CN115905938A - Storage tank safety monitoring method and system based on Internet of things - Google Patents
Storage tank safety monitoring method and system based on Internet of things Download PDFInfo
- Publication number
- CN115905938A CN115905938A CN202211306368.1A CN202211306368A CN115905938A CN 115905938 A CN115905938 A CN 115905938A CN 202211306368 A CN202211306368 A CN 202211306368A CN 115905938 A CN115905938 A CN 115905938A
- Authority
- CN
- China
- Prior art keywords
- neural network
- storage tank
- output
- interval
- parameter
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000003860 storage Methods 0.000 title claims abstract description 158
- 238000012544 monitoring process Methods 0.000 title claims abstract description 83
- 238000000034 method Methods 0.000 title claims abstract description 47
- 238000001514 detection method Methods 0.000 claims abstract description 76
- 230000007175 bidirectional communication Effects 0.000 claims abstract description 3
- 238000004590 computer program Methods 0.000 claims abstract 2
- 238000013528 artificial neural network Methods 0.000 claims description 130
- 238000003062 neural network model Methods 0.000 claims description 116
- 230000001537 neural effect Effects 0.000 claims description 69
- 239000002245 particle Substances 0.000 claims description 47
- 238000005457 optimization Methods 0.000 claims description 40
- 230000004927 fusion Effects 0.000 claims description 37
- 238000005259 measurement Methods 0.000 claims description 34
- 238000000354 decomposition reaction Methods 0.000 claims description 32
- 230000003044 adaptive effect Effects 0.000 claims description 25
- 230000009467 reduction Effects 0.000 claims description 24
- 239000011159 matrix material Substances 0.000 claims description 12
- 230000005540 biological transmission Effects 0.000 claims description 3
- 230000003750 conditioning effect Effects 0.000 claims description 3
- 238000006073 displacement reaction Methods 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 abstract description 10
- 230000007547 defect Effects 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 57
- 230000008569 process Effects 0.000 description 21
- 239000013598 vector Substances 0.000 description 19
- 238000004422 calculation algorithm Methods 0.000 description 13
- 230000015654 memory Effects 0.000 description 13
- 238000013461 design Methods 0.000 description 11
- 238000013507 mapping Methods 0.000 description 10
- 238000004364 calculation method Methods 0.000 description 9
- 210000002569 neuron Anatomy 0.000 description 9
- 230000006854 communication Effects 0.000 description 8
- 238000012546 transfer Methods 0.000 description 8
- 238000004891 communication Methods 0.000 description 7
- 210000004027 cell Anatomy 0.000 description 6
- 238000012549 training Methods 0.000 description 6
- 230000002457 bidirectional effect Effects 0.000 description 5
- 238000011161 development Methods 0.000 description 5
- 230000003993 interaction Effects 0.000 description 5
- 230000001934 delay Effects 0.000 description 4
- 230000005284 excitation Effects 0.000 description 4
- 238000011478 gradient descent method Methods 0.000 description 4
- 230000003111 delayed effect Effects 0.000 description 3
- 230000007774 longterm Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 230000006837 decompression Effects 0.000 description 2
- 238000009795 derivation Methods 0.000 description 2
- 230000008034 disappearance Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000006855 networking Effects 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 230000001373 regressive effect Effects 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 201000004569 Blindness Diseases 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000003213 activating effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000002238 attenuated effect Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000008602 contraction Effects 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 239000010779 crude oil Substances 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013523 data management Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000003912 environmental pollution Methods 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000002045 lasting effect Effects 0.000 description 1
- 238000012886 linear function Methods 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005312 nonlinear dynamic Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Landscapes
- Testing Or Calibration Of Command Recording Devices (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a storage tank safety monitoring method and system based on the Internet of things, wherein the system comprises a detection terminal and a field monitoring terminal, the detection terminal is responsible for collecting parameter information of a detected storage tank, a storage tank safety monitoring subsystem is arranged in the field monitoring terminal, bidirectional communication among the detection terminal, the field monitoring terminal, a cloud platform and a mobile phone APP is realized through a gateway node, the parameter collection of the detected storage tank and the classification of the storage tank safety type are realized, and the detection terminal and the field monitoring terminal are loaded with computer program steps for realizing the storage tank safety monitoring method based on the Internet of things. According to the invention, aiming at the technical defects of instability, inaccuracy and untimely property of the monitoring parameters of the storage tank, the technology of the Internet of things, the intelligent control technology, the sensor technology and the big data technology are applied to the online monitoring system of the storage tank, so that the running state of the storage tank can be known in time, and meanwhile, the running data of the storage tank related to the time domain space can be detected and early warned, and the reliability, stability and accuracy of the large data monitoring system of the storage tank of the Internet of things can be improved.
Description
Technical Field
The invention relates to the technical field of storage tank safety monitoring, in particular to a storage tank safety monitoring method and system based on the Internet of things.
Background
The storage tank is a very important device in the petrochemical industry, and with the development of economy, the safety problem of the storage tank is more and more prominent. The storage tank is used as a container for storing crude oil or other chemical products, and the potential danger of the storage tank is mainly leakage, fire and explosion due to the falling-behind safety monitoring and management technology of the storage tank, so that once an accident occurs, huge economic loss and serious environmental pollution problems are caused. Therefore, corresponding technical means need to be continuously developed and updated to carry out safety monitoring on the storage tank, so that accidents are fundamentally avoided, and safety production is ensured. The method for monitoring and controlling parameters of the tank body such as pressure, temperature and the like in real time is the most direct and effective method for ensuring the safety of the storage tank, a safe and reliable safety monitoring system is established, and a scientific and comprehensive monitoring system is established, so that the method has great practical significance for effectively reducing accident potential in a petrochemical industry reservoir area, preventing accidents, preventing casualties and economic loss and ensuring the sustainable development of national economy and society. With the development of information science and technology and automatic production, the requirements of people on the stable operation of the storage tank are also improved, the problems are solved by using a traditional method, and the operation monitoring of the storage tank is difficult to completely meet the operation requirements of cost monitoring and directional operation and maintenance. In the actual storage tank operation, the storage tank often appears taking trouble operation, and storage tank running state is not up to standard, also leads to storage tank running cost to promote and the security performance reduces. How to realize the intelligent monitoring of storage tank just become present technological problem that awaits solution urgently, use internet of things, intelligent control technique and big data to combine together, pushed storage tank monitoring to intellectuality, informationization, pushed storage tank monitoring to intelligent service, promoted storage tank monitoring's perception ability for reliability and accuracy nature of monitoring storage tank operation improve greatly.
Disclosure of Invention
According to the invention, aiming at the technical defects of instability, inaccuracy and untimely property of the monitoring parameters of the storage tank, the technology of the Internet of things, the intelligent control technology, the sensor technology and the big data technology are applied to the online monitoring system of the storage tank, so that the running state of the storage tank can be known in time, and meanwhile, the running data of the storage tank related to the time domain space can be detected and early warned, and the reliability, stability and accuracy of the large data monitoring system of the storage tank of the Internet of things can be improved.
In order to solve the problems, the invention adopts the following technical scheme:
1. the storage tank safety monitoring method comprises the following steps:
1. constructing a parameter detection model, wherein the parameter detection model consists of a particle swarm optimization adaptive wavelet neural network-NARX neural network model, a TDL beat-to-beat delay line A, a wavelet decomposition model, a noise reduction self coding neural network-NARX neural network model, an LSTM neural network-NARX neural network model, a TDL beat-to-beat delay line B, a TDL beat-to-beat delay line C, a TDL beat-to-beat delay line D and a BAM neural network-AANN auto-associative neural network model with interval hesitation fuzzy numbers, the output of a parameter sensor is used as the input of the particle swarm optimization adaptive wavelet neural network-NARX neural network model, the output of the particle swarm optimization adaptive wavelet neural network-NARX neural network model is used as the input of the TDL beat-to-beat delay line A, and the output of the TDL beat-delay line A is used as the input of the wavelet decomposition model, the low-frequency trend part and a plurality of high-frequency fluctuation parts output by the wavelet decomposition model are respectively used as the input of an LSTM neural network-NARX neural network model and a noise-reducing self-coding neural network-NARX neural network model, the output of the LSTM neural network-NARX neural network model and the output of the noise-reducing self-coding neural network-NARX neural network model are respectively used as the input of a TDL beat-by-beat delay line B and a TDL beat-by-beat delay line C, the output of the TDL beat-by-beat delay line B, the output of the TDL beat-by-beat delay line C and the output of the TDL beat-by-beat delay line D are respectively used as the corresponding input of a BAM neural network-AANN self-linking neural network model with interval hesitation fuzzy number, 4 parameters output by the BAM neural network-AANN self-linking neural network model with interval hesitation fuzzy number are respectively a, B, C and D, a, B are used as the input of the TDL beat-by-delay line D, a and B form an interval number (a, b) The BAM neural network-AANN auto-associative neural network model output is used as the output of the parameter detection model, wherein the minimum value of the detected parameter is used as c and d to form interval numbers (c, d), the maximum value of the detected parameter is used as c, b) and the interval numbers (c, d) form [ (a, b), (c, d) ] and are used as interval hesitation fuzzy numbers of the detected parameter; the structure of the parameter detection model is shown in figure 1.
2. Constructing an interval hesitation fuzzy number fusion model, wherein the interval hesitation fuzzy number fusion model comprises the following steps: the method comprises the steps that interval hesitation fuzzy numbers output by a parameter detection model of a plurality of parameter measurement sensors in a period of time form a time sequence interval hesitation fuzzy number array, the distance and the correlation degree of the time sequence interval hesitation fuzzy numbers of the parameter measurement sensors are defined, a distance matrix and a correlation matrix are constructed, the ratio of the reciprocal of the distance average of the time sequence interval hesitation fuzzy numbers of each measurement sensor to the reciprocal of the distance average of the time sequence hesitation fuzzy numbers of all the parameter measurement sensors is the distance fusion weight of the time sequence interval hesitation fuzzy numbers of the parameter measurement sensors, the ratio of the correlation average of the time sequence interval hesitation fuzzy numbers of each measurement sensor to the correlation fusion weight of the time sequence interval hesitation fuzzy numbers of all the measurement sensors is the correlation fusion weight of the time sequence interval hesitation fuzzy numbers of the parameter measurement sensors, and the correlation fusion weight of the time sequence interval hesitation numbers of the parameter measurement sensors are used as the fusion weights of the parameter measurement sensors from small to large fuzzy number; the sum of the fusion values of the time series interval hesitation fuzzy numbers of all the parameter measuring sensors is obtained by adding the products of the time series interval hesitation fuzzy numbers of each parameter measuring sensor at the same moment and the interval number fusion weights of the time series interval hesitation fuzzy numbers of the parameter measuring sensors; the structure of the interval hesitation fuzzy number fusion model is shown in figure 2.
3. Constructing a storage tank safety monitoring subsystem, wherein the storage tank safety monitoring subsystem is composed of a parameter detection module, a parameter detection model, an interval hesitation fuzzy number fusion model, a TDL beat-by-beat delay line, a noise reduction self-coding neural network model, a particle swarm optimization self-adaptive wavelet neural network, a RBF neural network model and a BAM neural network-LSTM neural network storage tank safety classifier of interval numbers, the parameter detection module comprises a plurality of parameter detection models, a period of storage tank temperature values output by a plurality of temperature sensors serve as corresponding inputs of a plurality of parameter detection models of the corresponding parameter detection module, and outputs of the plurality of parameter detection models serve as outputs of the parameter detection module; the pressure values of the storage tank output by the pressure sensors for a period of time are used as corresponding inputs of a plurality of parameter detection models of the corresponding parameter detection module, and the outputs of the parameter detection models are used as the outputs of the parameter detection module; the output of 2 parameter detection modules is used as the input of a corresponding interval hesitation fuzzy number fusion model, the output of a flow sensor is used as the input of the parameter detection model, the output of the parameter detection model and the output of 2 interval hesitation fuzzy number fusion models are respectively used as the corresponding TDL to be input according to a beat delay line, the output of 3 TDL is respectively used as the corresponding input of a noise reduction self-coding neural network model, a particle swarm optimization self-adaptive wavelet neural network and an RBF neural network model according to a beat delay line, the output of the noise reduction self-coding neural network model, the output of the particle swarm optimization self-adaptive wavelet neural network and the output of the RBF neural network model are respectively used as interval numbers of a BAM neural network-LSTM neural network storage tank safety classifier, the 2 parameters output by the BAM neural network-LSTM neural network storage tank safety classifier of the interval numbers are respectively e and f, the interval numbers (e, f) of the e and the f are respectively used as the interval numbers of the detected safety storage tank, and the interval numbers output by the BAM neural network-LSTM neural network storage tank safety classifier of the interval numbers correspond to the storage tank good condition, the good storage tank condition, the general storage tank condition, the poor storage tank condition, and the poor storage tank type; the structure of the storage tank safety monitoring subsystem is shown in figure 2.
4. The outputs of the plurality of temperature sensors, the plurality of pressure sensors and the flow sensor are used as the inputs of the storage tank safety monitoring subsystem, and the storage tank safety monitoring subsystem outputs the safety type of the monitored storage tank.
5. The particle swarm optimization adaptive wavelet neural network-NARX neural network model, the noise reduction self-coding neural network-NARX neural network model, the LSTM neural network-NARX neural network model, the BAM neural network-AANN self-association neural network model and the BAM neural network-LSTM neural network storage tank safety classifier are characterized in that the particle swarm optimization adaptive wavelet neural network is connected with the NARX neural network model in series, the noise reduction self-coding neural network is connected with the NARX neural network model in series, the LSTM neural network is connected with the NARX neural network model in series, the BAM neural network is connected with the AANN self-association neural network model in series, and the BAM neural network is connected with the LSTM neural network storage tank safety classifier in series.
2. Based on thing networking storage tank safety monitoring system:
the storage tank big data monitoring system of the Internet of things comprises a gateway node, a detection terminal, an on-site monitoring end, a cloud platform and a mobile phone APP, wherein the detection terminal is responsible for collecting parameter information of a detected storage tank, bidirectional communication of the detection terminal, the on-site monitoring end, the cloud platform and the mobile phone APP is realized through the gateway node, a storage tank safety monitoring subsystem is arranged in the on-site monitoring end to realize parameter collection of the detected storage tank and classification of storage tank safety types, and the storage tank safety monitoring subsystem is realized in the right 1. The structure of the internet-of-things storage tank big data monitoring system is shown in figure 3.
The detection terminal comprises displacement, vibration, temperature, flow, pressure and tilt sensors for acquiring the safety parameters of the storage tank, a corresponding signal conditioning circuit, an STM32 single chip microcomputer and a CC2530 wireless transmission module.
Compared with the prior art, the invention has the following obvious advantages:
1. aiming at the uncertainty and randomness of the problems of precision error, interference, abnormal measurement and the like of the parameter measurement sensor in the parameter measurement process, the output value of the parameter measurement sensor is converted into a BAM neural network-NARX neural network model form of interval hesitation fuzzy numbers through a parameter detection model, the ambiguity, the dynamics and the uncertainty of the parameter measurement are effectively processed, and the objectivity and the reliability of the parameter detection of the parameter sensor are improved.
2. The BAM neural network of the BAM neural network-LSTM neural network model adopted by the invention is an associative memory neural network model, the output of the BAM neural network is used as the input of the LSTM neural network model, the BAM neural network can realize a bidirectional different associative model, the input sample data pair of the BAM neural network summarized in advance is stored through a bidirectional associative memory matrix, when new information is input into the BAM neural network, the BAM neural network parallelly remembers and associates out a corresponding output result to be used as the input of the LSTM neural network model, the bidirectional associative memory network of the BAM neural network is a two-layer nonlinear feedback neural network and has the functions of associative memory, distributed storage, self-learning and the like, and when an input signal is added into one layer of the BAM neural network, the other layer of the BAM neural network can obtain an output signal.
3. The invention improves the scientificity and reliability of the intelligent grade classifier of the storage tank safety, the output types of the interval number BAM neural network-LSTM neural network storage tank safety classifiers are respectively 5 storage tank safety types of good storage tank condition, general storage tank condition, poor storage tank condition and poor storage tank condition according to the safety parameters, expert experience and storage tank safety related national standards of the detected storage tank, the interval number BAM neural network-LSTM neural network storage tank safety classifiers realize more scientific and precise classification of the storage tank safety grade, and the scientificity and reliability of the safety classification of the detected storage tank are improved.
4. The invention adopts the particle swarm optimization adaptive wavelet neural network model, the transfer function of the hidden layer in the model uses the wavelet function, and the time-frequency characteristic of the signal can be more effectively input and extracted by adaptively adjusting the parameters of the wavelet function. The blindness of BP network in structural design, the linear distribution of network weight coefficients and the convexity of learning objective functions are avoided, the problems of local optimization and the like in the training process of the network are fundamentally avoided, the algorithm concept is simple, the convergence speed is high, the function learning capability is strong, and any nonlinear function can be approached with high precision.
5. The invention adopts the particle swarm optimization adaptive wavelet neural network, avoids the requirement of activating the function to be microminiaturized in the gradient descent method and the calculation of the derivation process of the function, and the iterative formula is simple when each particle is searched, thus the calculation speed is much faster than that of the gradient descent method. By adjusting the parameters in the iterative formula, local extreme values can be well jumped out, global optimization is carried out, and the training speed of the network is simply and effectively improved. The self-adaptive wavelet neural network model of the particle swarm optimization algorithm has smaller error, faster convergence speed and stronger generalization capability. The method has the advantages of simple algorithm, stable structure, high calculation convergence speed, strong global optimization capability, high identification precision and strong generalization capability.
6. The invention adopts an LSTM neural network model, which is a recurrent neural network with 4 interaction layers in a repetitive network. The LSTM neural network model can not only extract information from sequence data of input parameters like a standard recurrent neural network, but also retain information of long-term correlation of input parameters from previous remote steps, the input parameters have long-term spatial and temporal correlation, and the LSTM neural network model has enough long-term memory to process the spatiotemporal relationship between the previous input parameters, so that the accuracy and the robustness of processing the input parameters of the LSTM neural network model are improved.
Drawings
FIG. 1 is a parameter detection model of the present invention;
FIG. 2 is a tank safety monitoring subsystem of the present invention;
FIG. 3 is a tank safety monitoring system based on the Internet of things;
FIG. 4 is a diagram of a detection terminal according to the present invention;
FIG. 5 is a gateway node of the present invention;
fig. 6 shows the site monitoring software according to the present invention.
Detailed Description
For better explanation of the present invention and for better understanding, the technical solutions of the present invention will be described in detail below with reference to fig. 1 to 6. The following examples are illustrative of the present invention, and the present invention is not limited to the following examples.
1. The method for monitoring the safety of the storage tank comprises the following steps:
1. constructing a parameter detection model, wherein the parameter detection model consists of a particle swarm optimization adaptive wavelet neural network-NARX neural network model, a TDL beat-by-beat delay line A, a wavelet decomposition model, a noise reduction self-coding neural network-NARX neural network model, an LSTM neural network-NARX neural network model, a TDL beat-by-beat delay line B, a TDL beat-by-beat delay line C, a TDL beat-by-beat delay line D and a BAM neural network-AANN self-association neural network model of interval hesitation fuzzy numbers; the structure of the construction parameter detection module is shown in figure 1.
(1) Particle swarm optimization self-adaptive wavelet neural network-NARX neural network model design
The output of the parameter sensor is used as the input of a particle swarm optimization adaptive wavelet neural network-NARX neural network model, the output of the particle swarm optimization adaptive wavelet neural network-NARX neural network model is used as the input of a TDL beat delay line A, and the output of the particle swarm optimization adaptive wavelet neural network is used as the input of the NARX neural network model. The adaptive wavelet neural network is a feedforward network based on wavelet theory and adopting nonlinear wavelet base to replace common nonlinear Sigmoid function and combining artificial neural network, and the transfer function of hidden layer uses wavelet function and adopts self-adaptive wavelet functionThe parameters of the wavelet function are adaptively adjusted, and the time-frequency characteristics of the input signals of the parameter sensor can be more effectively extracted. The method takes a wavelet function as an excitation function of a neuron, and the expansion and contraction, translation factors and connection weights of the wavelet are adaptively adjusted in the optimization process of an error energy function. Let the input signal of the wavelet neural network be expressed as a one-dimensional vector x output by a parameter sensor i (i =1,2,. Ang., n), the output signal being denoted as y k (k =1, 2.. Said., m), the calculation formula of the wavelet neural network output layer value is:
in the formula omega ij Inputting the connection weight between the i node of the layer and the j node of the hidden layer,
as wavelet basis functions, b j Is a shift factor of the wavelet basis function, a j Scale factor, omega, of wavelet basis functions jk And the connection weight between the j node of the hidden layer and the k node of the output layer is shown. The weight and threshold correction algorithm of the self-adaptive wavelet neural network in the patent adopts a gradient correction method to update the network weight and the wavelet basis function parameters, so that the predicted output of the wavelet neural network continuously approaches the expected output. The particle swarm optimization adaptive wavelet neural network is adopted, the requirement that the activation function is microminiaturible in a gradient descent method and the process calculation of derivation of the function are avoided, and the iterative formula is simple when each particle is searched, so that the calculation speed is much higher than that of the gradient descent method. And local extreme values can be well jumped out by adjusting parameters in the iterative formula. A population of random particles is initialized and then an optimal solution is found through iteration. In each iteration, the particle updates itself by tracking two "extrema". The first is the optimal solution pbest found by the particle itself, this solution is called the individual extremum; the other is the best solution currently found for the whole population, this solution is called global extremum gbest. The optimization of the adaptive wavelet neural network by particle swarm is the firstFirstly, various parameters of the self-adaptive wavelet neural network are listed as position vectors X of particles, a mean square error energy function is set as an objective function for optimization, iteration is carried out through a basic formula of a particle swarm optimization algorithm, and an optimal solution is sought. The particle swarm optimization adaptive wavelet neural network training algorithm is as follows:
A. initializing a network structure, determining the number of neurons in a hidden layer of the network, and determining the dimension D of a target search space.
B. Determining the number m of the particles, initializing position vectors and velocity vectors of the particles, substituting the position vectors and the velocity vectors of the particles into an algorithm iterative formula for updating, performing optimization calculation by taking an error energy function as a target function, and recording the optimal position pbest searched by each particle so far and the optimal position gbest searched by the whole particle swarm so far.
C. Searching the whole particle swarm to the optimal position gbest so far, mapping the optimal position gbest to a network weight and a threshold value for the learning, and performing the chemical calculation by taking an error energy function as the fitness of the particle.
D. If the error energy function value is within the error range allowed by the actual problem, the iteration is finished; otherwise, the algorithm is switched back to continue the iteration.
The NARX neural network model (Nonlinear Auto-Regression with External input neural network) is a dynamic feedforward neural network, the NARX neural network is a Nonlinear autoregressive network with External input, the NARX neural network has a dynamic characteristic of multistep time delay and is connected with a plurality of layers of closed networks through feedback, the NARX neural network model is a regressive dynamic neural network which is most widely applied in a Nonlinear dynamic system, and the performance of the NARX neural network model is generally superior to that of a full regressive neural network. The NARX neural network model mainly comprises an input layer, a hidden layer, an output layer and an input and output delay layer, wherein before application, the delay order and the number of hidden layer neurons of the input and the output are generally determined in advance, and the output of the NARX neural network model at the time not only depends on the output y (t-n) of the NARX neural network model in the past, but also depends on the output vector X (t) of the current particle swarm optimization adaptive wavelet neural network of the NARX neural network model, the delay order of the output vector of the particle swarm optimization adaptive wavelet neural network and the like. The output of the particle swarm optimization adaptive wavelet neural network of the NARX neural network model is transmitted to the hidden layer through the time delay layer, the hidden layer processes signals output by the particle swarm optimization adaptive wavelet neural network and then transmits the processed signals to the output layer, the output layer linearly weights the output signals of the hidden layer to obtain final output signals of the neural network, and the time extension layer delays signals fed back by the NARX neural network model and the output signals of the particle swarm optimization adaptive wavelet neural network of the input layer and then transmits the delayed signals to the hidden layer. The NARX neural network model has the characteristics of nonlinear mapping capability, good robustness, adaptability and the like, and is suitable for predicting the output of the parameter sensor. x (t) represents the input of a NARX neural network model, namely the output of the particle swarm optimization adaptive wavelet neural network; m represents the delay order of the external input; y (t) is the output of the neural network, namely the predicted value of the output of the particle swarm optimization adaptive wavelet neural network in the next time interval; n is the output delay order; s is the number of hidden layer neurons; the output of the jth implicit element can thus be found as:
in the above formula, w ji As a connection weight between the ith input and the jth implicit neuron, b j Is the bias value for the jth implicit neuron.
(2) Wavelet decomposition model design
The output of the TDL beat-to-beat delay line A is used as the input of a wavelet decomposition model, a low-frequency trend part and a plurality of high-frequency fluctuation parts output by the wavelet decomposition model are respectively used as the input of an LSTM neural network-NARX neural network model and a denoising self-coding neural network-NARX neural network model, the output of the TDL beat-to-beat delay line A is divided into a low-frequency component and a plurality of high-frequency components by the wavelet decomposition model, the output predicted value of a time sequence parameter sensor output by the TDL beat-to-beat delay line A is decomposed by a wavelet decomposition method, and the decomposed information of each layer is subjected to autocorrelation and cross-correlation analysis; smoothing the signals in the wavelet decomposition process, respectively establishing a corresponding LSTM neural network-NARX neural network model and a corresponding denoising self-coding neural network-NARX neural network model according to the analyzed characteristics of the signals of each layer to predict the output of the parameter sensor again, wherein the wavelet multi-resolution decomposition process generally adopts a Mallat algorithm, and the decomposition relation of the algorithm is expressed as follows:
h in formula (3) 0 、h 1 A low-pass decomposition filter and a high-pass decomposition filter, respectively. m is p 、n p Respectively, the low frequency coefficient and the high frequency coefficient, and the algorithm reconstructs the relationship as follows:
in the formula (4), g 0 、g 1 A low-pass reconstruction filter and a high-pass reconstruction filter, respectively. A. The p 、D p Respectively low frequency components and high frequency components. The Mallat algorithm decomposes each layer of decomposed low-frequency signal part into high-frequency and low-frequency signals again, thus carrying out layer-by-layer decomposition. The result obtained after p-layer decomposition of the original signal X is:
X＝D 1 +D 2 +...D p +A p (5)
a in the formula (5) p For the part of the low-frequency signal after the decomposition of the p-th layer, D p The high-frequency part after the decomposition of the p-th layer. The wavelet decomposition can decompose the time series signals output by the TDL according to the beat delay line A into different resolution spaces, and the effect of the wavelet decomposition is that the predicted value of the output of the parameter sensor decomposed into each resolution space is divided into a plurality of high-frequency fluctuation parts and low-frequency trend parts.
(3) Design of noise reduction self-coding neural network-NARX neural network model
And the output of the wavelet decomposition model is used as the input of a TDL beat-by-beat delay line C. The noise reduction self-coding neural network is a dimension reduction method, and high-dimensional data is converted into low-dimensional data by training a multi-layer neural network with a small central layer. A noise-reducing self-coding neural network (DAE) is a typical three-layer neural network with an encoding process between a hidden layer and an input layer and a decoding process between an output layer and the hidden layer. The noise-reducing self-coding neural network obtains a coding representation (coder) through a coding operation on input data, and obtains reconstructed input data (decoder) through an output decoding operation on a hidden layer, wherein the data of the hidden layer is dimension reduction data. A reconstruction error function is then defined to measure the learning effect of the noise-reduced self-coding neural network. Constraints can be added based on the error function to generate various types of noise-reducing self-coding neural networks, and the encoder and decoder and the loss function are as follows:
an encoder: h = δ (Wx + b) (6)
A decoder:
loss function:
the training process of AE is similar to BP neural network, W and W 'are weight matrix, b and b' are offset, h is output value of hidden layer, x is input vector,
to output the vector, δ is the excitation function, typically using a Sigmoid function or a tanh function. The noise reduction self-coding neural network is divided into a coding process and a decoding process, wherein the coding process is from an input layer to a hidden layer, and the decoding process is from the hidden layer to an output layer. The objective of the noise reduction self-coding neural network is to make the input and output as close as possible by using an error function, and obtain self-coding by minimizing the error function through back propagationAnd (4) optimal weight and bias of the code network, and preparation is made for establishing a deep self-coding network model. According to the self-coding network coding and decoding principle, coded data and decoded data are obtained by using data containing noise measurement, an error function is constructed through the decoded data and original data, and an optimal network weight and an optimal bias are obtained through a back propagation minimized error function. The measurement data is corrupted by adding noise and the corrupted data is then input into the neural network as an input layer. The reconstruction result of the noise reduction self-coding neural network is similar to the original data of the measurement parameters, and by the method, the disturbance of the measurement parameters can be eliminated and a stable structure can be obtained. The original measurement parameters are input into the encoder to obtain the characteristic expression, and then mapped to the output layer through the decoder.
The current output of the NARX neural network model not only depends on the past NARX neural network model output y (t-n), but also depends on the current noise-reduced self-coding neural network output vector X (t) of the NARX neural network model, the delay order of the noise-reduced self-coding neural network output vector, and the like. The output of the noise reduction self-coding neural network of the NARX neural network model is transmitted to the hidden layer through the time delay layer, the hidden layer processes signals output by the noise reduction self-coding neural network and transmits the processed signals to the output layer, the output layer linearly weights the output signals of the hidden layer to obtain final neural network output signals, and the time delay layer delays signals fed back by the NARX neural network model and signals output by the noise reduction self-coding neural network of the input layer and transmits the delayed signals to the hidden layer. The NARX neural network model has the characteristics of nonlinear mapping capability, good robustness, self-adaptability and the like, and is suitable for further processing a plurality of high-frequency fluctuation parts of the output predicted values of the parameter sensors.
(4) LSTM neural network-NARX neural network model design
The low-frequency trend part output by the wavelet decomposition model is used as the input of an LSTM neural network-NARX neural network model, the LSTM neural network-NARX neural network model is used as a TDL beat-to-beat delay line B, the LSTM neural network output of the LSTM neural network-NARX neural network model is used as the input of the NARX neural network model, and a memory sheet is introduced into the LSTM neural networkMechanisms of element (Memory Cell) and hidden layer State (Cell State) to control the information transfer between hidden layers. The memory unit of an LSTM neural network is internally provided with 3 Gates (Gates) of calculation structures, namely an Input Gate (Input Gate), a forgetting Gate (Forget Gate) and an Output Gate (Output Gate). The input gate can control the adding information or filtering of the low-frequency trend value part of the time sequence output by the wavelet decomposition model; the forgetting gate can forget the low frequency trend value part information of the time series output by the wavelet decomposition model needing to be discarded and the low frequency trend value part information of the time series output by the wavelet decomposition model which is useful in the past; the output gate enables the memory unit to output only information related to the current time step. The 3 gate structures carry out operations such as matrix multiplication, nonlinear summation and the like in the memory unit, so that the memory still cannot be attenuated in continuous iteration. The long-short term memory unit (LSTM) structure unit is composed of a unit (Cell), an Input Gate (Input Gate), an Output Gate (Output Gate) and a forgetting Gate (Forget Gate). The LSTM neural network is a model capable of lasting for a long time and short-term memory, is suitable for predicting a low-frequency trend value part of a time sequence output by a wavelet decomposition model, effectively prevents gradient disappearance during RNN training, and is a special RNN. The LSTM neural network can learn the dependence information among the low-frequency trend value parts of the time sequence of the output predicted values of the long-term parameter sensor, and meanwhile, the problem of gradient disappearance is avoided. The LSTM adds a structure called Memory Cell (Memory Cell) in a neural node of a hidden layer of a neuron internal structure RNN to memorize dynamic change information of a low frequency trend value part of a time series Output by a wavelet decomposition model in the past, and adds three gate (Input, form, output) structures to control use of their history information of the low frequency trend value part of the time series Output by the wavelet decomposition model. The time series value of the low frequency trend value part of the time series output by the input wavelet decomposition model is set as (x) 1 ，x 2 ，…，x T ) The hidden layer state is (h) 1 ，h 2 ，...，h T ) Then, time t has:
i t ＝sigmoid(W hi h t-1 +W xi X t ) (9)
f t ＝sigmoid(W hf h t-1 +W hf X t ) (10)
c t ＝f t ⊙c t-1 +i t ⊙tanh(W hc h t-1 +W xc X t ) (11)
O t ＝sigmoid(W ho h t-1 +W hx X t +W co c t ) (12)
h t ＝o t ⊙tanh(c t ) (13)
wherein i t 、J t 、O t Stands for input, forget, and output doors, c t Represents cell unit, wh represents weight of recursive connection, wx represents weight of input layer to hidden layer, sigmoid and tanh are two activation functions. The current output of the NARX neural network model depends not only on the past NARX neural network model output y (t-n), but also on the current LSTM neural network output vector X (t) of the NARX neural network model, the delay order of the LSTM neural network output vector, and so on. The output of the LSTM neural network of the NARX neural network model is transmitted to the hidden layer through the time delay layer, the hidden layer processes signals output by the LSTM neural network and then transmits the processed signals to the output layer, the output layer linearly weights the output signals of the hidden layer to obtain final output signals of the neural network, and the time extension layer delays signals fed back by the NARX neural network model and signals output by the LSTM neural network of the input layer and then transmits the delayed signals to the hidden layer. The NARX neural network model has the characteristics of nonlinear mapping capability, good robustness, adaptability and the like, and is suitable for further processing the trend part of the output predicted value of the parameter sensor.
(5) BAM neural network-AANN self-association neural network model design of interval hesitation fuzzy number
The outputs of the TDL beat-to-beat delay line B, the TDL beat-to-beat delay line C and the TDL beat-to-beat delay line D are respectively used as corresponding inputs of a BAM neural network-AANN auto-associative neural network model with interval hesitation fuzzy numbers, and 4 parameters output by the BAM neural network-AANN auto-associative neural network model with the interval hesitation fuzzy numbers are respectivelya. b, c and D, a, b, c and D as input of TDL beat-per-beat delay line D, a and b constituting interval number (a, b) as minimum value of detected parameter, c and D constituting interval number (c, D) as maximum value of detected parameter, interval number (a, b) and interval number (c, D) constituting [ (a, b), (c, D)]And outputting a BAM neural network-AANN auto-associative neural network model as the interval hesitation fuzzy number of the detected parameter and outputting the interval hesitation fuzzy number as the output of the parameter detection model. The BAM neural network-AANN auto-associative neural network model is that BAM neural network output is used as AANN auto-associative neural network model input, in the topological structure of the BAM neural network model, the initial mode of the network input end is x (t), and the initial mode is obtained by a weight matrix W 1 Weighted and then reaches the y end of the output end, and passes through the transfer characteristic f of the output node y Non-linear transformation of (1) and (W) 2 The matrix is weighted and returns to the input end x, and then the transfer characteristic f of the output node at the x end is passed x The nonlinear transformation of the BAM neural network model is changed into the output of the input terminal x, and the operation process is repeated, so that the state transition equation of the BAM neural network model is shown in an equation (14).
The BAM neural network output is an AANN Auto-associative neural network model input, the AANN Auto-associative neural network model is a feedforward Auto-associative neural network (AANN) with a special structure, and the AANN Auto-associative neural network model structure comprises an input layer, a certain number of hidden layers and an output layer. The method comprises the steps of firstly realizing compression of BAM neural network output data information through an input layer, a mapping layer and a bottleneck layer of parameters, extracting a most representative low-dimensional subspace reflecting BAM neural network output from a high-dimensional parameter space output by the BAM neural network, effectively filtering noise and measurement errors in BAM neural network output data, realizing decompression of the BAM neural network output through the bottleneck layer, the demapping layer and the output layer, and restoring the compressed information to each input value, thereby realizing reconstruction of the BAM neural network output data. In order to achieve the purpose of compressing output information of the BAM neural network, the number of nodes of a bottleneck layer of an AANN auto-associative neural network model is obviously smaller than that of an input layer, and in order to prevent the formation of simple single mapping between output of the BAM neural network and an output layer of the AANN auto-associative neural network, except that an excitation function of the output layer of the AANN auto-associative neural network adopts a linear function, other layers all adopt nonlinear excitation functions. In essence, the first layer of the hidden layer of the AANN auto-associative neural network model is called as a mapping layer, and the node transfer function of the mapping layer can be an S-shaped function or other similar nonlinear functions; the second layer of the hidden layer is called a bottleneck layer, the dimension of the bottleneck layer is the minimum in the network, the transfer function of the second layer can be linear or nonlinear, the bottleneck layer avoids the mapping relation that the output and the input are equal and can be easily realized in a one-to-one mode, the bottleneck layer enables the network to encode and compress the output signal of the BAM neural network, and the decoding and the decompression of the output data of the BAM neural network are carried out after the bottleneck layer to generate the estimation value of the output signal of the BAM neural network; the third layer or the last layer of the hidden layer is called a demapping layer, the node transfer function of the demapping layer is a generally nonlinear sigmoid function, and the AANN self-association neural network is trained by an error back propagation algorithm.
2. Design of interval hesitation fuzzy number fusion model
(1) Constructing a time sequence interval hesitation fuzzy numerical array of the parameter measurement sensor
Interval hesitation fuzzy numbers output by parameter detection models of a plurality of parameter measuring sensors at a period of time form a time sequence interval hesitation fuzzy value array, interval hesitation fuzzy values of n parameter measuring sensors and nm parameter measuring sensors at m moments form a time sequence interval hesitation fuzzy value array of n rows and m columns of parameter measuring sensors, and interval hesitation fuzzy values at different moments of the same parameter measuring sensor are set as A ij (t)，A ij (t+1)，...，A ij (m), then the time series interval hesitation fuzzy value array of all parameter measurement sensors is:
(2) distance fusion weight alpha of hesitation fuzzy number of time series interval of parameter measurement sensor i Is obtained by
Because of the randomness that the detection environment parameters are interfered by various factors, the expression form of the distance dij between the time series interval hesitation fuzzy number of the ith measuring sensor and the time series interval hesitation fuzzy number of the jth measuring sensor is as follows:
the distances between the time series interval hesitant ambiguity numbers output by the n different measurement sensors form a distance array of the n measurement sensors as follows:
distance matrix R where d ij Only the distance between the time series interval hesitation fuzzy number of the ith measuring sensor and the time series interval hesitation fuzzy number of the jth measuring sensor is represented, and the average value of the distances between the time series interval hesitation fuzzy number of each measuring sensor and the time series interval hesitation fuzzy numbers of other measuring sensors can be obtained according to the distance matrix R as follows:
the distance fusion weight of the ratio of the reciprocal of the distance average of the time series interval hesitation fuzzy numbers of each measuring sensor to the sum of the reciprocals of the distance average of the time series interval hesitation fuzzy numbers of all the parameter measuring sensors is as follows:
(3) and the correlation fusion weight beta of the hesitation fuzzy number in the time series interval of the parameter measurement sensor i Is obtained by
The expression form of the correlation degree Kij of the time series interval hesitation fuzzy number of the ith measuring sensor and the time series interval hesitation fuzzy number of the jth measuring sensor is as follows:
wherein:
then the correlation among the time series interval hesitant fuzzy numbers output by the n different measuring sensors forms a correlation array S of the n measuring sensors as follows:
the average value of the correlation degrees of the hesitation fuzzy numbers output by each measuring sensor in each row of the matrix S is as follows:
determining the correlation fusion weight of the hesitation fuzzy numbers of the time sequence intervals of the parameter measuring sensors according to the ratio of the correlation average value of the hesitation fuzzy numbers of the time sequence intervals of each measuring sensor to the correlation average value sum of the hesitation fuzzy numbers of the time sequence intervals of all measuring sensors:
(4) combining weight w i Is obtained by
Obtaining the interval number fusion weight of the hesitation fuzzy value of the time sequence interval of the parameter measurement sensor according to the formula (19) and the formula (23) as w i ：
w i ＝[min(α i ，β i )，max(α i ，β i )] (24)
From the formula (18), it can be known that the distance fusion weight of the time series interval hesitation fuzzy number of each parameter measurement sensor and the correlation fusion weight of the time series interval hesitation fuzzy number of the parameter measurement sensor are used as the interval number fusion weight of the time series interval hesitation fuzzy number of the parameter measurement sensor, wherein the interval numbers are formed by sequencing the interval numbers from small to large; the sum of the products of the time series interval hesitation fuzzy numbers of each parameter measuring sensor and the interval number fusion weight of the time series interval hesitation fuzzy values of the parameter measuring sensors at the same moment is the interval hesitation fuzzy value of the time series of the measured parameters, and the fusion value of the time series interval hesitation fuzzy numbers output by all the parameter measuring sensors is the interval hesitation fuzzy value of the time series of the measured parameters, and the interval hesitation fuzzy value of the time series is as follows:
the structure of the interval hesitation fuzzy number fusion model is shown in figure 2.
And constructing a storage tank safety monitoring subsystem, wherein the storage tank safety monitoring subsystem consists of a parameter detection module, a parameter detection model, an interval hesitation fuzzy number fusion model, a TDL beat-to-beat delay line, a noise reduction self-coding neural network model, a particle swarm optimization self-adaptive wavelet neural network, a RBF neural network model and an interval BAM neural network-LSTM neural network storage tank safety classifier, and the interval BAM neural network-LSTM neural network storage tank safety classifier takes the output of the BAM neural network as the input of the LSTM neural network storage tank safety classifier.
The radial basis vector of the RBF neural network model is H = [ H ] 1 ，h 2 ,…；h p ] T ，h p For basis functions, a commonly used radial basis function in a radial basis function neural network is a gaussian function, and its expression is:
wherein X is the time sequence output of 3 TDLs according to the beat delay line T, C is the coordinate vector of the central point of Gaussian basis function of neuron in hidden layer, delta j The width of the Gaussian base function of the jth neuron of the hidden layer; the output connection weight vector of the network is w ij The output expression of the RBF neural network model is as follows:
the 2 parameters output by the BAM neural network-LSTM neural network storage tank safety classifier with the interval number are e and f respectively, the interval number (e, f) is formed by the e and the f and is used as the interval number of the safety level of the detected storage tank, and the output of the LSTM neural network storage tank safety classifier at the time not only depends on the past output y (t-n) of the LSTM neural network storage tank safety classifier, but also depends on the current output vector X (t) of the BAM neural network of the LSTM neural network storage tank safety classifier, the delay order of the output vector of the BAM neural network and the like. The output of the BAM neural network of the LSTM neural network storage tank safety classifier is transmitted to the hidden layer through the time delay layer, the hidden layer processes signals output by the BAM neural network and then transmits the signals to the output layer, the output layer linearly weights the hidden layer output signals to obtain final neural network output signals, and the time delay layer delays signals fed back by the LSTM neural network storage tank safety classifier and signals output by the BAM neural network of the input layer and then transmits the signals to the hidden layer. The LSTM neural network storage tank security classifier has the characteristics of non-linear mapping capability, good robustness, self-adaptability and the like. The interval number output by the BAM neural network-LSTM neural network storage tank safety classifier of the interval number respectively corresponds to 5 storage tank safety types of good storage tank condition, common storage tank condition, poor storage tank condition and poor storage tank condition; the relation between the output interval number of the BAM neural network-LSTM neural network storage tank safety classifier and the storage tank safety level is shown in a corresponding table 1, and the structure of a storage tank safety monitoring subsystem is shown in a figure 2.
TABLE 1 relationship between number of intervals and safety level of storage tank
| Serial number | Storage tank safety class | Number of intervals |
| 1 | The storage tank has good condition | (0.0，0.2) |
| 2 | The storage tank has better condition | (0.2，0.4) |
| 3 | General condition of storage tank | (0.4，0.6) |
| 4 | Poor storage tank condition | (0.6，0.8) |
| 5 | Poor storage tank condition | (0.8，1.0) |
3. Storage tank safety monitoring system based on Internet of things
Based on thing networking storage tank safety monitoring system includes test terminal, the on-the-spot control end, the gateway node, cloud platform and cell-phone APP, test terminal is responsible for gathering by storage tank safety parameter information, realize test terminal through the gateway node, the on-the-spot control end, cloud platform and cell-phone APP's both way communication, there is storage tank safety monitoring subsystem to realize gathering by storage tank safety parameter and the classification of storage tank safety class in the on-the-spot control end, cell-phone APP realizes gathering and monitoring storage tank safety parameter through 5G network access cloud platform. The specific implementation mode is as follows:
1. design of overall system function
The storage tank safety monitoring system based on the Internet of things realizes detection of storage tank safety parameters and classification of storage tank safety grades, and detection terminals of a plurality of storage tank safety parameters of the system are constructed into a wireless measurement and control network in a self-organizing manner to realize bidirectional wireless communication of the detection terminals, gateway nodes, field monitoring terminals, cloud platforms and mobile phone APPs; the detection terminal sends the detected storage tank safety parameters to the field monitoring terminal and the cloud platform through the gateway node, and a storage tank safety monitoring subsystem of the field monitoring terminal processes the storage tank safety parameters and classifies the storage tank safety level; and the storage tank safety parameters and the grade are monitored in real time at the APP end of the mobile phone through accessing the cloud platform. A storage tank safety monitoring system based on the Internet of things is shown in figure 3.
2. Design of detection terminal
A large number of detection terminals based on the Internet of things are used as storage tank safety parameter sensing terminals, and the mutual information interaction between the on-site monitoring terminal and the cloud platform is realized through the self-organizing wireless network by the detection terminals and the gateway nodes. The detection terminal comprises displacement, vibration, temperature, flow, pressure and tilt sensors for collecting storage tank safety parameters, a corresponding signal conditioning circuit, an STM32 single chip microcomputer and a CC2530 wireless transmission module; software of the detection terminal mainly realizes wireless communication and acquisition and pretreatment of storage tank safety parameters. The software is designed by adopting a C language program, so that the compatibility degree is high, the working efficiency of software design and development is greatly improved, and the reliability, readability and transportability of program codes are enhanced. The structure of the detection terminal is shown in figure 4.
3. Gateway node design
The gateway node comprises a CC2530 module, an NB-IoT module, an STM32 single chip microcomputer and an RS232 interface, the gateway node comprises a self-organizing communication network which is used for realizing communication with the detection terminal through the CC2530 module, the NB-IoT module realizes data bidirectional interaction between the gateway and the cloud platform, and the RS232 interface is connected with the field monitoring terminal to realize information interaction between the gateway and the field monitoring terminal. The gateway node is shown in figure 5.
3. Site monitoring terminal software
The on-site monitoring terminal is an industrial control computer, mainly collects storage tank safety parameters and a storage tank safety grade classifier, realizes information interaction with the detection terminal, the gateway node and the cloud platform, and mainly has the functions of communication parameter setting, data analysis and data management and a storage tank safety monitoring subsystem. The management software selects Microsoft Visual + +6.0 as a development tool, calls the Mscomm communication control of the system to design a communication program, and the functions of the field monitoring end software are shown in FIG. 6.
The above embodiments are only for illustrating the technical idea and features of the present invention, and the purpose of the embodiments is to enable those skilled in the art to understand the content of the present invention and implement the present invention, and not to limit the protection scope of the present invention by this means. All equivalent changes and modifications made according to the spirit of the present invention should be covered within the protection scope of the present invention.
Claims (7)
1. A storage tank safety monitoring method based on the Internet of things is characterized by comprising the following steps:
step 1, constructing a parameter detection model: the parameter detection model comprises a particle swarm optimization adaptive wavelet neural network-NARX neural network model, a TDL beat-by-beat delay line A, a wavelet decomposition model, a noise reduction self-coding neural network-NARX neural network model, an LSTM neural network-NARX neural network model, a TDL beat-by-beat delay line B, a TDL beat-by-beat delay line C, a TDL beat-by-beat delay line D and a BAM neural network-AANN auto-association neural network model of interval hesitation fuzzy numbers;
step 2, constructing an interval hesitation fuzzy number fusion model, forming a time sequence interval hesitation fuzzy number array by interval hesitation fuzzy numbers output by parameter detection models of a plurality of parameter measurement sensors in a period of time, defining the distance and the correlation degree of the time sequence interval hesitation fuzzy numbers of the parameter measurement sensors, and constructing a distance matrix and a correlation matrix;
step 3, constructing a storage tank safety monitoring subsystem, wherein the storage tank safety monitoring subsystem is composed of a parameter detection module, a parameter detection model, an interval hesitation fuzzy number fusion model, a TDL beat-to-beat delay line, a noise reduction self-coding neural network model, a particle swarm optimization self-adaptive wavelet neural network, a RBF neural network model and a BAM neural network-LSTM neural network storage tank safety classifier of the interval number;
and 4, outputting by using a plurality of temperature sensors, a plurality of pressure sensors and a flow sensor as the input of the storage tank safety monitoring subsystem, and outputting the safety type of the monitored storage tank by the storage tank safety monitoring subsystem.
2. The Internet of things-based storage tank security monitoring method according to claim 1, wherein in the step 1, the output of the parameter sensor is used as the input of a particle swarm optimization adaptive wavelet neural network-NARX neural network model, the output of the particle swarm optimization adaptive wavelet neural network-NARX neural network model is used as the input of a TDL beat-to-beat delay line A, the output of the TDL beat-to-beat delay line A is used as the input of a wavelet decomposition model, the low-frequency trend part and the high-frequency fluctuation parts output by the wavelet decomposition model are respectively used as the input of an LSTM neural network-NARX neural network model and a noise reduction self-coding neural network-NARX neural network model, the outputs of the LSTM neural network-NARX neural network model and the noise reduction self-coding neural network-NARX neural network model are respectively used as the input of a TDL beat-to-delay line B and a TDL beat-to-delay line C, the outputs of the TDL beat-to-delay line B, the TDL beat to-NARX neural network model and the TDL to delay line D are respectively used as the corresponding inputs of a BAM neural network-AANN self-associated neural network model, and the output of the fuzzy BANN neural network output as fuzzy BANN parameters of the AAb-NN associated neural network model, and the TDL associated neural network output.
3. The Internet of things-based tank safety monitoring method according to claim 2, wherein the number of intervals (a, b) formed by a and b is used as the minimum value of the detected parameter, the number of intervals (c, d) formed by c and d is used as the maximum value of the detected parameter, and the number of intervals (a, b) and the number of intervals (c, d) are used as [ (a, b), (c, d) ] fuzzy numbers of intervals of the detected parameter.
4. The Internet of things-based storage tank safety monitoring method according to claim 1, wherein in the step 2, the ratio of the reciprocal of the distance average of the time-series interval hesitation fuzziness of each measuring sensor to the sum of the distance average reciprocals of the time-series interval hesitation fuzziness of all the parameter measuring sensors is the distance fusion weight of the time-series interval hesitation fuzziness of the parameter measuring sensor, the ratio of the correlation average of the time-series interval hesitation fuzziness of each measuring sensor to the sum of the correlation average values of the time-series interval hesitation fuzziness of all the measuring sensors is the correlation fusion weight of the time-series interval hesitation fuzziness of the parameter measuring sensor, and the distance fusion weight of the time-series interval hesitation fuzziness of each parameter measuring sensor and the correlation fusion weight of the time-series interval hesitation fuzziness of the parameter measuring sensor are sequentially arranged from small to large as the fusion weight of the time-series interval hesitation fuzziness of the parameter measuring sensor; and the sum is the fusion value of the time series interval hesitation fuzzy numbers of all the parameter measuring sensors according to the sum of products of the time series interval hesitation fuzzy number of each parameter measuring sensor at the same moment and interval number fusion weights of the time series interval hesitation fuzzy numbers of the parameter measuring sensors.
5. The Internet of things-based storage tank safety monitoring method according to claim 1, wherein in the step 3, the parameter detection module comprises a parameter detection model, the storage tank temperature value output by the temperature sensor for a period of time is used as corresponding input of a plurality of parameter detection models corresponding to the parameter detection module, and the output of the plurality of parameter detection models is used as the output of the parameter detection module; the storage tank pressure values output by the pressure sensors for a period of time are used as corresponding inputs of a plurality of parameter detection models of the corresponding parameter detection module, and the outputs of the parameter detection models are used as the outputs of the parameter detection module; the output of 2 parameter detection modules is used as the input of a corresponding interval hesitation fuzzy number fusion model, the output of a flow sensor is used as the input of the parameter detection model, the output of the parameter detection model and the output of the 2 interval hesitation fuzzy number fusion models are respectively used as the corresponding TDL input according to a beat delay line, the output of 3 TDL outputs according to a beat delay line are respectively used as the corresponding input of a noise reduction self-coding neural network model, a particle swarm optimization self-adaptive wavelet neural network and an RBF neural network model, the output of the noise reduction self-coding neural network model, the output of the particle swarm optimization self-adaptive wavelet neural network and the output of the RBF neural network model are respectively used as the interval number BAM neural network-LSTM neural network storage tank safety classifier, the 2 parameters output by the interval number BAM neural network-LSTM neural network storage tank safety classifier are respectively used as the interval number of a detected safety storage tank, and the interval number output by the interval number BAM neural network-LSTM neural network storage tank safety classifier is respectively corresponding to 5 safety types.
6. Storage tank safety monitoring system based on Internet of things, which is characterized in that the detection system comprises a detection terminal and a field monitoring terminal, the detection terminal is responsible for collecting parameter information of a detected storage tank, a storage tank safety monitoring subsystem is arranged in the field monitoring terminal, bidirectional communication of the detection terminal, the field monitoring terminal, a cloud platform and a mobile phone APP is realized through a gateway node, the parameter collection of the detected storage tank and the classification of the safety type of the storage tank are realized, and the detection terminal and the field monitoring terminal are loaded with computer program steps for realizing the storage tank safety monitoring method based on Internet of things according to any one of claims 1-5.
7. The Internet of things-based storage tank safety monitoring system according to claim 6, wherein the detection terminal comprises displacement, vibration, temperature, flow, pressure and tilt sensors for acquiring storage tank safety parameters, a corresponding signal conditioning circuit, an STM32 single chip microcomputer and a CC2530 wireless transmission module.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211306368.1A CN115905938B (en) | 2022-10-24 | 2022-10-24 | Storage tank safety monitoring method and system based on Internet of things |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211306368.1A CN115905938B (en) | 2022-10-24 | 2022-10-24 | Storage tank safety monitoring method and system based on Internet of things |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115905938A true CN115905938A (en) | 2023-04-04 |
| CN115905938B CN115905938B (en) | 2024-04-05 |
Family
ID=86490627
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211306368.1A Active CN115905938B (en) | 2022-10-24 | 2022-10-24 | Storage tank safety monitoring method and system based on Internet of things |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115905938B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110702852A (en) * | 2019-09-10 | 2020-01-17 | 淮阴工学院 | Intelligent monitoring system of multiple oil-gas concentration sensors based on Internet of things |
| CN110716012A (en) * | 2019-09-10 | 2020-01-21 | 淮阴工学院 | Oil gas concentration intelligent monitoring system based on field bus network |
| CN111220202A (en) * | 2018-11-23 | 2020-06-02 | 中国科学院大连化学物理研究所 | Dangerous chemical solution safety early warning remote monitoring system and method based on Internet of things |
| CN112903929A (en) * | 2021-01-13 | 2021-06-04 | 淮阴工学院 | Food quality detection system |
| CN112904756A (en) * | 2021-01-13 | 2021-06-04 | 淮阴工学院 | Pipe network big data detection system |
| CN114839881A (en) * | 2022-06-17 | 2022-08-02 | 淮阴工学院 | Intelligent garbage cleaning and environmental parameter big data internet of things system |
| CN114911185A (en) * | 2022-06-17 | 2022-08-16 | 淮阴工学院 | Security big data Internet of things intelligent system based on cloud platform and mobile terminal App |
| CN114995248A (en) * | 2022-06-17 | 2022-09-02 | 淮阴工学院 | Intelligent maintenance and environmental parameter big data internet of things system |
| CN115016276A (en) * | 2022-06-17 | 2022-09-06 | 淮阴工学院 | Intelligent moisture regulation and environmental parameter Internet of things big data system |
-
2022
- 2022-10-24 CN CN202211306368.1A patent/CN115905938B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111220202A (en) * | 2018-11-23 | 2020-06-02 | 中国科学院大连化学物理研究所 | Dangerous chemical solution safety early warning remote monitoring system and method based on Internet of things |
| CN110702852A (en) * | 2019-09-10 | 2020-01-17 | 淮阴工学院 | Intelligent monitoring system of multiple oil-gas concentration sensors based on Internet of things |
| CN110716012A (en) * | 2019-09-10 | 2020-01-21 | 淮阴工学院 | Oil gas concentration intelligent monitoring system based on field bus network |
| CN112903929A (en) * | 2021-01-13 | 2021-06-04 | 淮阴工学院 | Food quality detection system |
| CN112904756A (en) * | 2021-01-13 | 2021-06-04 | 淮阴工学院 | Pipe network big data detection system |
| CN114839881A (en) * | 2022-06-17 | 2022-08-02 | 淮阴工学院 | Intelligent garbage cleaning and environmental parameter big data internet of things system |
| CN114911185A (en) * | 2022-06-17 | 2022-08-16 | 淮阴工学院 | Security big data Internet of things intelligent system based on cloud platform and mobile terminal App |
| CN114995248A (en) * | 2022-06-17 | 2022-09-02 | 淮阴工学院 | Intelligent maintenance and environmental parameter big data internet of things system |
| CN115016276A (en) * | 2022-06-17 | 2022-09-06 | 淮阴工学院 | Intelligent moisture regulation and environmental parameter Internet of things big data system |
Non-Patent Citations (3)
| Title |
|---|
| ABDULLAH ALHARBI ET AL.: "Analyzing the Impact of Cyber Security Related Attributes for Intrusion Detection Systems", SUSTAINABILITY, vol. 13 * |
| 杨威超等: "数据驱动的物联网安全威胁检测与建模", 中国优秀硕士学位论文全文数据库信息科技辑, no. 02 * |
| 马从国等: "国内养殖池塘溶解氧智能监测与调控研究进展", 中国农机化学报, vol. 37, no. 03 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115905938B (en) | 2024-04-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112904756B (en) | Pipe network big data detection system | |
| CN113126676B (en) | Livestock and poultry house breeding environment parameter intelligent control system | |
| CN112665656B (en) | Big data detection system of agricultural product growth environment | |
| CN111476278B (en) | Intelligent gas concentration detection system | |
| CN111461187B (en) | Intelligent building settlement detection system | |
| CN115016276B (en) | Intelligent water content adjustment and environment parameter Internet of things big data system | |
| CN110716012B (en) | Oil gas concentration intelligent monitoring system based on field bus network | |
| CN112903929A (en) | Food quality detection system | |
| CN115630101A (en) | Hydrological parameter intelligent monitoring and water resource big data management system | |
| CN115687995A (en) | Big data environmental pollution monitoring method and system | |
| CN113281465A (en) | Livestock and poultry house breeding environment harmful gas detection system | |
| CN114911185A (en) | Security big data Internet of things intelligent system based on cloud platform and mobile terminal App | |
| CN115128978A (en) | Internet of things environment big data detection and intelligent monitoring system | |
| CN112911533B (en) | Temperature detection system based on remove end App | |
| CN113283642A (en) | Poultry feed detection and formula system | |
| CN113255739A (en) | Fish feed detection and formula system | |
| CN117221352A (en) | Internet of things data acquisition and intelligent big data processing method and cloud platform system | |
| CN114358244B (en) | Big data intelligent detection system of pressure based on thing networking | |
| CN114970745B (en) | Intelligent security and environment big data system of Internet of things | |
| CN115659201A (en) | Gas concentration detection method and monitoring system for Internet of things | |
| CN115905938A (en) | Storage tank safety monitoring method and system based on Internet of things | |
| CN114386672B (en) | Environment big data Internet of things intelligent detection system | |
| CN115616163A (en) | Gas accurate preparation and concentration measurement system | |
| CN114995248A (en) | Intelligent maintenance and environmental parameter big data internet of things system | |
| CN116108361B (en) | Intelligent oil quality detection method and monitoring system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |