain the final predicted output signal of the neural network, and the delay layer will delay the signal fed back by the network and the signal output by the input layer, and then send it to the hidden layer. The NARX neural network prediction model has the characteristics of nonlinear mapping ability, good robustness and adaptability, and is suitable for predicting input parameters. x(t) represents the external input of the neural network, that is, the time series triangular fuzzy number neural network model 1 outputs the lower limit value a of the detected point parameter of the S triangular fuzzy number; m represents the delay order of the external input a; y(t) is the output of the neural network, that is, the predicted value of a; n is the output delay order; s is the number of neurons in the hidden layer; from this, the output of the jth hidden unit can be obtained as:

In the above formula, w _ji is the connection weight between the i-th input and the j-th hidden neuron, b _j is the bias value of the j-th hidden neuron, and the output y of the NARX neural network prediction model ( t+1) represents the predicted value of a:

y(t+1)=f[y(t),y(t-1),…,y(t-n),x(t),x(t-1),…,x(t-m+1) ;W] (13)

NARX neural network prediction model 2 and NARX neural network prediction model 3 respectively predict the maximum possible value b of the detected point parameter of the time series triangular fuzzy number neural network model 1 output S triangular fuzzy number and the upper limit value c of the detected point parameter , they are designed in a similar way to the NARX neural network prediction model 1.

⑵, Elman neural network prediction model design of phase space reconstruction technology

The Elman neural network prediction model of the three phase space reconstruction techniques composed of the phase space reconstruction technology and the Elman neural network 1, the Elman neural network prediction model of the phase space reconstruction technology 2 and the Elman neural network prediction of the phase space reconstruction technology Model 3 is respectively for the lower limit a of the triangular fuzzy number S of the detected point and the residual output of the NARX neural network prediction model 1, the maximum possible value b of the triangular fuzzy number S of the detected point and the output of the NARX neural network prediction model 2 The residual sum of the upper limit c of the triangular fuzzy number S of the detected point and the residual output of the NARX neural network prediction model 3 are predicted; the Elman neural network prediction model of the three phase space reconstruction techniques The Elman neural network prediction model 2 of the reconstruction technology and the Elman neural network prediction model 3 of the phase space reconstruction technology are respectively used for the lower limit a, the maximum possible value b and the upper limit c of the triangular fuzzy number S of the detected point. The residual error is predicted, and the Elman neural network prediction model of the three phase space reconstruction technology is trained with enough training samples to simulate the lower limit a, the maximum possible value b and the upper limit of the triangular fuzzy number S for predicting the detected point. Limit the variation of the residual error of c, so as to realize the compensation for the prediction of the triangular fuzzy number S of the detected point.

The specific steps of the prediction method of the Elman neural network prediction model reconstructed in three phase spaces are as follows:

Step 1: Collect the lower limit a, the maximum possible value b and the upper limit c of the output triangular fuzzy number S of the time series triangular fuzzy number neural network model 1 corresponding to the detected point sensor and three NARX neural networks respectively The three series differences of the output value of prediction model 1, the output value of NARX neural network prediction model 2, and the output value of NARX neural network prediction model 3 constitute three residual time series data corresponding to a, b and c, respectively.

Step 2: Determine the optimal delay constant τ and the embedding dimension m according to the routine.

Step 3: Build the Elman neural network prediction model. The Elman neural network prediction model is a forward neural network with local memory units and local feedback connections. The association layer receives feedback signals from the hidden layer. Each hidden layer node has a The corresponding association layer nodes are connected. The association layer uses the state of the hidden layer at the previous moment together with the network input at the current moment as the input of the hidden layer as the state feedback. The transfer function of the hidden layer is generally a sigmoid function, and the correlation layer and the output layer are linear functions. Let the number of input layer, output layer and hidden layer of Elman neural network prediction model be m, n and r respectively; w ₁ , w ₂ , w ₃ and w ₄ represent the structure layer unit to the hidden layer, the input layer to the hidden layer, respectively. layer, the connection weight matrix from the hidden layer to the output layer, and the structure layer to the output layer, the output value expressions of the hidden layer, the correlation layer and the output layer of the network are:

c _p (k) = x _p (k-1) (15)

3 Elman neural network prediction models of phase space reconstruction technology 1, Elman neural network prediction model of phase space reconstruction technology 2 and Elman neural network prediction model of phase space reconstruction technology 3 The input data is triangular fuzzy number neural network model The lower limit value a, the maximum possible value b and the upper limit value c of the output triangular fuzzy number S of 1 and the 3 NARX neural network prediction model 1 output value, NARX neural network prediction model 2 output value and NARX neural network prediction model are respectively 3 residual time series data corresponding to a, b and c composed of 3 series differences of output values, 3 Elman neural network prediction models of phase space reconstruction technology 1, Elman neural network prediction of phase space reconstruction technology The outputs of model 2 and Elman neural network prediction model 3 of the phase space reconstruction technique are the predicted compensation values of a, b and c, respectively. The input dimension of the Elman neural network prediction model is equal to the embedding dimension m, and the time difference between each input data is τ time points, that is, the residual time series data of a, b and c are used as the input of the Elman neural network prediction model; implicitly The layer is a single layer, and the number is determined by the 2m+1 method; the output layer contains a neuron, and its output is the residual prediction value of the time point to be predicted. If one of the residual time series data is X(t ₀ ), X(t ₁ ),…,X(t _i ),…,X(t _n ), first use MatlabR2012a programming to reconstruct the time series in phase space, The output of the phase space reconstruction is as follows:

Step 4: Elman neural network prediction model network training, select some data from the original data of the residual time series of a, b and c, and conduct network training until the training meets the requirements.

Step 5: Select test samples from the original data of the residual time series of a, b and c. If the requirements are met, you can go to step 6 for prediction. If the test error is large, return to step 4 for retraining, or return to The third step is to redesign the network structure.

Step 6: Select the prediction time point and apply the Elman neural network prediction model established earlier to make predictions.

The predicted value can be obtained by the following formula:

Y=f(X(t ₀ +mτ),X(t ₁ +mτ)…X(t _i +mτ)…X(t _n +τ)) (18)

4. The design of wavelet neural network oil and gas leakage concentration classifier

The output of the triangular fuzzy number prediction module is the triangular fuzzy number prediction value of the detected oil and gas concentration in the oil tank area of the gas station. The predicted value is used as the input of the wavelet neural network oil and gas leakage concentration classifier. The wavelet neural network oil and gas leakage concentration classifier The dynamic change status of the detected oil and gas concentration prediction value in the oil tank area of the gas station is divided into the detected oil and gas leakage concentration in the oil tank area of the gas station into the high oil and gas leakage concentration, the high oil and gas leakage concentration, the relatively high oil and gas leakage concentration, There are 5 types of leakage levels: normal oil and gas leakage concentration and very low oil and gas leakage concentration. The output of the wavelet neural network oil and gas leakage concentration classifier is a triangular fuzzy value representing the oil and gas leakage concentration level; According to the engineering practice and the national "Gas Station Seepage and Leakage Pollution Control Standard", the wavelet neural network oil and gas leakage concentration classifier classifies the detected oil and gas concentration leakage concentration level of the gas station tank area, and calculates the oil and gas leakage through the wavelet neural network respectively. The output of the concentration classifier is similar to the triangular fuzzy numbers representing the five concentration levels of oil and gas leakage in the gas station oil tank environment. The oil and gas leakage level corresponding to the triangular fuzzy number with the largest similarity is the detected gas station oil tank. The concentration level of oil and gas leakage in the area environment is the current concentration level of the detected oil and gas leakage in the oil tank area of the gas station. Wavelet Neural Network Oil and Gas Leak Concentration Classifier is a gas station oil and gas leakage concentration classifier constructed on the basis of wavelet neural network WNN (Wavelet Neural Networks). A feedforward network. In the wavelet neural network oil and gas leakage concentration classifier, the wavelet scaling, translation factor and connection weight are adaptively adjusted in the process of optimizing the error energy function. Set the clear value of the predicted value of the triangular fuzzy number of the oil and gas leakage concentration in a continuous time period as the input one-dimensional vector x _i (i=1,2,...,n) of the oil and gas leakage concentration classifier of the wavelet neural network, and the output is expressed as The triangular fuzzy value y _k (k=1,2,...,m) of the oil and gas leakage concentration level of the gas station, where m is equal to 3, represents the calculation of the output layer output of the oil and gas leakage concentration classifier of the wavelet neural network of the oil and gas leakage concentration level of the gas station The formula is:

为小波基函数，b_j为小波基函数的平移因子，a_j小波基函数的伸缩因子，ω_jk为隐含层j节点和输出层k节点间的连接权值。本专利中的小波神经网络油气泄露浓度分类器的权值和阈值的修正算法采用梯度修正法来更新网络权值和小波基函数参数，从而使小波神经网络油气泄露浓度分类器输出不断逼近期望输出。分别求取小波神经网络油气泄露浓度分类器输出与代表5种加油站油罐区环境油气泄露浓度等级语言变量对应的三角模糊数的相似度，相似度最大的三角模糊数对应的油气泄露浓度等级为被检测加油站油罐区环境油气泄露等级。In the formula, ω _ij is the connection weight between the input layer i node and the hidden layer j node,

is the wavelet basis function, b _j is the translation factor of the wavelet basis function, a _j is the scaling factor of the wavelet basis function, ω _jk is the connection weight between the hidden layer j node and the output layer k node. The correction algorithm for the weights and thresholds of the wavelet neural network oil and gas leakage concentration classifier in this patent uses the gradient correction method to update the network weights and wavelet basis function parameters, so that the output of the wavelet neural network oil and gas leakage concentration classifier is constantly approaching the expected output. . The similarity between the output of the wavelet neural network oil and gas leakage concentration classifier and the linguistic variables representing five kinds of oil and gas leakage concentration levels in the gas station tank area environment is obtained respectively, and the oil and gas leakage concentration level corresponding to the triangular fuzzy number with the largest similarity is obtained. It is the level of oil and gas leakage in the oil tank area of the gas station to be detected.

According to the engineering practice of the detected oil and gas leakage concentration in the oil tank area of the gas station and the national "Gas Station Seepage and Leakage Pollution Control Standard", the language variables of 5 concentration levels are established to evaluate the environmental oil and gas leakage concentration of the detected gas station oil tank area. The corresponding relationship table with triangular fuzzy numbers is shown in Table 1.

serial number Oil and gas leakage level triangular fuzzy number 1 The concentration of oil and gas leakage is very low (0.00, 0.00, 0.25) 2 The concentration of oil and gas leakage is normal (0.00, 0.25, 0.50) 3 The concentration of oil and gas leakage is relatively high (0.25, 0.50, 0.75) 4 High concentration of oil and gas leakage (0.50, 0.75, 1.00) 5 High concentration of oil and gas leakage (0.75, 1.00, 1.0)

5. An example of the design of the environmental parameter detection system in the oil tank farm of the gas station

According to the environment of the gas station oil tank area, the system has arranged the layout and installation diagram of the detection node 1 and the on-site monitoring terminal 2, in which the detection node 1 is evenly arranged in the environment of the oil tank area of the gas station to be tested, and the overall system layout is shown in the figure 6. Through this system, the collection of environmental parameters of the gas station oil tank area and the intelligent classification of the environmental oil and gas concentration of the gas station oil tank area are realized.

The technical means disclosed in the solution of the present invention are not limited to the technical means disclosed in the above embodiments, but also include technical solutions composed of any combination of the above technical features. It should be pointed out that for those skilled in the art, without departing from the principle of the present invention, several improvements and modifications can be made, and these improvements and modifications are also regarded as the protection scope of the present invention.

Claims (5)

1. a kind of oil and gas concentration intelligent monitoring system based on field bus network, it is characterized in that: described system is leaked by the gas station oil tank area environment parameter acquisition platform and the gas station oil tank area environment multi-point oil and gas concentration based on CAN field bus network It is composed of a classification subsystem, which realizes intelligent detection of the concentration of oil and gas leakage in the oil tank area of the gas station and classification of the concentration level of oil and gas leakage in the oil tank area of the gas station. Concentration sensor, multiple time series triangular fuzzy number neural network, multi-point oil and gas concentration fusion model of gas station tank area environment, triangular fuzzy number prediction module and wavelet neural network oil and gas leakage concentration classifier are composed of five parts. The concentration sensor senses the concentration of oil and gas leakage at the detected point. The output of the oil and gas concentration sensor at each detection point is used as the input of the corresponding time series triangular fuzzy number neural network, and the output of the time series triangular fuzzy number neural network is used as the gas station tank area. The input of the environmental multi-point oil and gas concentration fusion model, the output of the environmental multi-point oil and gas concentration fusion model of the gas station tank area is used as the input of the triangular fuzzy number prediction module, and the output of the triangular fuzzy number prediction module is used as the wavelet neural network oil and gas leakage concentration classifier. Input, the wavelet neural network oil and gas leakage concentration classifier divides the detected gas station oil and gas leakage concentration into different levels, and the multi-point oil and gas concentration leakage classification subsystem in the gas station tank area environment realizes the detection and fuzzy detection of the gas station oil and gas leakage concentration. Quantification, multi-point fusion, prediction and classification process of oil and gas leakage concentration levels; The time series triangular fuzzy number neural network is composed of each time series triangular fuzzy number neural network corresponding to each detection point. The input of the radial basis neural network, the triangular fuzzy value of the radial basis neural network and the oil and gas concentration of the detected point are composed as the output of the radial basis neural network, and the triangular fuzzy value output by the radial basis neural network respectively represents the Lower limit value, possible value and upper limit value of oil and gas concentration; time series triangular fuzzy number neural network converts a conventional time series value of oil and gas concentration at the detected point into detected oil and gas according to the dynamic change characteristics of oil and gas concentration at the detected point The density is represented by the triangular blur value. 2. a kind of oil and gas concentration intelligent monitoring system based on fieldbus network according to claim 1, is characterized in that: described gas station oil tank area environment multi-point oil and gas concentration fusion model is composed of oil and gas concentration time series triangular fuzzy number array, Calculate the relative closeness of the predicted value of the oil and gas concentration triangle fuzzy number to the positive ideal value and negative ideal value, and calculate the fusion value of the triangle fuzzy number of the oil and gas concentration. It consists of three parts, and the predicted value of the triangle fuzzy number of the oil and gas concentration of the multi-parameter detection unit for a period of time Construct the triangular fuzzy number array of oil and gas concentration time series, determine the positive ideal value and negative ideal value of the oil and gas concentration time series triangular fuzzy number array, and calculate the oil and gas concentration time series triangular fuzzy number prediction value of each detection unit and the oil and gas concentration time series triangular fuzzy number respectively. The distance between the positive ideal value and the negative ideal value of the fuzzy number array, the distance of the negative ideal value of the time series triangular fuzzy number prediction value of each detection unit divided by the negative ideal value of the time series triangular fuzzy number prediction value of each detection unit The quotient obtained from the sum of the distance between the positive ideal value of the predicted value of the time series triangular fuzzy number of each detection unit is the relative closeness of the predicted value of the time series triangular fuzzy number of each detection unit, and the time of each detection unit The quotient obtained by dividing the relative closeness of the predicted value of the sequential triangular fuzzy number by the sum of the relative closeness of the predicted value of the time series triangular fuzzy number of all detection units is the fusion weight of the predicted value of the time series triangular fuzzy number of each detection unit. The sum of the products of the time series triangular fuzzy number predicted value of each detection unit and the fusion weight of the time series triangular fuzzy number prediction value of this detection unit obtains the fusion value of the time series triangular fuzzy predicted values of multiple detection points; The average value of the triangular fuzzy number predicted value of the unit oil and gas concentration constitutes the positive ideal value of the triangular fuzzy number array of the oil and gas concentration time series. The predicted value constitutes the negative ideal value of the triangular fuzzy number array of the oil and gas concentration time series. 3. a kind of oil and gas concentration intelligent monitoring system based on field bus network according to claim 1, is characterized in that: described triangular fuzzy number prediction module is composed of three NARX neural network prediction models and three phase space reconstruction techniques. The Elman neural network prediction model is composed. The lower limit, possible value and upper limit of the triangular fuzzy number of the detected environmental oil and gas concentration output by the multi-point oil and gas concentration fusion model of the gas station tank area environment are respectively corresponding to the NARX neural network prediction model. Input, the difference between the lower limit, possible value and upper limit of the triangular fuzzy number of the detected environmental oil and gas concentration output by the multi-point oil and gas concentration fusion model of the gas station tank area environment and the output of the corresponding NARX neural network prediction model respectively corresponds to The input of the Elman neural network prediction model of the phase space reconstruction technology and the output of the NARX neural network prediction model are respectively added with the output of the corresponding Elman neural network prediction model of the phase space reconstruction technology, and the sum is used as the triangular blur of the detected environmental oil and gas concentration. The predicted value of the triangular fuzzy number is output as the triangular fuzzy number prediction module. 4. a kind of oil and gas concentration intelligent monitoring system based on field bus network according to claim 1, is characterized in that: described wavelet neural network oil and gas leakage concentration classifier, according to the detected gas station oil tank area environment oil and gas leakage concentration Engineering practice and the state's "Gas Station Seepage and Leakage Pollution Control Standards", establish a correspondence table between five concentration levels of linguistic variables and five different triangular fuzzy numbers to evaluate the detected gas station oil tank area environmental oil and gas leakage concentration, which will be used. The detected oil and gas leakage concentration in the oil tank area of the gas station is divided into five leakage levels: high oil and gas leakage concentration, high oil and gas leakage concentration, relatively high oil and gas leakage concentration, normal oil and gas leakage concentration, and very low oil and gas leakage concentration; wavelet neural network The oil and gas leakage concentration classifier classifies the detected oil and gas leakage concentration levels in the oil tank area of the gas station. The output of the oil and gas leakage concentration classifier of the wavelet neural network is a triangular fuzzy value representing the oil and gas leakage concentration level. By calculating the wavelet neural network respectively The similarity between the output of the oil and gas leakage concentration classifier and the five triangular fuzzy numbers representing the five oil and gas leakage concentration levels of the detected gas station tank area environment, where the oil and gas leakage level corresponding to the triangular fuzzy number with the largest similarity is the The current concentration level of the detected oil and gas leakage in the tank area of the gas station. 5. a kind of oil and gas concentration intelligent monitoring system based on field bus network according to claim 1, is characterized in that: described gas station oil tank farm environment parameter acquisition platform based on CAN field bus network is composed of a plurality of parameter detection nodes and It is composed of on-site monitoring terminals, and the information communication between them is realized through CAN field bus network; It can manage the environmental parameters of the gas station and the parameter management of the multi-point detection of the gas station tank area environment, integrate the oil and gas concentration of multiple detection points, predict and classify the oil and gas leakage concentration level.

Images