CN112651176A - GA-BP algorithm-based shielded door motor output force prediction method - Google Patents

GA-BP algorithm-based shielded door motor output force prediction method Download PDF

Info

Publication number
CN112651176A
CN112651176A CN202011553375.2A CN202011553375A CN112651176A CN 112651176 A CN112651176 A CN 112651176A CN 202011553375 A CN202011553375 A CN 202011553375A CN 112651176 A CN112651176 A CN 112651176A
Authority
CN
China
Prior art keywords
output
neural network
algorithm
shield door
data
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.)
Pending
Application number
CN202011553375.2A
Other languages
Chinese (zh)
Inventor
贺德强
袁誉钊
陈彦君
李先旺
靳震震
班玉友
马瑞
邹雪妍
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangxi University
Original Assignee
Guangxi University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Guangxi University filed Critical Guangxi University
Priority to CN202011553375.2A priority Critical patent/CN112651176A/en
Publication of CN112651176A publication Critical patent/CN112651176A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Data Mining & Analysis (AREA)
  • Artificial Intelligence (AREA)
  • Biophysics (AREA)
  • Health & Medical Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Evolutionary Biology (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Mathematical Physics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Computer Hardware Design (AREA)
  • Geometry (AREA)
  • Physiology (AREA)
  • Genetics & Genomics (AREA)
  • Power-Operated Mechanisms For Wings (AREA)

Abstract

The invention discloses a shield door motor output force prediction method based on a GA-BP algorithm, which comprises the steps of firstly obtaining subway shield door fault data, preprocessing the data, and then obtaining the shield door motor predicted output force according to the GA-BP algorithm; coding the weight and the threshold of the neural network to obtain an initial population; training a neural network, and calculating the output error of the neural network; introducing a genetic algorithm and calculating the fitness, and judging whether the output error meets a termination condition; and if the termination condition is met, decoding to obtain the optimal weight and the threshold. The method has the advantages of high calculation speed and high solving precision, and can effectively predict the output force of the motor of the shield door, implement adjustment and reduce the faults of the shield door.

Description

GA-BP algorithm-based shielded door motor output force prediction method
Technical Field
The invention belongs to the technical field of urban rail transit vehicle safety, particularly relates to a masking dream fault identification technology, and particularly relates to a masking door motor output force prediction method based on a GA-BP algorithm.
Background
The urban rail train has the advantages of convenience, rapidness, comfort, safety, high punctuality rate and the like, becomes an important travel mode for people, and plays an important role in relieving urban traffic pressure. In order to meet the traveling requirements of passengers, the number of travelling pairs of existing lines is continuously encrypted, the running speed of a train is increased, alternating load caused by tunnel piston wind is remarkably increased, and the most direct influence is that the opening and closing of a shielding door is slowed or even the door cannot be opened and closed. The shield door system is widely applied to an existing subway station environmental control system, and the conditions of the pressure bearing of the shield door are influenced by the factors of the number and the state of trains, the type of the trains, the speed of the train, the length and the structure of a tunnel, the form of a station platform (island type or side type), whether the upper heat extraction and the lower heat extraction of a station trackway area are opened or not and the opening state of piston air shafts at two ends of the station, so that the condition can be obtained to be more complex. When a train stops in a station and the train is started and accelerated afterwards, air between the two trains can form piston extrusion, and due to local resistance formed by train stopping, a dynamic pressure head part of longitudinal wind speed formed in a tunnel is converted into static pressure, so that the pressure born by the shield door is increased, and the root cause of unsmooth closing of the shield door is caused. In the prior art, the technical research for predicting the piston wind in the subway tunnel and adjusting the output force of the shield door system control motor is not mature. Therefore, it is necessary to provide a method for predicting the output force of the shield door motor to solve the above technical problems.
Disclosure of Invention
The invention aims to: aiming at the existing problems, the invention provides a shield door motor output force prediction method based on the GA-BP algorithm, which can predict the piston wind in the subway tunnel, effectively predict the shield door motor output force and implement adjustment, has high calculation speed and high solution precision, and reduces the shield door faults. In order to achieve the purpose, the invention adopts the following technical scheme:
the invention provides a shield door motor output force prediction method based on a GA-BP algorithm, which is characterized by comprising the following steps: the prediction method comprises the following steps:
step S1, acquiring data, preprocessing the data and determining the topological structure of the neural network;
step S2, encoding the weight and the threshold of the neural network to obtain an initial population;
step S3, training a neural network, and calculating the output error of the neural network;
step S4, introducing a genetic algorithm and calculating fitness, and judging whether the output error meets a termination condition;
and step S5, if the termination condition is met, decoding to obtain the optimal weight and threshold.
Further preferably, the step of acquiring the data and preprocessing the data includes: the data acquisition comprises the steps of dividing the data into training set data and test set data, and preprocessing the training set data and the test set data through a normalization formula and an inverse normalization formula.
Preferably, the step of determining the topology of the neural network includes constructing a BP neural network and constructing hidden layer nodes;
constructing the BP neural network through mapping from any cluster of n dimension to m dimension to form a three-layer BP neural network;
constructing an empirical design formula of the hidden layer nodes satisfying the number of the hidden layer nodes
Figure BDA0002857635530000021
In the formula, l is the number of hidden layer nodes, n is the number of input layer nodes, m is the number of output layer nodes, and alpha is an adjustment coefficient between 1 and 10.
Preferably, the step of coding the weight and the threshold of the neural network to obtain the initial population comprises: initializing the weight, threshold, error and learning rate of the network; in the invention, the number of nodes of a network input layer, the number of nodes of a hidden layer and the number of nodes of an output layer are determined according to a system input and output sequence, the connection weight among neurons of the input layer, the hidden layer and the output layer is initialized, the threshold of the hidden layer and the threshold of the output layer are initialized, and the learning rate and the neuron excitation function are given.
Further preferably, the calculating the output error of the neural network includes the following steps: taking a sigmoid function and a Purelin function as transfer functions of a hidden layer and an output layer, wherein the sigmoid function and the Purelin function are respectively as follows:
Figure BDA0002857635530000022
ψ(x)＝x；
the overall error function E of the network output is:
Figure BDA0002857635530000023
wherein S is the number of all samples; t is the experimental target output after the normalization process, and O is the neural network prediction output, wherein:
Figure BDA0002857635530000031
where w is the connection weight, y is the output of the hidden layer, x is the input to the network, b is the neuron bias, and wijAnd representing the connection weight between the input layer and the hidden layer, wherein i and j respectively represent different neuron nodes.
Further preferably, the introducing the genetic algorithm and calculating the fitness comprises: calculating individual adaptation values of the initial population, and judging whether the output error meets a termination condition or not according to the calculation of the individual adaptation values; and judging whether the output error is smaller than the threshold value, and if the output error is larger than the threshold value, returning to execute the step S3.
Preferably, the calculation of the individual fitness value includes fitness function of genetic algorithm, selection function calculation of genetic algorithm, crossover operation and mutation operation; the fitness function F of the genetic algorithm satisfies:
Figure BDA0002857635530000032
where n denotes the total number of training samples, oiIndicating the expected output value, y, of the ith sample dataiRepresenting the actual output value of the ith sample data network model;
the selection function of the genetic algorithm is:
Figure BDA0002857635530000033
Figure BDA0002857635530000034
wherein M represents the size of the selected population, fiDenotes individual fitness, PiRepresenting the probability that the individual may be selected;
the cross operation is to carry out cross pairing by randomly selecting chromosomes in a population, and then selecting a random position k to carry out the following operation:
Figure BDA0002857635530000035
in the formula, b is [0, 1 ]]Random number between, xkAnd ykEach representing two different chromosomes, x'kAnd y'kRespectively represent xkAnd ykChromosomes after crossover operation and the relationship x 'k + y' k as x before and after the crossover of paired chromosomesk+yk；
The mutation operation is to randomly find a chromosome X from the population,
wherein X is (X)1,x2,…,xk,…,xm) For its component xkMutation is carried out according to a certain mutation probability, and the chromosome after mutation is X' ═ (X)1,x2,…,x'k,…,xm)：
Figure BDA0002857635530000041
f(g)＝γ2(1-g/gnax)2；
In the formula, ak，bkIs a component xkG is the evolution generation number of the chromosome, g is the upper and lower bounds ofmaxIs the maximum evolution algebra
In summary, due to the adoption of the technical scheme, the invention has the following beneficial effects: compared with the prior art, the shield door motor output force prediction method based on the GA-BP algorithm, provided by the invention, is characterized in that a neural network prediction experiment is carried out through a piston wind numerical simulation sample, the prediction mean square error and the correlation coefficient of each group of samples are compared to determine the neural network topological structure and the genetic algorithm population scale, and finally, the validity and the applicability of the method are proved through verification and simulation. The method can provide effective and accurate shield door motor output force prediction, has high calculation speed and high solving precision, can effectively predict the shield door motor output force and implement adjustment, and reduces shield door faults.
Drawings
FIG. 1 is a flow chart of a method for predicting the output force of a motor of a shielded door based on GA-BP algorithm according to the present invention;
FIG. 2 is a block diagram of a three-layer BP neural network of the present invention;
FIG. 3 is a flow chart of the genetic algorithm operation of the present invention;
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail below with reference to the accompanying drawings by way of examples of preferred embodiments. It should be noted, however, that the numerous details set forth in the description are merely for the purpose of providing the reader with a thorough understanding of one or more aspects of the invention, even though such aspects of the invention may be practiced without these specific details.
As shown in fig. 1, according to the shield door motor output power prediction method based on the GA-BP algorithm of the present invention, firstly, failure data of a subway shield door is obtained, the data is preprocessed, and then, the predicted output power of the shield door motor is obtained according to the GA-BP algorithm, the output power prediction method includes the following steps:
step S1, acquiring data, preprocessing the data and determining the topological structure of the neural network;
step S2, encoding the weight and the threshold of the neural network to obtain an initial population; in the invention, the step of coding the weight and the threshold of the neural network to obtain the initial test population comprises the following steps: initializing weight, threshold, error and learning rate of the network, determining the number of nodes of a network input layer, the number of nodes of a hidden layer, the number of nodes of an output layer, initializing connection weights among neurons of the input layer, the hidden layer and the output layer, initializing the threshold of the hidden layer and the threshold of the output layer, and setting the learning rate and a neuron excitation function according to input and output sequences of a system (the input sequence of a GA-BP model is respectively train running speed, tunnel length between two trains and piston wind speed;
step S3, training a neural network, and calculating the output error of the neural network;
step S4, introducing a genetic algorithm and calculating fitness, and judging whether the output error meets a termination condition;
and step S5, if the termination condition is met, decoding to obtain the optimal weight and threshold.
In the present invention, the step of acquiring data and preprocessing the data includes: the data acquisition step includes dividing the data into training set data and test set data, and preprocessing the training set data and the test set data through a normalization formula and an inverse normalization formula; wherein the normalization formula is:
Figure BDA0002857635530000051
the inverse normalization formula is:
Figure BDA0002857635530000052
in the formula, x is original data; x is normalized data; x is the number ofmaxAnd xminIs the maximum and minimum values in the raw data; a and b are the upper and lower limits of the normalized data.
The step of determining the topological structure of the neural network comprises the steps of constructing a BP neural network and constructing hidden layer nodes;
the BP neural network is constructed through mapping from any cluster of n dimensions to m dimensions to form a three-layer BP neural network, the number of input neurons is defined to be 2, the number of output neurons is 3, the number of implicit neurons is 6, a three-layer network topological structure of 2-6-3 is obtained, the three-layer neural network topological structure is shown in figure 2, the three-layer neural network mainly comprises an input layer, an implicit layer and an output layer, wherein the connection between the layers is full connection, the three-layer neural network is divided into a forward propagation process, an error reverse transfer process and a weight updating process, the neural network can approximate any nonlinear function with any precision, and in the invention, an empirical design formula that the implicit layer nodes meet the number of the implicit layer nodes is constructed:
Figure BDA0002857635530000053
in the formula, l is the number of hidden layer nodes, n is the number of input layer nodes, m is the number of output layer nodes, and alpha is an adjustment coefficient between 1 and 10.
In the present invention, calculating the output error of the neural network comprises the steps of: taking a sigmoid function and a Purelin function as transfer functions of a hidden layer and an output layer, wherein the sigmoid function and the Purelin function are respectively as follows:
Figure BDA0002857635530000061
ψ(x)＝x； (5)；
the overall error function E of the network output is:
Figure BDA0002857635530000062
wherein S is the number of all samples; t is the experimental target output after the normalization process, and O is the neural network prediction output, wherein:
Figure BDA0002857635530000063
wherein, w is the connection weight, y is the output of the hidden layer, x is the input of the network, b is the neuron bias, i is 1,2, ·, n; j ═ 1,2, ·, m; wherein n and m are integers; w is aijAnd representing the connection weight between the input layer and the hidden layer, wherein i and j respectively represent different neuron nodes.
In the present invention, introducing the genetic algorithm and calculating the fitness include: calculating individual adaptation values of the initial population, wherein the initial population adopts a real number coding method to code a connection weight and a deviation in a BP neural network, and whether an output error meets a termination condition or not is judged according to the individual adaptation value calculation; judging whether the output error is smaller than the threshold value, and if the output error is larger than the threshold value, returning to execute the step S3; as shown in fig. 3, the individual fitness value calculation includes a fitness function of a genetic algorithm, a selection function calculation of the genetic algorithm, a crossover operation, and a mutation operation; the fitness function F of the genetic algorithm satisfies:
Figure BDA0002857635530000064
where n denotes the total number of training samples, oiIndicating the expected output value, y, of the ith sample dataiRepresenting the actual output value of the ith sample data network model;
the selection function of the genetic algorithm is:
Figure BDA0002857635530000071
Figure BDA0002857635530000072
wherein M represents the size of the selected population, fiDenotes individual fitness, PiRepresenting the probability that the individual may be selected;
the cross operation is to carry out cross pairing by randomly selecting chromosomes in a population, and then selecting a random position k to carry out the following operation:
Figure BDA0002857635530000073
in the formula (I), the compound is shown in the specification,b is [0, 1 ]]Random number between, xkAnd ykEach representing two different chromosomes, x'kAnd y'kRespectively represent xkAnd ykChromosomes after crossover operation and the relationship x 'k + y' k as x before and after the crossover of paired chromosomesk+yk；
The mutation operation is to randomly find a chromosome X from the population,
wherein X is (X)1,x2,…,xk,…,xm) The component xk is mutated according to a certain mutation probability, and the chromosome after mutation is X' ═ (X)1,x2,…,x'k,…,xm)：
Figure BDA0002857635530000074
f(g)＝γ2(1-g/gnax)2； (13)；
In the formula, ak，bkIs a component xkG is the evolution generation number of the chromosome, g is the upper and lower bounds ofmaxIs the maximum evolution algebra.
In the invention, data are acquired and preprocessed, and the wind pressure born by the screen door and the output force of a motor of the screen door system under different working conditions in the actual running process of the subway train are selected as sample data; the collected sample data comprises the running speed of the train, the length of a tunnel between two trains, the wind speed of piston wind vertical to the shielded door, the vertical pressure of the piston wind borne by the shielded door and the output force of a motor controlled by the shielded door system; in the example, 96-minute sample data is collected, 86 parts of the collected sample data are used as a training set, the other 10 parts of the collected sample data are used as a test set, and 10 groups of verification samples are shown in table 1; and normalizing the sorted sample data according to formula 1, wherein the data range adopted in the normalization process is [0.10,0.90 ].
TABLE 1
Figure BDA0002857635530000081
In the invention, whether the output error meets a termination condition is judged according to the introduction of a genetic algorithm and the calculation of fitness, if the output error meets the termination condition, decoding is carried out to obtain an optimal weight and a threshold, the obtained optimal solution (the optimal weight and the threshold) is brought into a neural network for training, and a prediction and output error value is calculated; the steps of obtaining the optimal weight and the threshold value are as follows: and (4) satisfying a termination condition, and decoding the function so as to obtain the optimal weight and the threshold. In conclusion, MSE (mean square error can reach 1.3134 x 10) in learning and testing processes by the method-3. As shown in Table 2, the performance of the designed GA-BP model is demonstrated.
TABLE 2
Figure BDA0002857635530000082
As can be seen from Table 2, R for the training, validation and test set2The values are all greater than 99%, indicating that the predicted values have significant agreement with the expected values. Meanwhile, the MSE value can also prove the effective performance of the method, so that the method for predicting the output force of the motor of the shield door based on the GA-BP algorithm can effectively predict the piston wind speed, the piston wind pressure amplitude of the platform shield door and the output force of the control motor of the shield door according to the train running speed and the length of the two-workshop tunnel, and the shield door can be normally closed.
The foregoing is only a preferred embodiment of the present invention, and it should be noted that those skilled in the art can make various improvements and modifications without departing from the principle of the present invention, and these improvements and modifications should also be construed as the protection scope of the present invention.

Claims (7)

1. A shield door motor output force prediction method based on GA-BP algorithm is characterized in that: the prediction method comprises the following steps:
step S1, acquiring data, preprocessing the data and determining the topological structure of the neural network;
step S2, encoding the weight and the threshold of the neural network to obtain an initial population;
step S3, training a neural network, and calculating the output error of the neural network;
step S4, introducing a genetic algorithm and calculating fitness, and judging whether the output error meets a termination condition;
and step S5, if the termination condition is met, decoding to obtain the optimal weight and threshold.
2. The method for predicting the output force of the shield door motor based on the GA-BP algorithm, as claimed in claim 1, wherein: the step of acquiring the data and preprocessing the data comprises the following steps: the data acquisition comprises the steps of dividing the data into training set data and test set data, and preprocessing the training set data and the test set data through a normalization formula and an inverse normalization formula.
3. The method for predicting the output force of the shield door motor based on the GA-BP algorithm, as claimed in claim 1, wherein: the step of determining the topological structure of the neural network comprises the steps of constructing a BP neural network and constructing hidden layer nodes;
constructing the BP neural network through mapping from any cluster of n dimension to m dimension to form a three-layer BP neural network;
constructing an empirical design formula of the hidden layer nodes satisfying the number of the hidden layer nodes
Figure FDA0002857635520000011
In the formula, l is the number of hidden layer nodes, n is the number of input layer nodes, m is the number of output layer nodes, and alpha is an adjustment coefficient between 1 and 10.
4. The method for predicting the output force of the shield door motor based on the GA-BP algorithm, as claimed in claim 1, wherein: the step of coding the weight value and the threshold value of the neural network to obtain the initial test population comprises the following steps: and initializing the weight, the threshold, the error and the learning rate of the network.
5. The method for predicting the output force of the shield door motor based on the GA-BP algorithm, as claimed in claim 1, wherein: calculating the output error of the neural network comprises the following steps: taking a sigmoid function and a Purelin function as transfer functions of a hidden layer and an output layer, wherein the sigmoid function and the Purelin function are respectively as follows:
Figure FDA0002857635520000021
ψ(x)＝x；
the overall error function E of the network output is:
Figure FDA0002857635520000022
wherein S is the number of all samples; t is the experimental target output after the normalization process, and O is the neural network prediction output, wherein:
Figure FDA0002857635520000023
where w is the connection weight, y is the output of the hidden layer, x is the input to the network, b is the neuron bias, and wijAnd representing the connection weight between the input layer and the hidden layer, wherein i and j respectively represent different neuron nodes.
6. The method for predicting the output force of the shield door motor based on the GA-BP algorithm, as claimed in claim 1, wherein: introducing the genetic algorithm and calculating fitness comprises: calculating individual adaptation values of the initial population, and judging whether the output error meets a termination condition or not according to the calculation of the individual adaptation values; and judging whether the output error is smaller than the threshold value, and if the output error is larger than the threshold value, returning to execute the step S3.
7. The method for predicting the output force of the shield door motor based on the GA-BP algorithm, as claimed in claim 6, wherein: the calculation of the individual adaptive value comprises a fitness function of a genetic algorithm, selection function calculation of the genetic algorithm, cross operation and mutation operation; the fitness function F of the genetic algorithm satisfies:
Figure FDA0002857635520000024
where n denotes the total number of training samples, oiIndicating the expected output value, y, of the ith sample dataiRepresenting the actual output value of the ith sample data network model;
the selection function of the genetic algorithm is:
Figure FDA0002857635520000025
Figure FDA0002857635520000031
wherein M represents the size of the selected population, fiDenotes individual fitness, PiRepresenting the probability that the individual may be selected;
the cross operation is to carry out cross pairing by randomly selecting chromosomes in a population, and then selecting a random position k to carry out the following operation:
Figure FDA0002857635520000032
in the formula, b is [0, 1 ]]Random number between, xkAnd ykEach representing two different chromosomes, x'kAnd y'kRespectively represent xkAnd ykChromosomes after crossover operation and the relationship x 'k + y' k as x before and after the crossover of paired chromosomesk+yk；
The mutation operation is to randomly find a chromosome X from the population,
wherein X is (X)1,x2,…,xk,…,xm) For its component xkMutation is carried out according to a certain mutation probability, and the chromosome after mutation is X' ═ (X)1,x2,…,x'k,…,xm)：
Figure FDA0002857635520000033
f(g)＝γ2(1-g/gnax)2；
In the formula, ak，bkIs a component xkG is the evolution generation number of the chromosome, g is the upper and lower bounds ofmaxIs the maximum evolution algebra.
CN202011553375.2A 2020-12-24 2020-12-24 GA-BP algorithm-based shielded door motor output force prediction method Pending CN112651176A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011553375.2A CN112651176A (en) 2020-12-24 2020-12-24 GA-BP algorithm-based shielded door motor output force prediction method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011553375.2A CN112651176A (en) 2020-12-24 2020-12-24 GA-BP algorithm-based shielded door motor output force prediction method

Publications (1)

Publication Number Publication Date
CN112651176A true CN112651176A (en) 2021-04-13

Family

ID=75362686

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011553375.2A Pending CN112651176A (en) 2020-12-24 2020-12-24 GA-BP algorithm-based shielded door motor output force prediction method

Country Status (1)

Country Link
CN (1) CN112651176A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113721121A (en) * 2021-09-02 2021-11-30 长江存储科技有限责任公司 Fault detection method and device for semiconductor process
CN113779868A (en) * 2021-08-13 2021-12-10 一汽奔腾轿车有限公司 Rectangular hole metal plate shielding effectiveness prediction method, system, terminal and storage medium

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107219837A (en) * 2016-03-21 2017-09-29 刘春梅 One kind is used for subway shield door failure monitoring and early warning intelligence system
CN110363344A (en) * 2019-07-11 2019-10-22 安徽理工大学 Probability integral parameter prediction method based on MIV-GP algorithm optimization BP neural network
CN111160520A (en) * 2019-12-06 2020-05-15 南京理工大学 BP neural network wind speed prediction method based on genetic algorithm optimization
CN111324989A (en) * 2020-03-19 2020-06-23 重庆大学 GA-BP neural network-based gear contact fatigue life prediction method

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107219837A (en) * 2016-03-21 2017-09-29 刘春梅 One kind is used for subway shield door failure monitoring and early warning intelligence system
CN110363344A (en) * 2019-07-11 2019-10-22 安徽理工大学 Probability integral parameter prediction method based on MIV-GP algorithm optimization BP neural network
CN111160520A (en) * 2019-12-06 2020-05-15 南京理工大学 BP neural network wind speed prediction method based on genetic algorithm optimization
CN111324989A (en) * 2020-03-19 2020-06-23 重庆大学 GA-BP neural network-based gear contact fatigue life prediction method

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
张江石: "《行为安全管理中的数学模型及应用》", 28 February 2016, 北京：煤炭工业出版社 *
杨伟超等: "地铁活塞风作用下屏蔽门的气动特性分析", 《郑州大学学报 ( 工学版 )》 *
黄育良: "地铁屏蔽门电机功率需求计算", 《科技创新导报》 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113779868A (en) * 2021-08-13 2021-12-10 一汽奔腾轿车有限公司 Rectangular hole metal plate shielding effectiveness prediction method, system, terminal and storage medium
CN113721121A (en) * 2021-09-02 2021-11-30 长江存储科技有限责任公司 Fault detection method and device for semiconductor process
CN113721121B (en) * 2021-09-02 2024-04-19 长江存储科技有限责任公司 Fault detection method and device for semiconductor process

Similar Documents

Publication Publication Date Title
CN112651176A (en) GA-BP algorithm-based shielded door motor output force prediction method
CN106600140A (en) Improved support vector machine-based gas pipeline fault prediction and early warning system and method
CN103714382B (en) A kind of multi-index comprehensive evaluation method for reliability of urban rail train security detection sensor network
CN113723010A (en) Bridge damage early warning method based on LSTM temperature-displacement correlation model
CN109086926B (en) Short-time rail transit passenger flow prediction method based on combined neural network structure
CN113487855B (en) Traffic flow prediction method based on EMD-GAN neural network structure
CN114202316A (en) Urban rail transit train schedule optimization method based on deep reinforcement learning
CN111063204A (en) Expressway vehicle speed prediction model training method based on toll station flow
CN112734094B (en) Intelligent city intelligent rail vehicle fault gene prediction method and system
CN112232561A (en) Power load probability prediction method based on constrained parallel LSTM quantile regression
CN111680413B (en) Tramcar timing energy-saving operation optimization method and system based on double-layer algorithm
Zhang et al. An AI based high-speed railway automatic train operation system analysis and design
CN112394729A (en) T-S fuzzy modeling and fuzzy tracking control method for motor train unit operation process
CN111353640A (en) Method for constructing wind speed prediction model by combination method
CN113326962B (en) Subway train ATO speed curve prediction method based on BP neural network
CN115169642A (en) IPSO-GRU network model-based train derailment coefficient prediction method
CN112967420B (en) Interval two-type based heavy-duty train operation process monitoring method and system
CN113312722B (en) Reliability prediction optimization method for key parts of urban rail train
CN114881204A (en) Road flatness prediction method based on road element splitting and GA-BP neural network model
Guo et al. An enhanced fault detection method for railway turnouts incorporating prior faulty information
CN111898195B (en) Rail transit train system polymorphism reliability analysis method based on improved d-MC
CN113255963A (en) Road surface use performance prediction method based on road element splitting and deep learning model LSTM
CN112685956A (en) Road feel simulation method based on CART regression tree
Huang et al. Short-term traffic flow parameters prediction based on multi-scale analysis and artificial neural network
CN117455073B (en) Prediction method for passenger evacuation efficiency in carriage type transportation means under flood scene

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20210413