CN111539175A - Railway electromagnetic compatibility fault prediction method - Google Patents
Railway electromagnetic compatibility fault prediction method Download PDFInfo
- Publication number
- CN111539175A CN111539175A CN202010558739.XA CN202010558739A CN111539175A CN 111539175 A CN111539175 A CN 111539175A CN 202010558739 A CN202010558739 A CN 202010558739A CN 111539175 A CN111539175 A CN 111539175A
- Authority
- CN
- China
- Prior art keywords
- output
- hidden layer
- layer
- parameter
- input
- 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
- 238000000034 method Methods 0.000 title claims abstract description 31
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 13
- 230000002068 genetic effect Effects 0.000 claims abstract description 7
- 238000012549 training Methods 0.000 claims description 22
- 238000010606 normalization Methods 0.000 claims description 13
- 230000005540 biological transmission Effects 0.000 claims description 6
- 238000013507 mapping Methods 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 6
- 238000002474 experimental method Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 2
- 238000012545 processing Methods 0.000 claims 1
- 238000013528 artificial neural network Methods 0.000 description 10
- 238000004458 analytical method Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 6
- 238000013178 mathematical model Methods 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 230000005672 electromagnetic field Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 230000003014 reinforcing effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/30—Circuit design
- G06F30/38—Circuit design at the mixed level of analogue and digital signals
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/30—Circuit design
- G06F30/32—Circuit design at the digital level
- G06F30/33—Design verification, e.g. functional simulation or model checking
- G06F30/3308—Design verification, e.g. functional simulation or model checking using simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/30—Circuit design
- G06F30/32—Circuit design at the digital level
- G06F30/337—Design optimisation
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
- Y04S10/52—Outage or fault management, e.g. fault detection or location
Landscapes
- Engineering & Computer Science (AREA)
- Computer Hardware Design (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a railway electromagnetic compatibility fault prediction method, which comprises two steps of defining related parameters and constructing an electromagnetic compatibility network prediction model; firstly, setting appropriate use parameters, then constructing an electromagnetic compatible network prediction model, comprehensively setting the network prediction model by adjusting input parameters, intermediate parameters and output parameters, and finally adjusting the weight value by a genetic algorithm.
Description
Technical Field
The invention relates to the field of railway electromagnetic compatibility, in particular to a railway electromagnetic compatibility fault prediction method.
Background
The prediction of electromagnetic compatibility is a necessary step for realizing the electromagnetic compatibility of the electronic equipment or system, and is also a main basis for designing the electromagnetic compatibility. The basic idea is to establish a corresponding prediction mathematical model according to the existing experiments, experiences and theoretical analysis aiming at three factors of electromagnetic interference, then carry out simulation calculation by using a proper electromagnetic field numerical method to obtain various calculation results of potential electromagnetic interference so as to make scientific and reasonable prediction and judgment on whether the electromagnetic energy emitted by an interference source can influence the normal work of sensitive equipment. It is necessary to research a fault detection method for railway electromagnetic compatibility.
Disclosure of Invention
In view of the above, the present invention provides a method for predicting a fault of electromagnetic compatibility of a railway, which solves or partially solves the above-mentioned problems.
In order to achieve the effect of the technical scheme, the technical scheme of the invention is as follows: a railway electromagnetic compatibility fault prediction method comprises the following steps:
firstly, defining parameters related to the method; firstly, let the interference power output by the interference source be PoutTransmission loss of interference source is PsThe effective power generated by the interference source on the sensitive equipment is PeThe three satisfy the relation Pe=Pout-PsP is interference power; interference power P output by interference sourceoutTransmission loss P of interference sourcesThe relevant parameter is used as the input parameter set | x1,x2,...,xnI effective power P generated by interference source on sensitive equipmenteTo aIs the output parameter set y1,y2,...,ym|, wherein ,x1,x2,...,xnFor specific parameters included in the input parameter set, x is the input parameter set, y1,y2,...,ymThe parameters are specific parameters contained in an output group, y is an output parameter group, n is the number of input parameter groups, m is the number of output parameter groups, and the parameters are positive integers and can be freely set according to actual conditions, the input parameter groups correspond to the specific parameters contained in the output parameter groups one by one, and the past experiment results are input as training samples; transmission loss P of interference sourcesParameter of interest, effective power P generated by interference source on sensitive equipmenteParameter of interest, effective power P generated by interference source on sensitive equipmenteThe relevant parameters may be freely specified by the user;
secondly, constructing an electromagnetic compatible network prediction model; the input of the network prediction model is an input parameter set | x1,x2,...,xnThe output of the network prediction model is an output parameter set y1,y2,...,ymThe network prediction model sequentially comprises an input layer, a hidden layer, a dynamic regulation layer and an output layer, wherein the input layer is used for receiving an input parameter group, and the output layer is used for receiving a value of an output parameter group;
b) preprocessing the training sample, firstly normalizing, mapping the data of the training sample into a designated interval range by a normalization formula, wherein the designated interval range can be designated with a normalization coefficient, namely the designated interval range is the normalization coefficient multiplied by a unit interval, and the unit interval is initially [ -1,1]When the training samples are set, the unit interval is equal to (a-a)min)/(amax-amin) Wherein a is data of training sample, aminAs a minimum value of the training sample, amaxIs the maximum value of the training sample;
the normalization formula is expressed asWhere k is a normalization coefficient and the value range is real numbers greater than 0(ii) a The normalization formula is reversely pushed, so that data of the training sample can be obtained;
setting the threshold value of the hidden layer of the j-th layer to be omegajω is the threshold of the hidden layer, ωjThe value range is (-omega)max,ωmax), wherein ,ωmaxIs the extreme upper limit of the threshold, and the weight is χijThe adjusting function of the dynamic adjusting layer is d (r), and r is an adjusting parameter and takes a positive real number;
wherein ,ΔωjThreshold omega for hidden layerjIs not greater than the threshold value omegajThe real number of (2); subscript i is the index of the input parameter set, the index of the input parameter set as the input of the network prediction model, and the index is the second group input of the network prediction model; subscript j is a label of the output of the hidden layer, indicates the number of layers where the hidden layer is located, and is a positive integer;
Δωjself-adjusting according to the error norm of the fitness function, wherein the fitness function is defined as the reciprocal of the error norm, and the error norm is defined as | | qj-oj||2,qjIs the expected output of the jth hidden layer, ojThe prediction output of the j-th hidden layer is obtained;
the adjustment function d (r) of the dynamic adjustment layer is defined as
wherein , is Δ ωjH is the set of all possible outputs of the hidden layer, hjA set of outputs representing a jth layer hidden layer; omega is all possible sets of thresholds for the hidden layer,representing that 2 values are respectively selected from all possible sets omega of the threshold value of the hidden layer and all possible output sets h of the hidden layer and multiplied, wherein a parameter n in the formula represents that the number of the output of the hidden layer is consistent with the number of the input parameter sets and is equal to n;representing the multiplication of the values of r selected from the set of all possible thresholds omega of the hidden layer and the set of all possible outputs h of the hidden layer,the method comprises the steps of sequentially adding products obtained by respectively selecting and multiplying 2 numerical values from all possible sets of threshold values of the hidden layer and all possible output sets of the hidden layer to products obtained by respectively selecting and multiplying 3 numerical values from all possible sets of threshold values of the hidden layer and all possible output sets of the hidden layer, and sequentially accumulating the products until products obtained by respectively selecting and multiplying r numerical values from all possible sets of threshold values of the hidden layer and all possible output sets of the hidden layer are added; the output of the output layer takes the adjusting parameter r in the adjusting function d (r) as a free variable, increases from 1 to m, and takes the increasing value as an output parameter group, wherein the specific process is the output parameter group | y1,y2,...,ymY in |)1D (1), …, in that order, ymD (m); calculating the mean square error between the actual output and the expected output of the training sample, and if the mean square error does not reach the expected value, namely the value of the mean square error is gradually reduced, using a genetic algorithm to reduce the weight χij(ii) a If the mean square error does not reach the expected value, i.e. the value thereof gradually increases, the weight χ is increased by using the genetic algorithmij。
The beneficial results of the invention are as follows: the invention provides a railway electromagnetic compatibility fault prediction method, which comprises the steps of firstly setting appropriate use parameters, then constructing an electromagnetic compatibility network prediction model, comprehensively setting the network prediction model by adjusting input parameters, intermediate parameters and output parameters in the electromagnetic compatibility network prediction model, and finally adjusting weight values by a genetic algorithm.
Detailed Description
In order to make the technical problems, technical solutions and advantageous effects to be solved by the present invention more apparent, the present invention is described in detail below with reference to the embodiments. It should be noted that the specific embodiments described herein are only for illustrating the present invention and are not to be construed as limiting the present invention, and products that can achieve the same functions are included in the scope of the present invention. The specific method comprises the following steps:
example (b): from the last 70 s, the electromagnetic compatibility prediction theory has been studied deeply domestically and abroad, and various analysis prediction models, such as a source model, a sensor model, a coupling model, an inter-system analysis model and an intra-system analysis model, are established. And a plurality of electromagnetic field numerical methods for solving the prediction model are provided, such as a moment method, a finite element method, a time domain finite difference method and the like. However, currently, many difficulties are faced in predicting; no mathematical model suitable for general prediction exists so far, so that the application of a prediction model is greatly limited; most predictive models are an approximation and simplified representation under ideal conditions; the solution process of the prediction model of the complex system is very complex. Even if the results are obtained under a large number of constraints and simplifications, there are large errors, which worsen the prediction accuracy. Therefore, the search for more reasonable prediction and analysis methods is always a hot spot studied by researchers in the field. The artificial neural network is an important component of modern nonlinear science, can realize nonlinear mapping of different dimensional spaces, and is widely applied in many fields in recent years. The method is tried to be applied to the prediction problem of electromagnetic compatibility by a writer, a network is adopted to construct a mapping relation between an input prediction factor and disturbance response of sensitive equipment, an electromagnetic field numerical calculation method is adopted to establish a training sample set and a test sample set, and a constructed neural network is trained to be used for rapidly predicting and evaluating the electromagnetic compatibility problem.
At present, the EMC theory is deeply researched at home and abroad, various analysis and prediction models such as a source model, a sensitive period model, a coupling model, an inter-system/intra-system analysis model and the like are established, and a plurality of numerical methods for solving the models such as a moment method, a finite element method, a time domain finite difference method and the like are provided. However, the current EMC prediction faces 3 drawbacks:
a mathematical model suitable for universal EMC prediction is not proposed up to now;
secondly, most prediction models are only an approximation under ideal conditions;
solving the prediction model of the complex system is very complex. Therefore, the result of the solution has a large error. Considering that Neural Network (NN) is an emerging discipline, it does not relate to the complex model of the original problem, but realizes the mapping of input and output data pairs. Therefore, a BP neural network (BPNN) is often employed to predict EMC problems. Considering that BP is easy to converge to a local optimal point, a particle swarm algorithm is adopted to realize weight optimization.
The EMC prediction model is as follows:
assuming that PT represents the interference power output by the interference source, LP represents the transmission loss of the interference signal, and PI represents the effective interference power generated by the interference source on the sensitive device, the mathematical model of EMC can be expressed as PI PT-LP. The interference source PT and the transmission loss LP can be used as the input of NN, the interference result PI is used as the output of the network, and the known actual measurement result is used as the training sample, so that the mapping from the interference environment to the interference response is realized. Since the wire is both an efficient electromagnetic interference receiving antenna and an efficient electromagnetic interference radiating antenna, it is a main reason for hindering EMC, and attracts a great deal of study of scholars.
In the prediction technology, a BP neural network technology and an Adaboost algorithm are often used; the BP neural network is a supervised learning multilayer feedforward neural network and is mainly characterized by signal forward transmission and error backward propagation. In the signal forward transmission process, an input signal enters from an input layer, is processed by a hidden layer and reaches an output layer. The neuron state of each layer only affects the neuron state of the next layer. And judging whether the result of the output layer is expected output, if not, turning to reverse propagation, and then adjusting the network weight and the threshold according to the prediction error so as to enable the prediction output of the BP neural network to continuously approach the expected output.
Adaboost is an iterative algorithm, and obtains a strong learning algorithm by reinforcing a weak learning algorithm, namely, a strong classifier with ideal classification capability is constructed by a weak classifier set containing key features. The Adaboost algorithm has the advantages that training data selected after weighting is used for replacing training samples selected randomly, weak classifiers are combined, and a weighted voting mechanism is used for replacing an average voting mechanism.
The beneficial results of the invention are as follows: the invention provides a railway electromagnetic compatibility fault detection method, which comprises the steps of firstly setting appropriate use parameters, then constructing an electromagnetic compatibility network prediction model, comprehensively setting the network prediction model by adjusting input parameters, intermediate parameters and output parameters in the electromagnetic compatibility network prediction model, and finally adjusting weight values by a genetic algorithm.
It is further noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The above description is only for the preferred embodiment of the present invention, and should not be used to limit the scope of the claims of the present invention. While the foregoing description will be understood and appreciated by those skilled in the relevant art, other equivalents may be made thereto without departing from the scope of the claims.
Claims (1)
1. A railway electromagnetic compatibility fault prediction method is characterized by comprising the following steps:
firstly, defining parameters related to the method; firstly, let the interference power output by the interference source be PoutTransmission loss of interference source is PsThe effective power generated by the interference source on the sensitive equipment is PeThe three satisfy the relation Pe=Pout-PsP is interference power; interference power P output by the interference sourceoutTransmission loss P of the interference sourcesThe relevant parameter is used as the input parameter set | x1,x2,...,xnEffective power P generated by the interference source on the sensitive equipmenteThe parameter concerned is the output parameter set y1,y2,...,ym|, wherein ,x1,x2,...,xnFor specific parameters included in the input parameter set, x is the input parameter set, y1,y2,...,ymThe input parameter groups are specific parameters contained in the output group, y is an output parameter group, n is the number of the input parameter groups, m is the number of the output parameter groups, the input parameter groups are positive integers and can be freely set according to actual conditions, the input parameter groups correspond to the specific parameters contained in the output parameter groups one by one, and the previous experiment results are input as training samples; transmission loss P of the interference sourcesParameter of interest, effective power P generated by said interference source on a sensitive deviceeParameter of interest, effective power P generated by said interference source on a sensitive deviceeThe relevant parameters may be freely specified by the user;
secondly, constructing an electromagnetic compatible network prediction model; the input of the network prediction model is the input parameter set | x1,x2,...,xnThe output of the network prediction model is the output parameter set y1,y2,...,ymI, the networkThe prediction model comprises an input layer, a hidden layer, a dynamic regulation layer and an output layer in sequence, wherein the input layer is used for receiving the input parameter group, and the output layer is used for receiving the value of the output parameter group;
a) preprocessing the training sample, firstly, performing normalization processing, mapping the data of the training sample into a specified interval range by a normalization formula, wherein the specified interval range can be specified with a normalization coefficient, namely the specified interval range is the result of multiplying the normalization coefficient by a unit interval, and the unit interval is initially [ -1,1 [ -1]When the training samples are set, the unit interval is equal to (a-a)min)/(amax-amin) Wherein a is the data of the training sample, aminIs the minimum value of the training sample, amaxIs the maximum value of the training sample;
the normalization formula is expressed asK is a normalization coefficient, and the value range of k is real numbers larger than 0; the normalization formula is reversely pushed, so that the data of the training sample can be obtained;
setting the threshold value of the hidden layer at the j-th layer to be omegajω is the threshold of the hidden layer, ωjThe value range is (-omega)max,ωmax), wherein ,ωmaxIs the extreme upper limit of the threshold, and the weight is χijThe adjusting function of the dynamic adjusting layer is d (r), and r is an adjusting parameter and takes a positive real number;
wherein ,ΔωjA threshold value ω for the hidden layerjIs not greater than the threshold value omegajThe real number of (2); subscript i is the index of the set of input parameters, the index of the input parameters being the inputs of the network prediction model, and the number of the set of inputs; the subscript j isThe output labels of the hidden layers represent the number of layers where the hidden layers are located, and are positive integers;
the Δ ωjSelf-adjusting according to an error norm of a fitness function, the fitness function being defined as a reciprocal of the error norm, the error norm being defined as | | qj-oj||2,qjFor the expected output of the hidden layer at layer j, ojA prediction output of the hidden layer for the j-th layer;
the adjustment function d (r) of the dynamic adjustment layer is defined as
wherein , is Δ ωjH is the set of all possible outputs of the hidden layer, hjA set of outputs representing a jth layer hidden layer; omega is all possible sets of thresholds for the hidden layer,representing that 2 values are respectively selected from all possible sets omega of the threshold value of the hidden layer and all possible output sets h of the hidden layer and multiplied, wherein a parameter n in the formula represents that the number of the output of the hidden layer is consistent with the number of the input parameter sets and is equal to n;representing the multiplication of the values of r selected from the set of all possible thresholds omega of the hidden layer and the set of all possible outputs h of the hidden layer,representing the sum of products obtained by sequentially selecting and multiplying 2 values from all possible sets of the threshold value of the hidden layer and all possible output sets of the hidden layer and 3 values from all possible sets of the threshold value of the hidden layer and all possible output sets of the hidden layer, and sequentially accumulating the products until products obtained by selecting and multiplying r values from all possible sets of the threshold value of the hidden layer and all possible output sets of the hidden layer are added; the output of the output layer takes the adjusting parameter r in the adjusting function d (r) as a free variable, and increases from 1 to m as the output parameter group, specifically, the process is the output parameter group | y1,y2,...,ymY in |)1D (1), …, in that order, ymD (m); calculating the mean square error between the actual output and the expected output of the training sample, and if the mean square error does not reach the expected value, namely the value of the mean square error is gradually reduced, using a genetic algorithm to reduce the weight χij(ii) a If the mean square error does not reach the desired value, i.e. its value gradually increases, then the weight χ is increased using a genetic algorithmij。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010558739.XA CN111539175B (en) | 2020-06-18 | 2020-06-18 | Railway electromagnetic compatibility fault prediction method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010558739.XA CN111539175B (en) | 2020-06-18 | 2020-06-18 | Railway electromagnetic compatibility fault prediction method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111539175A true CN111539175A (en) | 2020-08-14 |
CN111539175B CN111539175B (en) | 2023-06-02 |
Family
ID=71976508
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010558739.XA Active CN111539175B (en) | 2020-06-18 | 2020-06-18 | Railway electromagnetic compatibility fault prediction method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111539175B (en) |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101158588A (en) * | 2007-11-16 | 2008-04-09 | 北京航空航天大学 | MEMS gyroscopes error compensation method for micro satellite based on integration nerval net |
CN104636985A (en) * | 2013-11-08 | 2015-05-20 | 国家电网公司 | Method for predicting radio disturbance of electric transmission line by using improved BP (back propagation) neural network |
CN106874581A (en) * | 2016-12-30 | 2017-06-20 | 浙江大学 | A kind of energy consumption of air conditioning system in buildings Forecasting Methodology based on BP neural network model |
CN109508488A (en) * | 2018-11-07 | 2019-03-22 | 西北工业大学 | Contour peening technological parameter prediction technique based on genetic algorithm optimization BP neural network |
CN110429889A (en) * | 2019-08-07 | 2019-11-08 | 北京航空航天大学 | A kind of adjustable Square wave injection torque capacity electric current of amplitude compares motor control method |
AU2020100051A4 (en) * | 2020-01-10 | 2020-02-13 | Jin, Yawen Miss | Method of the mission planning for the communication between the small satellite clusters |
CN111126560A (en) * | 2019-11-07 | 2020-05-08 | 云南民族大学 | Method for optimizing BP neural network based on cloud genetic algorithm |
-
2020
- 2020-06-18 CN CN202010558739.XA patent/CN111539175B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101158588A (en) * | 2007-11-16 | 2008-04-09 | 北京航空航天大学 | MEMS gyroscopes error compensation method for micro satellite based on integration nerval net |
CN104636985A (en) * | 2013-11-08 | 2015-05-20 | 国家电网公司 | Method for predicting radio disturbance of electric transmission line by using improved BP (back propagation) neural network |
CN106874581A (en) * | 2016-12-30 | 2017-06-20 | 浙江大学 | A kind of energy consumption of air conditioning system in buildings Forecasting Methodology based on BP neural network model |
CN109508488A (en) * | 2018-11-07 | 2019-03-22 | 西北工业大学 | Contour peening technological parameter prediction technique based on genetic algorithm optimization BP neural network |
CN110429889A (en) * | 2019-08-07 | 2019-11-08 | 北京航空航天大学 | A kind of adjustable Square wave injection torque capacity electric current of amplitude compares motor control method |
CN111126560A (en) * | 2019-11-07 | 2020-05-08 | 云南民族大学 | Method for optimizing BP neural network based on cloud genetic algorithm |
AU2020100051A4 (en) * | 2020-01-10 | 2020-02-13 | Jin, Yawen Miss | Method of the mission planning for the communication between the small satellite clusters |
Non-Patent Citations (5)
Title |
---|
[1] 王迪: "基于自适应遗传算法的智能桁杆件架多目标优化设计" * |
JIN,DA-WEI. LU,JUN.: "An improved determination approach to the structure and parameters of dynamic structure-based neural networks" * |
李晓峰: "基于数据挖掘的城市轨道交通车站突发客流预测研究" * |
李永明,祝言菊,李旭等: "电磁兼容的人工神经网络预测技术分析" * |
马昭钰: "三相异步电动机基本参数自动识别方法的研究" * |
Also Published As
Publication number | Publication date |
---|---|
CN111539175B (en) | 2023-06-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Wei et al. | Some dependent aggregation operators with 2-tuple linguistic information and their application to multiple attribute group decision making | |
CN110334875B (en) | Wind power combination probability prediction method considering evaluation index conflict | |
CN107392368B (en) | Meteorological forecast-based office building dynamic heat load combined prediction method | |
An et al. | Fuzzy rough regression with application to wind speed prediction | |
CN108445752B (en) | Random weight neural network integrated modeling method for self-adaptively selecting depth features | |
CN111046187B (en) | Sample knowledge graph relation learning method and system based on confrontation type attention mechanism | |
CN110286586A (en) | A kind of MR damper hybrid modeling method | |
Yang et al. | Synchronization for fractional-order reaction–diffusion competitive neural networks with leakage and discrete delays | |
CN109242212A (en) | A kind of wind-powered electricity generation prediction technique based on change Mode Decomposition and length memory network | |
Su et al. | Prediction model of hot metal temperature for blast furnace based on improved multi-layer extreme learning machine | |
CN113095598A (en) | Multi-energy load prediction method, system, device and medium | |
CN116526450A (en) | Error compensation-based two-stage short-term power load combination prediction method | |
Abdulsalam et al. | Electrical energy demand forecasting model using artificial neural network: A case study of Lagos State Nigeria | |
Jaipuria et al. | A study on behaviour of bullwhip effect in (R, S) inventory control system considering DWT-MGGP demand forecasting model | |
CN103020452B (en) | A kind of Radar emitter threat level determination methods | |
Wahyudi et al. | Backpropagation network optimization using one step secant (oss) algorithm | |
Zhang et al. | Blind source separation based on quantum slime mould algorithm in impulse noise | |
CN117134346A (en) | Short-term load prediction method and device based on VMD-GA-PSO-LSTM model | |
CN111539175A (en) | Railway electromagnetic compatibility fault prediction method | |
Wang et al. | The research and application of tax risk identification based on pso-bp algorithm | |
CN109829574A (en) | Methods of electric load forecasting based on ACDE and BP neural network | |
CN112653751B (en) | Distributed intrusion detection method based on multilayer extreme learning machine in Internet of things environment | |
CN109474258A (en) | The Optimization Method of Kernel Parameter of random Fourier feature core LMS based on nuclear polarization strategy | |
CN114118163B (en) | Optical fiber intrusion detection method based on machine learning | |
Hu et al. | Improved BP neural network based on genetic algorithm |
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 |