CN112734097A - Unmanned train energy consumption prediction method, system and storage medium - Google Patents
Unmanned train energy consumption prediction method, system and storage medium Download PDFInfo
- Publication number
- CN112734097A CN112734097A CN202011639921.4A CN202011639921A CN112734097A CN 112734097 A CN112734097 A CN 112734097A CN 202011639921 A CN202011639921 A CN 202011639921A CN 112734097 A CN112734097 A CN 112734097A
- Authority
- CN
- China
- Prior art keywords
- energy consumption
- wolf
- train
- value
- neural network
- 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
- 238000005265 energy consumption Methods 0.000 title claims abstract description 175
- 238000000034 method Methods 0.000 title claims abstract description 56
- 238000003860 storage Methods 0.000 title claims abstract description 8
- 230000008569 process Effects 0.000 claims abstract description 28
- 239000002245 particle Substances 0.000 claims description 125
- 241000282461 Canis lupus Species 0.000 claims description 113
- 238000013528 artificial neural network Methods 0.000 claims description 73
- 238000004422 calculation algorithm Methods 0.000 claims description 58
- 238000012549 training Methods 0.000 claims description 36
- 238000005457 optimization Methods 0.000 claims description 31
- 239000013598 vector Substances 0.000 claims description 22
- 238000004364 calculation method Methods 0.000 claims description 12
- 238000010845 search algorithm Methods 0.000 claims description 10
- 230000007613 environmental effect Effects 0.000 claims description 8
- 238000011156 evaluation Methods 0.000 claims description 8
- 230000001133 acceleration Effects 0.000 claims description 7
- 239000011159 matrix material Substances 0.000 claims description 6
- 230000002028 premature Effects 0.000 claims description 5
- 238000012545 processing Methods 0.000 claims description 5
- 230000005484 gravity Effects 0.000 claims description 4
- 241000894007 species Species 0.000 claims 1
- 230000006870 function Effects 0.000 description 11
- 230000002068 genetic effect Effects 0.000 description 11
- 238000012360 testing method Methods 0.000 description 7
- 210000000349 chromosome Anatomy 0.000 description 5
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 5
- 230000009194 climbing Effects 0.000 description 4
- 238000011161 development Methods 0.000 description 4
- 238000012795 verification Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 3
- 238000004140 cleaning Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000004927 fusion Effects 0.000 description 3
- 238000013507 mapping Methods 0.000 description 3
- 230000035772 mutation Effects 0.000 description 3
- 241000282421 Canidae Species 0.000 description 2
- 241000208422 Rhododendron Species 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 230000002457 bidirectional effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000003071 parasitic effect Effects 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000033228 biological regulation Effects 0.000 description 1
- 235000005770 birds nest Nutrition 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013524 data verification Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000013486 operation strategy Methods 0.000 description 1
- 230000035515 penetration Effects 0.000 description 1
- 238000012827 research and development Methods 0.000 description 1
- 230000008054 signal transmission Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
- 238000009423 ventilation Methods 0.000 description 1
- 235000005765 wild carrot Nutrition 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- 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
- G06N3/086—Learning methods using evolutionary algorithms, e.g. genetic algorithms or genetic programming
Abstract
The invention discloses a method, a system and a storage medium for predicting the energy consumption of an unmanned train, which are used for realizing the minimum traction energy consumption prediction of the unmanned train with high reliability and high precision by fusing various collected parameters such as train operation data, in-train and station passenger data, out-train environment data and the like on the basis of ensuring the requirements of safety, comfort, timekeeping and the like in the operation process of the unmanned train.
Description
Technical Field
The invention relates to the field of artificial intelligence, in particular to a method and a system for predicting energy consumption of an unmanned train and a storage medium.
Background
With continuous breakthrough and innovation of rail transit technology, the automation level of rail vehicles is gradually improved. Under the large background that the problem of energy consumption is increasingly serious, the development of high-efficiency energy-saving technology and operation technology is urgently needed in the application of novel unmanned trains. Meanwhile, a high-efficiency and environment-friendly rail transit system is developed to actively deal with the energy crisis and meet the development requirement of building a green city smart city in the future. If the energy consumption of the rail transportation department is too high, not only can huge pressure be brought to national energy supply, but also the economic benefits of enterprises are limited, and the sustainable development is influenced.
In the railway department, the energy consumption of the transportation department accounts for more than 80% of the energy consumption of the railway, and the train traction power consumption is the main mode of the energy consumption of the rail transit system. Therefore, the research on the effective energy-saving train control method has important significance for reducing the energy consumption of the railway, and simultaneously promotes the improvement of the parking precision and the train punctuality, thereby realizing the continuous improvement of the automatic train control system. The method has important significance for reducing railway transportation cost, improving the efficiency of the railway transportation industry and realizing the sustainable development of railways.
The running process of the unmanned train between stations needs to select proper working conditions, and factors of running energy consumption caused by passenger flow change and extra energy consumption caused by environment change outside the train need to be considered besides energy consumption of normal running of the train on a fixed track. Therefore, the number of the operation conditions can be flexibly selected according to the operation experience of big data and the line condition during the operation of the unmanned train. The CN102360401B patent proposes an energy-saving scheduling method for urban rail transit based on genetic algorithm, and CN106740998A proposes a method for realizing energy-saving operation by finding speed switching points through vehicle-mounted ATO, and although energy-saving is realized by adjusting stop time or switching points, influence factors inside and outside urban rail transit vehicles and the condition of operation on the whole line are ignored.
Disclosure of Invention
The invention aims to solve the technical problem that the prior art is insufficient, and provides a method, a system and a storage medium for predicting the energy consumption of an unmanned train, which are used for realizing the minimum traction energy consumption prediction of the unmanned train with high reliability and high precision by fusing various collected parameters such as train operation data, passenger data in and at a station, environment data outside the train and the like on the basis of ensuring the requirements of safety, comfort, timekeeping and the like in the operation process of the unmanned train.
In order to solve the technical problems, the technical scheme adopted by the invention is as follows: a method for predicting energy consumption of an unmanned train comprises the following steps:
1) collecting the operation data of the unmanned train, the passenger data in the train and at the station and the environment data outside the train; the unmanned train operation data comprises a stable running speed energy consumption value, a running distance and road gradient loss power of the train in a specified time interval in the running process; the passenger data in the train and at the station comprises passenger flow, riding comfort evaluation indexes, consignment luggage weight and in-train seat-up rate; the external environment data of the train comprises rainfall resistance, road ponding resistance, wind resistance energy consumption, accumulated snow energy consumption and temperature energy consumption in a specified time interval;
2) the unmanned train operation data is used as input of a DBN deep confidence neural network, the DBN deep confidence neural network is trained, and an unmanned train vehicle running energy consumption control prediction model is obtained; taking the data of passengers in the train and in the station as the input of a BILSTM deep neural network, training the BILSTM deep neural network, and obtaining a passenger flow energy consumption prediction model of the unmanned train; taking the external environment data of the train as the input of a GRU deep neural network, training the GRU deep neural network, and obtaining an unmanned train environment energy consumption prediction model;
3) fusing the unmanned train vehicle running energy consumption control prediction model, the unmanned train passenger flow energy consumption prediction model and the unmanned train environment energy consumption prediction model to obtain an energy consumption prediction model;
4) and inputting the unmanned train operation data, the train inside and station passenger data and the train outside environment data which are acquired in real time into the energy consumption prediction model to predict the train energy consumption.
The steps fully consider the influence factors of various human-computer loop combinations during the operation of the unmanned train, and fuse the collected parameters to realize the prediction and optimization process of the lowest traction energy consumption with high reliability and high precision, so that the energy-saving optimization problem of the unmanned train can be effectively solved.
In step 2), the specific training process of the unmanned train vehicle running energy consumption control prediction model comprises the following steps: and searching the optimal weight and the threshold of the DBN deep belief neural network by using a gray wolf optimization algorithm, wherein the DBN deep belief neural network corresponding to the optimal weight and the threshold is the vehicle running energy consumption control prediction model of the unmanned train.
The model mainly predicts different historical data in the energy consumption of the unmanned train vehicle in running, the DBN deep confidence neural network can enable the whole neural network to generate training data according to the maximum probability by training the weight among the neurons, and meanwhile, the data characteristics can be effectively extracted and the prediction precision can be improved by combining the application of an optimization algorithm.
The specific implementation process of finding the optimal weight and the threshold of the DBN deep belief neural network by utilizing the gray wolf optimization algorithm comprises the following steps:
a1, randomly placing the position of each gray wolf in wolf clusters in a solution space, setting the population quantity of the gray wolfs as m, the characteristic quantity of the original data set of the vehicle running energy consumption as d, and the position matrix of the wolf clusters as oneA m x d dimensional binary matrix, wherein the location of the ith gray wolf is designated as CGi＝(CGi1,CGi2,...CGid)，CGi1,CGi2,...CGidCoding each dimension of each gray wolf in the wolf group; processing an input data set of the DBN deep confidence neural network according to the position information in the wolf individuals to obtain a new data set, calculating and evaluating the fitness value fitness of each wolf individual, and taking the maximum iteration number iter of the individual wolfmaxThe position obtained by the secondary iteration is used as a local optimal solution, and the local optimal solution is the weight and the threshold of the DBN neural network;
a2, finding and recording the positions X of alpha wolf, beta wolf and delta wolf according to the fitness value from big to small in sequenceα、Xβ、Xδ；
A3, disturbing the positions of the alpha wolf, the beta wolf and the delta wolf, comparing the disturbed positions of the alpha wolf, the beta wolf and the delta wolf with the original positions of the alpha wolf, the beta wolf and the delta wolf, reserving the position closer to the optimal solution, and gradually approaching the optimal solution by utilizing the subsequent iteration process;
a4, for each gray wolf individual in the wolf group, calculating the positions of the alpha wolf, the beta wolf and the delta wolf by using a gray wolf algorithm, and obtaining the position of the updated gray wolf individual closer to the optimal solution by using the following formula:
wherein ,
represents the position value of the ith dimension of the ith Hui wolf individual in the process of the (t +1) th iteration, Xα、Xβ、XδRespectively, the positions of alpha wolf, beta wolf and delta wolf, random is [0, 1]]BetweenCounting;
a5, randomly updating the position of the wolf population according to a preset probability value Pa, calculating the updated probability r of finding the wolf prey, and if r is greater than Pa, finding the wolf prey and updating the position of the wolf individual; according to the fact that the updated position and the original position are evaluated to be good or bad, the position closer to the optimal solution is reserved, and the step A6 is carried out; otherwise, go directly to step A6;
a6, judging whether the iteration times reach the maximum iteration times, if so, stopping the iteration, and outputting a global optimum value obtained according to the alpha wolf position and the fitness value fitness thereof; otherwise, returning to step A2 until a global optimum is found; the global optimal values are the optimal weight and the threshold of the DBN deep confidence neural network.
The inherent defects of low learning speed and local minimum point of the artificial neural network are effectively overcome. The position of the population is updated in the iterative process, so that the algorithm is prevented from being partially optimized, and the optimization precision and the convergence speed of the algorithm are improved.
In the step 2), the process of obtaining the passenger flow energy consumption prediction model of the unmanned train comprises the following steps: and searching the optimal weight and threshold of the BILSTM deep neural network by using a quantum particle group algorithm, wherein the BILSTM deep neural network corresponding to the optimal weight and threshold is the passenger flow energy consumption prediction model of the unmanned train.
The model mainly predicts different historical data in passenger flow energy consumption of unmanned train vehicles, and the BILSTM deep neural network can better capture bidirectional transmitted information compared with a unidirectional LSTM neural network, so that the effectiveness of passenger flow data feature extraction is enhanced, and the prediction precision is further improved by combining with the application of an optimization algorithm.
The specific implementation process of finding the optimal weight and threshold of the BILSTM deep neural network by using the quantum particle swarm algorithm comprises the following steps:
b1, taking the position vector of each quantum particle individual in the quantum particle swarm as the weight and the threshold of the BILSTM deep neural network, and initializing the position vector parameters of the quantum particle swarm individual into random numbers of [ -1,1 ]; initializing input parameters of a quantum particle swarm algorithm;
b2, setting a fitness function, and determining an initial optimal quantum particle individual position vector and iteration times; substituting the weight and the threshold corresponding to the quantum particle individual position vector into the passenger flow energy consumption prediction model of the unmanned train based on the BILSTM deep network, determining the type of the identification vector label by using the passenger flow energy consumption prediction model of the unmanned train based on the BILSTM deep network determined by the quantum particle individual position vector, and determining the weight calculation results of passenger flow, riding comfort evaluation index, consignment luggage weight and in-train boarding rate in the train and station passenger data by using the quantum particle individual position;
b3, if the variance of the population fitness of the quantum particle swarm is smaller than the premature convergence judgment threshold, carrying out variation on the particles with the worst fitness and the extreme particles of the swarm in the quantum particle swarm, and taking the particles with the best fitness as the global optimal quantum particle individuals, namely the extreme values of the swarm;
b4, when the iteration times are larger than the iteration times of the elite population, extracting extreme values of each population through information sharing among the populations to construct the elite population, and turning to the step B6, otherwise, turning to the step B6;
b5, updating the position parameters of each particle swarm by comparing the individual extreme value with the global extreme value, namely recalculating and comparing the fitness value of each particle, and updating the individual extreme value if the fitness value is superior to the current individual extreme value; comparing the global extreme value particles, if the particle fitness value is superior to the current population extreme value, updating the global extreme value particles, adding 1 to the iteration number, and turning to the step B3; the global extreme example is an extreme value obtained by evolution in all particle individuals including an elite population;
b6, continuing iterative evolution of the elite population according to the steps B3-B5, judging whether the maximum iteration times are met, if so, ending, otherwise, adding 1 to the iteration times, and transferring to the step B3 until a global optimum value is found, wherein the global optimum value is the optimal weight and the threshold of the BILSTM deep neural network.
The quantum particle swarm optimization is an improved particle swarm optimization, effectively solves the problem of insufficient global convergence performance of the particle swarm optimization, enables an example to be searched in the whole feasible solution space, thereby searching a global optimal solution and improving the global convergence and the searching capability of the algorithm.
In step 2), the obtaining process of the unmanned train environment energy consumption prediction model comprises the following steps: and searching the optimal weight and threshold of the GRU deep neural network by using a hybrid particle swarm algorithm and an universal gravitation search algorithm, wherein the GRU deep neural network corresponding to the optimal weight and threshold is the passenger flow energy consumption prediction model of the unmanned train.
Compared with the LSTM, the GRU algorithm is easier to train and improves the operation speed, can quickly meet the training requirements while retaining various important data characteristics in the processing of a large amount of unmanned train environment data, and outputs a prediction result, thereby facilitating the optimization training.
In step 3), the energy consumption prediction model
The expression of (a) is:
wherein ,
respectively obtaining a prediction result of an unmanned train vehicle running energy consumption control prediction model, a prediction result of an unmanned train passenger flow energy consumption prediction model and a prediction result of an unmanned train environment energy consumption prediction model; w is a1、w2、w3Are weight coefficients. The fused model can further improve the prediction precision.
The concrete implementation process of finding the optimal weight and threshold of the GRU deep neural network by using a hybrid particle swarm algorithm and a universal gravitation search algorithm comprises the following steps:
c1, initializing input parameters of a PSOGSA gravity search algorithm and a particle swarm algorithm;
c2, taking the position vector of the particle individual as the weight and the threshold of the GRU deep neural network, setting the individual extreme value of each particle as the current position, taking the reciprocal of the Mean Square Error (MSE) of the predicted value and the actual value as a fitness function, calculating the fitness value of each particle by using the fitness function, and taking the individual extreme value corresponding to the particle with the highest fitness value as the initial global extreme value; sequentially substituting parameter values corresponding to the particle individual position vectors into an unmanned train environment energy consumption prediction model based on a GRU depth network, and calculating weight calculation results of train rainfall resistance energy consumption, road ponding resistance energy consumption, wind resistance energy consumption and ponding energy consumption determined by the particle individual positions, and calculating to obtain the sum of environment energy consumption;
c3, calculating the fitness value of each particle after each iteration according to the fitness function set by C2, namely the reciprocal of the Mean Square Error (MSE) of the predicted value and the actual value;
c4, comparing the fitness value of the single particle with the fitness value of the individual extreme value thereof, if the fitness value of the single particle is better, updating the individual extreme value, otherwise, keeping the original value;
c5, updating the gravity coefficient and the inertia mass of the particles, and calculating the speed and the acceleration based on the fitness value of each particle to update the position of each particle;
c6, calculating the global optimal values of the positions and the fitness of the particles by using a GSA optimization algorithm;
c7, judging whether a termination condition is met, if so, exiting, otherwise, turning to the step C3 until a global optimal value is found, wherein the global optimal value is the optimal weight and the threshold of the GRU deep neural network.
The hybrid particle swarm algorithm and the universal gravitation search algorithm are improved algorithms of the particle swarm algorithm, the advantages that the global optimizing capability of the universal gravitation search algorithm and the particle swarm algorithm can increase the memory and the social information exchange capability of the particles are utilized, the optimizing capability of the hybrid algorithm is improved, and meanwhile the defect of algorithm stagnation is effectively overcome.
The invention also provides an energy consumption prediction system of the unmanned train, which comprises computer equipment; the computer device is configured or programmed for performing the steps of the method of the invention.
The present invention also provides a computer storage medium storing a program; the program is configured or programmed for performing the steps of the method of the invention.
Compared with the prior art, the invention has the beneficial effects that:
1. the invention considers the energy consumption conditions of the unmanned vehicle in various complex environments inside and outside the train in addition to the common running energy consumption, realizes the comprehensive calculation and prediction of the unmanned train influence parameters, organically combines the real-time train detection data, the passenger data and the external environment influence, and provides more complete consideration factors for the energy-saving research and development field of the unmanned vehicle.
2. The invention carries out human-computer loop fusion construction besides the traditional train mechanics analysis and train running state research, fully considers each subsystem element on the basis of the actual train running experience, provides human-computer loop matching conditions and standards, and establishes a multi-target unmanned train energy-saving optimization model. Meanwhile, different energy-saving regulation strategies are selected through the collected data, and comprehensive detection measures and control schemes are provided. The unmanned train is operated in an energy-saving optimization mode on the basis of personnel, vehicle and environmental parameters, so that the operation cost of the train can be saved, and the convenience and the energy-saving benefit of the unmanned train are improved to the greatest extent.
3. The invention comprehensively considers various factors such as speed energy consumption value, driving distance, road slope loss power, in-vehicle load and comfort energy consumption, accumulated snow water and wind power and the like under the stable driving speed of the unmanned train, performs weighted fusion by utilizing various prediction models and optimization algorithms, establishes a parameter-based traction energy consumption model for energy consumption analysis and prediction, and ensures the effectiveness of energy consumption strategy selection and energy consumption real-time adjustment instructions. The configuration of unmanned train energy consumption multi-objective optimization and the lowest energy consumption of the unmanned train running route are realized, and the energy consumption of the whole urban rail transit network can be reduced.
Drawings
FIG. 1 is a schematic block diagram of an embodiment of the present invention.
Detailed Description
As shown in fig. 1, the embodiment of the present invention mainly includes the following steps:
step 1: unmanned train human-machine loop parameter signal acquisition
The method comprises the following steps of train operation data acquisition, in-train and station passenger data acquisition and internal and external environment data acquisition, wherein the specific contents are as follows:
1) unmanned train operation data acquisition
The training data collected by the vehicle-mounted equipment, the trackside equipment and the like of the unmanned train comprise a stable running speed energy consumption value, a running distance and road gradient loss power of the unmanned train in a specified time interval in the running process.
The real-time running speed and the running distance within the interval time T of the unmanned train can be obtained by the vehicle-mounted equipment. The road gradient loss power is calculated by combining the vehicle body offset angle information collected by a level meter arranged at the bottom of the vehicle with the vehicle running speed. And integrating the time by using a climbing factor of the unmanned train in the interval time T.
2) In-train and station passenger data acquisition for driverless train
The training data collected by the unmanned train vehicle-mounted equipment, the station fixing equipment and the like comprise passenger flow, riding comfort evaluation indexes, consignment luggage weight, in-train seat-boarding rate and the like. And the comfort evaluation is obtained by fusion calculation of temperature data measured by the temperature sensor and the train ventilation volume, and the rest are obtained by the load sensor to obtain train load data.
3) Unmanned train exterior environment data acquisition
The training data are collected through unmanned train vehicle-mounted equipment, railway monitoring stations and the like, and comprise rainfall resistance, road ponding resistance, wind resistance energy consumption, accumulated snow energy consumption, temperature energy consumption and the like in a specified time interval.
The rainfall resistance calculation module utilizes data collected by a force-sensitive sensor arranged on the surface of the unmanned train to fuse and obtain the rainfall resistance of the surface of the vehicle within the interval time T.
The vehicle-mounted camera is used for collecting road accumulated water images, accumulated water information is collected by the aid of the infrared detector beside the rail, and accumulated water energy consumption is calculated by combining vehicle water penetration resistance.
The wind resistance energy consumption is obtained by combining a trackside wind measuring station and an auxiliary wind measuring station thereof to acquire wind speed data in real time to obtain a wind speed sample set along the railway and accumulating relative wind speeds in an interval time T.
The energy consumption generated by the accumulated snow on the road surface is obtained by acquiring a road surface image by using an image acquisition device, performing triangulation on an infrared image and a reference image to obtain a depth image, and identifying and extracting characteristics of reconstructed three-dimensional information by using a three-dimensional reconstruction method.
The temperature energy consumption value is obtained by accumulating the difference value of the temperature inside the unmanned vehicle and the temperature outside the unmanned vehicle within the interval time T;
step 2: unmanned train man-machine ring parameter signal transmission and preprocessing
Each carriage of the unmanned train is provided with a wireless transmission device for connecting a vehicle-mounted data acquisition module, a vehicle-mounted data acquisition module and a platform data processing center, so that data acquisition and storage and data transmission of different modules are realized, and wireless network can be adopted for transmission.
And a central computer is arranged on each row of unmanned trains to form a data processing model to receive key data acquired from the monitoring range, respectively perform data preprocessing and model training, and output a model training result in real time. And meanwhile, the man-machine interaction end is used for receiving the instruction sent by the platform data center and displaying the instruction on the interaction port to guide the train to operate in the next step.
Meanwhile, original data are preprocessed through methods of eliminating error information, supplementing missing values, designing labels and the like. The validity and diversity of the original data are ensured. Meanwhile, the original data is divided into a training set, a verification set and a test set. The neural network is fully trained, the performance of the model is accurately tested, the multi-class data are further integrated and processed by utilizing an optimization algorithm, and the prediction performance is improved.
And step 3: human-computer loop parameter information processing model of unmanned train
Step 3.1: control and prediction model for training vehicle running energy consumption of unmanned train
And training the energy consumption of the unmanned train by adopting a DBN deep belief neural network to train historical data. The inputs of the model are the energy consumption value for stable running of the train, the energy consumption values for climbing and descending slopes, the power consumption and historical data of running distance. 2000 pieces of time series of energy consumption for driving of the unmanned train are collected, and the obtained data are divided into a training set, a verification set and a test set. The weight process of the unmanned train operation energy consumption value obtained by performing integrated optimization on the input parameters by adopting a Huilus optimization (CS-GWO) algorithm fused with rhododendron search based on DBN deep belief neural network prediction is as follows:
a1: obtaining history data of running energy consumption of unmanned train
The energy consumption historical data comprises a stable running energy consumption value of the train, a climbing and descending energy consumption value, a power consumption and a running distance. And loading a driving data set of the unmanned train and pre-cleaning training data.
A2: and the energy consumption value of stable running of the train, the energy consumption value of climbing and descending, the power loss and the running distance which are concentrated by historical running data verification are used as input data, the energy consumption after the interval time T is used as output data, and the weight and the threshold of the DBN neural network in the energy consumption prediction model of the unmanned vehicle based on the DBN neural network are trained and optimized by adopting a wolf optimization (CS-GWO) algorithm which is fused with rhododendron search.
Setting the maximum iteration times m of the algorithm, setting the size of the wolf population, and setting the probability Pa of the wolf population to be discovered; the dimension n of the data set is set.
A3: randomly placing the position of each gray wolf in a wolf cluster in a solution space, setting the population number of the gray wolfs as m, the characteristic number of a vehicle driving energy consumption original data set as d, and a position matrix of the wolf cluster as an m x d dimensional binary matrix, and setting the position of the ith gray wolf as CGi＝(CGi1,CGi2,...CGid)，CGi1,CGi2,...CGidCoding each dimension of each gray wolf in the wolf group; processing an input data set of the DBN deep confidence neural network according to the position information in the wolf individuals to obtain a new data set, calculating and evaluating the fitness value fitness of each wolf individual, and taking the maximum iteration number iter of the individual wolfmaxThe position obtained by the secondary iteration is used as a local optimal solution, and the local optimal solution is the weight and the threshold of the DBN neural network;
a4: finding and recording the positions X of the alpha wolf, the beta wolf and the delta wolf from large to small according to the fitness valueα、Xβ、Xδ。
A5: disturbing the positions of the alpha wolf, the beta wolf and the delta wolf in the step 4 through the following formula, wherein the value of the gray wolf is decimal number
wherein ,xi(t +1) is the updated position of the ith bird nest, alpha is the variable of the control step length, and L (lambda) is the random number representing the Lewy distribution as shown in the following formula
L*u＝t-λ,1＜i≤n
The updated gray wolf location is then obtained by mapping it to binary using the following notations. And comparing the disturbed position with the original position, and reserving the position which is more close to the prey.
in the formula ,
represents the value of the d-dimension component of the i-th parasitic nest in the process of the algorithm to the i-th generation, and random is [0, 1]]A random number in between; the next generation value of the parasitic nest is converted into 0 or 1 through mapping, and the conversion from decimal number to binary value is realized
A6: for each gray wolf individual in the wolf group, the GWO algorithm is used for updating the formula to calculate the position X of the alpha wolf, the beta wolf and the delta wolfα、Xβ、XδAnd at the moment, the values of the gray wolves are decimal numbers, and the positions of the updated gray wolves are obtained by using the following formula.
A7: and calculating the updated probability Pa of finding the wolf target, and if r is greater than Pa, finding the wolf target and updating the position of the wolf individual. And evaluating the quality of the updated position and the original position according to the fixness, and reserving a position which is closer to the optimal solution of the prey and is better. Otherwise, directly entering the next step.
A8: and updating the positions of the recorded alpha wolf, beta wolf and delta wolf. Judging whether the iteration times of the algorithm reach the maximum iteration times max or not, if so, stopping the iteration of the algorithm, and outputting a global optimal value obtained according to the alpha wolf position and the fitness value fitness of the alpha wolf position; otherwise, executing A5 again until a global optimal value is found, and outputting the optimal weight and threshold of the DBN neural network.
A9: and inputting the data of the unmanned driving historical driving data training set into the DBN neural network for training, outputting a prediction result and evaluating the performance of the model by using a test set.
Step 3.2: model for predicting passenger flow energy consumption of training unmanned train
The BILSTM deep neural network is adopted to train passenger flow historical data of the unmanned train, and the BILSTM is an abbreviation of a bidirectional long-short term memory and structurally consists of a forward LSTM and a backward LSTM. The BILSTM may connect two hidden layers with different directions for output, which means that it contains both forward and reverse data. The output layer may obtain past and future information in the input data through the structure. The BILSTM deep neural network is capable of efficiently analyzing time series data with long and short term dependencies.
The training data collected by the unmanned train vehicle-mounted equipment, the station fixing equipment and the like comprise passenger flow, riding comfort evaluation indexes, consignment luggage weight and in-train seat-in rate. 2000 passenger flow energy consumption time sequences of the unmanned train are collected, and the obtained data are divided into a training set, a verification set and a test set. The integrated optimization process of the prediction of the input parameters based on the BILSTM deep neural network by adopting the quantum particle swarm optimization is as follows:
b1: obtaining passenger flow energy consumption historical data of unmanned train
And loading a passenger flow energy consumption data set of the unmanned train, and pre-cleaning training data. And (3) verifying concentrated passenger flow, riding comfort evaluation indexes, consignment luggage weight and in-vehicle seat occupancy as input data by using historical passenger flow energy consumption data, and training and optimizing the weight and threshold of the BILSTM neural network in the BILSTM neural network-based unmanned vehicle passenger flow energy consumption prediction model by using a quantum particle swarm algorithm by using energy consumption after an interval time T as output data.
B2: taking the position vector of each quantum particle individual in the quantum particle swarm as the weight and the threshold of the BILSTM neural network, and initializing the position vector parameters of the quantum particle swarm individual into random numbers of [ -1,1 ]; the value range of the number of the quantum particle swarm is [10,100], the value range of the number of the particles of the quantum particle swarm is [3,60], the value range of the maximum iteration frequency is [200,1200], the value range of the iteration frequency for constructing the elite swarm is [20,200], the value range of the premature convergence judgment threshold is [0.02,0.5], and the value range of the variation proportion of the worst particles of the swarm is [ 1%, 5% ];
b3: setting a fitness function, determining an initial optimal individual quantum particle position vector and iteration times t, wherein t is 1, sequentially bringing parameter values corresponding to the individual quantum particle position vector, utilizing weight calculation results of energy consumption parameters of each train determined by the individual quantum particle position, and taking the reciprocal of Mean Square Error (MSE) of the calculation results and actual values as a second fitness function f2(x)，f2(x)＝1/MSE；
B4: calculating the group fitness variance of each quantum particle swarm, and performing premature convergence judgment;
if the group fitness variance of the quantum particle swarm is smaller than the premature convergence judgment threshold gamma, delta% of worst fitness particles and group extreme value particles in the quantum particle swarm are varied, and the particles with the best current fitness are taken as global optimal quantum particle individuals;
b5: judging whether to establish an elite population; when the iteration times are larger than the iteration times of the elite population, extracting extreme values of all populations through information sharing among the populations to establish the elite population, and turning to the step B9, otherwise, turning to the step B6;
b6: updating parameters of each population of particles;
b7: recalculating and comparing the fitness value of each particle, and updating the individual extreme value if the fitness value is superior to the current individual extreme value; comparing the global extreme value particles, if the particle fitness value is superior to the current group extreme value, updating the global extreme value particles, enabling t to be t +1, and turning to the step B4;
b8: the elite population continues to evolve;
b9: and judging whether the maximum iteration times are met, if so, exiting, otherwise, enabling t to be t +1, turning to a step B4 until a global optimal value is found, and outputting the optimal weight and threshold of the passenger flow energy consumption prediction model of the corresponding BILSTM deep network.
B10: and inputting the data of the unmanned passenger flow energy consumption data training set into the BILSTM neural network for training, outputting a prediction result and evaluating the performance of the model by using a test set.
Step 3.3: model for predicting environmental energy consumption of training unmanned train
Energy consumption prediction caused by natural environment based on GRU deep neural network is adopted, and the input of the model is rainfall resistance energy consumption, road ponding resistance energy consumption, wind resistance energy consumption and ponding energy consumption data during the operation of the unmanned train. 2000 pieces of unmanned train environment energy consumption time series are collected, and the obtained data are divided into a training set, a verification set and a test set. The process of optimizing and selecting the weight of each parameter prediction result of the model by adopting a hybrid particle swarm optimization algorithm and a universal gravitation search algorithm (PSOGSA) is as follows:
c1: obtaining the environmental energy consumption historical data of the unmanned train
And loading a passenger flow energy consumption data set of the unmanned train, and pre-cleaning training data. And verifying concentrated rainfall resistance energy consumption, road ponding resistance energy consumption, wind resistance energy consumption and ponding energy consumption as input data by using the historical passenger flow energy consumption data, taking energy consumption after an interval time T as output data, and training and optimizing the weight and threshold of the GRU neural network in the unmanned vehicle environment energy consumption prediction model based on the GRU neural network by adopting a hybrid particle swarm algorithm and a universal gravitation search algorithm.
C2: initializing input parameters of a PSOGSA particle swarm algorithm, setting the size of a swarm to be 50, setting the maximum iteration number to be 1000, setting the function dimension to be 30, setting the inertia weight to be 0.9, setting the acceleration factor to be c1 to be 0.5, and setting the c2 to be 1.5; the calculation formula of PSOGSA is as follows:
Vi(t+1)＝w×Vi(t)+d1×rand×aci(t)+d2×rand×(gbest-Xi(t))
Xi(t+1)＝Xi(t)+Vi(t+1)
in the above formula, ViDenotes speed, Xi(t) is the current position of the ith particle, t is the number of iterations, the inertial weight is the weight, d1 and d2Is the acceleration coefficient, rand is [0, 1]]A random variable between, aci(t) is the acceleration of the ith particle in t iterations, gbest is the current best solution, ViIs the velocity of the ith particle.
C3: calculating a fitness value, taking the position vector of each particle as the weight and the threshold of the GRU neural network, setting the individual extreme value of each particle as the current position, calculating the fitness value of each particle by using a fitness function, and taking the individual extreme value corresponding to the particle with good fitness as the initial global extreme value. Sequentially bringing in parameter values corresponding to the particle individual position vectors, utilizing the weight calculation result of the energy consumption parameters of each train determined by the particle individual positions, and taking the Mean Square Error (MSE) of the calculation result and the actual value as a fitness function
C4: calculating the fitness value of each particle after each iteration according to the fitness function of the particles;
c5: comparing the fitness value of each particle with the fitness value of the individual extreme value of each particle, if the fitness value is better, updating the individual extreme value, otherwise, keeping the original value;
c6: updating gravitational coefficient and inertial mass of particle, calculating velocity and acceleration to update particle position
C7: using global optima for calculating particle position and fitness
C8: and judging whether a termination condition is met, if so, exiting, otherwise, turning to the step C4 until a global optimum value is found, and outputting the optimal weight and threshold of the corresponding GRU depth network environmental energy consumption prediction model.
C9: and inputting the data of the unmanned environment energy consumption data training set into the GRU neural network for training, outputting a prediction result and evaluating the performance of the model by using a test set.
And 4, step 4: energy consumption adjustment for unmanned train
After obtaining information of the unmanned train, such as vehicle running energy consumption, passenger flow energy consumption, environmental energy consumption and the like, the vehicle-mounted central computer of the unmanned train and the platform data center output real-time energy consumption information, comprehensively compare factors of running conditions, such as central energy-saving adjustment instructions, central departure intervals, horizontal line conditions, vertical line conditions, vehicle braking modes, train running positions and the like, which influence expert strategies in the unmanned train, guide the next-step running of the train according to the existing prediction results and the line real-time states in time, adopt a framework based on multi-population genetic algorithm MPGA exceeding and only relying on a single population to carry out genetic evolution, and introduce multiple populations to carry out optimization search simultaneously.
And 3, the three deep networks are used for completing prediction of different types of energy consumption sequences, and different from the traditional shallow neural network, the deep neural network has stronger learning and modeling capabilities. The multi-population genetic algorithm is used for integrating the prediction results of the three deep networks. The final prediction result is obtained by integrating the prediction results of the three deep networks, and the model set is realized by setting the weight coefficients of the prediction results of the deep networks. In addition, the integration of multiple deep learning methods can effectively improve the adaptability and robustness of the model.
In the running process of the unmanned train, the main running stages are an acceleration stage, a cruise stage, a coasting stage and a braking stage. The energy-saving control of the unmanned train is realized by completing a conversion instruction based on the energy consumption of the unmanned train, the passenger flow energy consumption and the environmental energy consumption brought by the influence of real-time man-machine ring parameters during the running of the train, and adjusting the running state of the train. The multi-population genetic algorithm MPGA is used for calculating working condition switching points, integrates various energy consumption predicted values, and feeds back a center to determine an optimal energy consumption position, namely, the position of each stage switching.
The algorithm for solving the unmanned train energy-saving operation strategy model comprises the following steps:
d1: reading the basic simulation data and calculating corresponding parameters. And reading corresponding predicted values of the vehicle running energy consumption, passenger flow energy consumption and environmental energy consumption of the unmanned train. Training and optimizing the predicted value weights of the three optimized neural networks by using a multi-population genetic algorithm MPGA (Luhua, Zhoucleng, Juanjuan, Zhengjinhua.) based on a multi-population evolved genetic algorithm [ J ] computer engineering and application, 2010,46(28): 57-60.).
D2: randomly generating a plurality of initial populations, setting population number, initial population individual number and individual length, and generating the initial populations P (t) by taking the position represented by the chromosome in the elite population as an optimal weight coefficient. And expanding and dividing into various populations according to the existing information:
P(t)＝{P1(t),P2(t),P3(t)} (1)
d3: and (4) determining control parameters. According to the method, different control parameters are taken to ensure the differential evolution of each population according to individuals in the discrete initial population under different operating conditions of the unmanned train. The main control parameter is the crossover probability PcAnd the mutation probability PbThe value of the search balance between the global search and the local search of the algorithm is maintained, and the calculation formula is as follows
Pco，PboRespectively, initial cross probability and mutation probability; g is the population number; c, b is the interval length of the crossover and mutation operation; f. ofrandAs a function of the generation of the random numbers. PcGenerally in the range of [0.7, 0.9 ]]Randomly generated within a range, PbGenerally in the range of [0.001,0.05 ]]And randomly generating in the interval, and taking the optimal chromosome in the multi-population genetic algorithm as the human-computer loop energy consumption integration weight.
D4: and optimizing each population by adopting a genetic algorithm, transferring chromosomes among the populations, preferentially selecting the optimal chromosomes from the optimized populations respectively, and adding the optimal chromosomes into the elite population, so that the elite population contains not only a local optimal solution but also a global optimal solution.
D5: selecting elite individuals according to the elite retention strategy and generating new populations. MPGA decides algorithm termination and extracts global optimal solution according to the elite population. And comparing the optimized result with the result before optimization, and selecting the optimal weight coefficient as the global optimal solution of the optimization problem.
D6: integrating the information of the energy consumption of the unmanned train vehicle, the energy consumption of passenger flow and the energy consumption of environment according to the result obtained by the multi-population genetic algorithm MPGA to generate an energy-saving solution, wherein w in the following formulaiAre the weighting coefficients of the three deep networks,
is the predicted outcome for each deep network.
And the operation result is fed back to the vehicle-mounted central computer of the unmanned train and the platform data center to output a real-time driving state conversion instruction, so that the tractive force energy consumption value of the unmanned train is effectively controlled. Compared with the traditional genetic algorithm, the MPGA is not easy to fall into local optimization, and vehicle running energy consumption, passenger flow energy consumption and environment energy consumption information of the unmanned train are integrated comprehensively aiming at different actual running environments of the unmanned train, so that not only is the self-adaptive adjustment of the optimal solution of each data characteristic realized, but also the overall optimization of the relevant characteristics is realized, and the reliability of the optimal solution is enhanced.
Claims (10)
1. The method for predicting the energy consumption of the unmanned train is characterized by comprising the following steps of:
1) collecting the operation data of the unmanned train, the passenger data in the train and at the station and the environment data outside the train; the unmanned train operation data comprises a stable running speed energy consumption value, a running distance and road gradient loss power of the train in a specified time interval in the running process; the passenger data in the train and at the station comprises passenger flow, riding comfort evaluation indexes, consignment luggage weight and in-train seat-up rate; the external environment data of the train comprises rainfall resistance, road ponding resistance, wind resistance energy consumption, accumulated snow energy consumption and temperature energy consumption in a specified time interval;
2) the unmanned train operation data is used as input of a DBN deep confidence neural network, the DBN deep confidence neural network is trained, and an unmanned train vehicle running energy consumption control prediction model is obtained; taking the data of passengers in the train and in the station as the input of a BILSTM deep neural network, training the BILSTM deep neural network, and obtaining a passenger flow energy consumption prediction model of the unmanned train; taking the external environment data of the train as the input of a GRU deep neural network, training the GRU deep neural network, and obtaining an unmanned train environment energy consumption prediction model;
3) fusing the unmanned train vehicle running energy consumption control prediction model, the unmanned train passenger flow energy consumption prediction model and the unmanned train environment energy consumption prediction model to obtain an energy consumption prediction model; preferably, the method further comprises the following steps:
4) and inputting the unmanned train operation data, the train inside and station passenger data and the train outside environment data which are acquired in real time into the energy consumption prediction model to predict the train energy consumption.
2. The method for predicting the energy consumption of the unmanned train according to claim 1, wherein in the step 2), the specific training process of the unmanned train vehicle driving energy consumption control prediction model comprises: and searching the optimal weight and the threshold of the DBN deep belief neural network by using a gray wolf optimization algorithm, wherein the DBN deep belief neural network corresponding to the optimal weight and the threshold is the vehicle running energy consumption control prediction model of the unmanned train.
3. The method for predicting the energy consumption of the unmanned train as claimed in claim 2, wherein the specific implementation process of finding the optimal weight and threshold of the DBN deep belief neural network by using the grey wolf optimization algorithm comprises:
a1, randomly placing the position of each gray wolf in wolf clusters in a solution space, setting the population number of the gray wolfs as m, the characteristic number of the original data set of the vehicle running energy consumption as d, and the position matrix of the wolf clusters as a m x d dimensional binary matrix, wherein the position of the ith gray wolf is CGi＝(CGi1,CGi2,...CGid)，CGi1,CGi2,...CGidCoding each dimension of each gray wolf in the wolf group; processing an input data set of the DBN deep confidence neural network according to the position information in the wolf individuals to obtain a new data set, calculating and evaluating the fitness value fitness of each wolf individual, and taking the maximum iteration number iter of the individual wolfmaxThe position obtained by the secondary iteration is used as a local optimal solution, and the local optimal solution is the weight and the threshold of the DBN neural network;
a2, finding and recording the positions X of alpha wolf, beta wolf and delta wolf according to the fitness value from big to small in sequenceα、Xβ、Xδ；
A3, disturbing the positions of the alpha wolf, the beta wolf and the delta wolf, comparing the disturbed positions of the alpha wolf, the beta wolf and the delta wolf with the original positions of the alpha wolf, the beta wolf and the delta wolf, and reserving the position closer to the optimal solution;
a4, for each gray wolf individual in the wolf group, calculating the positions of the alpha wolf, the beta wolf and the delta wolf by using a gray wolf algorithm, and obtaining the position of the updated gray wolf individual closer to the optimal solution by using the following formula:
wherein ,
represents the position value of the ith dimension of the ith Hui wolf individual in the process of the (t +1) th iteration, Xα、Xβ、XδRespectively, the positions of alpha wolf, beta wolf and delta wolf, random is [0, 1]]A random number in between;
a5, randomly updating the position of the wolf population according to a preset probability value Pa, calculating the updated probability r of finding the wolf prey, and if r is greater than Pa, finding the wolf prey and updating the position of the wolf individual; according to the fact that the updated position and the original position are evaluated to be good or bad, the position closer to the optimal solution is reserved; otherwise, go directly to step A6;
a6, judging whether the iteration times reach the maximum iteration times, if so, stopping the iteration, and outputting a global optimum value obtained according to the alpha wolf position and the fitness value fitness thereof; otherwise, returning to step A2 until a global optimum is found; the global optimal values are the optimal weight and the threshold of the DBN deep confidence neural network.
4. The method for predicting the energy consumption of the unmanned train according to claim 1, wherein in the step 2), the obtaining process of the passenger flow energy consumption prediction model of the unmanned train comprises: and searching the optimal weight and threshold of the BILSTM deep neural network by using a quantum particle group algorithm, wherein the BILSTM deep neural network corresponding to the optimal weight and threshold is the passenger flow energy consumption prediction model of the unmanned train.
5. The method for predicting the energy consumption of the unmanned train according to claim 4, wherein the specific implementation process of finding the optimal weight and threshold of the BILSTM deep neural network by using a quantum particle group algorithm comprises the following steps:
b1, taking the position vector of each quantum particle individual in the quantum particle swarm as the weight and the threshold of the BILSTM deep neural network, and initializing the position vector parameters of the quantum particle swarm individual into random numbers of [ -1,1 ]; initializing input parameters of a quantum particle swarm algorithm;
b2, setting a fitness function, and determining an initial optimal quantum particle individual position vector and iteration times by taking the reciprocal of the Mean Square Error (MSE) of a predicted value and an actual value as the fitness function; substituting the weight and the threshold corresponding to the quantum particle individual position vector into the passenger flow energy consumption prediction model of the unmanned train based on the BILSTM deep network, determining the type of the identification vector label by using the passenger flow energy consumption prediction model of the unmanned train based on the BILSTM deep network determined by the quantum particle individual position vector, determining the passenger flow, the riding comfort evaluation index, the weight of the consignment luggage and the weight calculation result of the passenger occupancy in the train and the station passenger data by using the quantum particle individual position, and calculating to obtain the sum of the passenger flow energy consumption;
b3, if the variance of the population fitness of the quantum particle swarm is smaller than the premature convergence judgment threshold, carrying out variation on the particles with the worst fitness and the extreme particles of the swarm in the quantum particle swarm, and taking the particles with the best fitness as the global optimal quantum particle individuals, namely the extreme values of the swarm;
b4, when the iteration times are more than the iteration times of the elite population, extracting each species through information sharing among the populations
Establishing an elite population by the extreme value of the population, and switching to the step B6, otherwise, switching to the step B6;
b5, updating the position parameters of each particle swarm by comparing the individual extreme value with the global extreme value, namely recalculating and comparing the fitness value of each particle, and updating the individual extreme value if the fitness value is superior to the current individual extreme value; comparing the global extreme value particles, if the particle fitness value is superior to the current population extreme value, updating the global extreme value particles, adding 1 to the iteration number, and turning to the step B3; the global extreme example is an extreme value obtained by evolution in all particle individuals including an elite population;
b6, repeating the steps B3-B5, judging whether the maximum iteration times are met, if so, ending, otherwise, adding 1 to the iteration times, and shifting to the step B3 until a global optimal value is found, wherein the global optimal value is the optimal weight and the threshold of the BILSTM deep neural network.
6. The method for predicting the energy consumption of the unmanned train according to claim 1, wherein in the step 2), the obtaining process of the model for predicting the environmental energy consumption of the unmanned train comprises: and searching the optimal weight and threshold of the GRU deep neural network by using a hybrid particle swarm algorithm and an universal gravitation search algorithm, wherein the GRU deep neural network corresponding to the optimal weight and threshold is the passenger flow energy consumption prediction model of the unmanned train.
7. The method for predicting the energy consumption of the unmanned train according to claim 6, wherein the concrete implementation process of finding the optimal weight and the threshold of the GRU deep neural network by using a hybrid particle swarm algorithm and a universal gravitation search algorithm comprises the following steps:
c1, initializing input parameters of a PSOGSA gravity search algorithm and a particle swarm algorithm;
c2, taking the position vector of the particle individual as the weight and the threshold of the GRU deep neural network, setting the individual extreme value of each particle as the current position, taking the reciprocal of the Mean Square Error (MSE) of the predicted value and the actual value as a fitness function, calculating the fitness value of each particle by using the fitness function, and taking the individual extreme value corresponding to the particle with the highest fitness value as the initial global extreme value; sequentially substituting parameter values corresponding to the particle individual position vectors into an unmanned train environment energy consumption prediction model based on a GRU depth network for iteration;
c3, calculating the fitness value of each particle after each iteration according to the fitness function set by C2;
c4, comparing the fitness value of the single particle with the fitness value of the individual extreme value thereof, if the fitness value of the single particle is better, updating the individual extreme value, otherwise, keeping the original value;
c5, updating the gravity coefficient and the inertia mass of the particles, and calculating the speed and the acceleration based on the fitness value of each particle to update the position of each particle;
c6, calculating the global optimal values of the positions and the fitness of the particles by using a GSA optimization algorithm;
c7, judging whether a termination condition is met, if so, exiting, otherwise, turning to the step C3 until a global optimal value is found, wherein the global optimal value is the optimal weight and the threshold of the GRU deep neural network.
8. The method for predicting the energy consumption of the unmanned train according to any one of claims 1 to 7, wherein in the step 3), the energy consumption prediction model is adopted
The expression of (a) is:
wherein ,
respectively obtaining a prediction result of an unmanned train vehicle running energy consumption control prediction model, a prediction result of an unmanned train passenger flow energy consumption prediction model and a prediction result of an unmanned train environment energy consumption prediction model; w is a1、w2、w3Are weight coefficients.
9. An energy consumption prediction system for an unmanned train is characterized by comprising computer equipment; the computer device is configured or programmed for carrying out the steps of the method according to one of claims 1 to 7.
10. A computer storage medium characterized by storing a program; the program is configured or programmed for carrying out the steps of the method according to one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011639921.4A CN112734097B (en) | 2020-12-31 | 2020-12-31 | Unmanned train energy consumption prediction method, system and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011639921.4A CN112734097B (en) | 2020-12-31 | 2020-12-31 | Unmanned train energy consumption prediction method, system and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112734097A true CN112734097A (en) | 2021-04-30 |
| CN112734097B CN112734097B (en) | 2023-09-05 |
Family
ID=75608917
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011639921.4A Active CN112734097B (en) | 2020-12-31 | 2020-12-31 | Unmanned train energy consumption prediction method, system and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112734097B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113361792A (en) * | 2021-06-21 | 2021-09-07 | 吉林大学 | Urban electric bus travel energy consumption estimation method based on multivariate nonlinear regression |
| CN113687242A (en) * | 2021-09-29 | 2021-11-23 | 温州大学 | Lithium ion battery SOH estimation method for optimizing and improving GRU neural network based on GA algorithm |
| CN113822473A (en) * | 2021-09-03 | 2021-12-21 | 浙江浙大中控信息技术有限公司 | Traction energy consumption reasonable interval prediction method based on multidimensional data |
| CN113997915A (en) * | 2021-11-26 | 2022-02-01 | 北京大象科技有限公司 | Big data-based automatic train operation ATO (automatic train operation) accurate parking control method |
| CN114545280A (en) * | 2022-02-24 | 2022-05-27 | 苏州市职业大学 | New energy automobile lithium battery life prediction method based on optimization algorithm |
| CN117195484A (en) * | 2023-08-08 | 2023-12-08 | 广东贝能达交通设备有限公司 | Rail transit management method and system |
| CN117408720A (en) * | 2023-11-14 | 2024-01-16 | 武汉亿量科技有限公司 | Method, device, equipment and medium for preventing complaint marks of outer calls of electric pins |
| CN117195484B (en) * | 2023-08-08 | 2024-05-03 | 广东贝能达交通设备有限公司 | Rail transit management method and system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108572648A (en) * | 2018-04-24 | 2018-09-25 | 中南大学 | A kind of automatic driving vehicle power supply multi-source fusion prediction technique and system |
| CN108846571A (en) * | 2018-06-08 | 2018-11-20 | 福州大学 | A kind of net connectionization electric car macroscopic view energy consumption estimation method |
| CN110852498A (en) * | 2019-10-31 | 2020-02-28 | 西安交通大学 | Method for predicting data center energy consumption efficiency PUE based on GRU neural network |
| CN111356620A (en) * | 2017-08-24 | 2020-06-30 | 图森有限公司 | System and method for autonomous vehicle control to minimize energy costs |
-
2020
- 2020-12-31 CN CN202011639921.4A patent/CN112734097B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111356620A (en) * | 2017-08-24 | 2020-06-30 | 图森有限公司 | System and method for autonomous vehicle control to minimize energy costs |
| CN108572648A (en) * | 2018-04-24 | 2018-09-25 | 中南大学 | A kind of automatic driving vehicle power supply multi-source fusion prediction technique and system |
| CN108846571A (en) * | 2018-06-08 | 2018-11-20 | 福州大学 | A kind of net connectionization electric car macroscopic view energy consumption estimation method |
| CN110852498A (en) * | 2019-10-31 | 2020-02-28 | 西安交通大学 | Method for predicting data center energy consumption efficiency PUE based on GRU neural network |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113361792A (en) * | 2021-06-21 | 2021-09-07 | 吉林大学 | Urban electric bus travel energy consumption estimation method based on multivariate nonlinear regression |
| CN113822473A (en) * | 2021-09-03 | 2021-12-21 | 浙江浙大中控信息技术有限公司 | Traction energy consumption reasonable interval prediction method based on multidimensional data |
| CN113822473B (en) * | 2021-09-03 | 2023-12-26 | 浙江中控信息产业股份有限公司 | Traction energy consumption reasonable interval prediction method based on multidimensional data |
| CN113687242A (en) * | 2021-09-29 | 2021-11-23 | 温州大学 | Lithium ion battery SOH estimation method for optimizing and improving GRU neural network based on GA algorithm |
| CN113997915A (en) * | 2021-11-26 | 2022-02-01 | 北京大象科技有限公司 | Big data-based automatic train operation ATO (automatic train operation) accurate parking control method |
| CN113997915B (en) * | 2021-11-26 | 2022-06-24 | 北京大象科技有限公司 | Big data-based automatic train operation ATO (automatic train operation) accurate parking control method |
| CN114545280A (en) * | 2022-02-24 | 2022-05-27 | 苏州市职业大学 | New energy automobile lithium battery life prediction method based on optimization algorithm |
| CN114545280B (en) * | 2022-02-24 | 2022-11-15 | 苏州市职业大学 | New energy automobile lithium battery life prediction method based on optimization algorithm |
| CN117195484A (en) * | 2023-08-08 | 2023-12-08 | 广东贝能达交通设备有限公司 | Rail transit management method and system |
| CN117195484B (en) * | 2023-08-08 | 2024-05-03 | 广东贝能达交通设备有限公司 | Rail transit management method and system |
| CN117408720A (en) * | 2023-11-14 | 2024-01-16 | 武汉亿量科技有限公司 | Method, device, equipment and medium for preventing complaint marks of outer calls of electric pins |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112734097B (en) | 2023-09-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112734097B (en) | Unmanned train energy consumption prediction method, system and storage medium | |
| CN108197739B (en) | Urban rail transit passenger flow prediction method | |
| CN105868942B (en) | The orderly charging schedule method of electric car | |
| CN112686453B (en) | Intelligent prediction method and system for locomotive energy consumption | |
| CN108647813B (en) | High-speed train dynamic interval energy-saving optimization method based on elastic force adjustment | |
| CN109508751B (en) | Deep neural network model modeling method for high-speed railway train late time prediction | |
| Liu et al. | An on-line energy management strategy based on trip condition prediction for commuter plug-in hybrid electric vehicles | |
| CN110299011A (en) | A kind of traffic flow forecasting method of the highway arbitrary cross-section based on charge data | |
| Kim et al. | Idle vehicle relocation strategy through deep learning for shared autonomous electric vehicle system optimization | |
| Guo et al. | A novel energy consumption prediction model with combination of road information and driving style of BEVs | |
| CN110220725A (en) | A kind of wheel for metro vehicle health status prediction technique integrated based on deep learning and BP | |
| CN103280100A (en) | Bus IC card passenger transfer station judgment method combining intelligent dispatching data | |
| CN106056238B (en) | Planning method for train interval running track | |
| Valera et al. | Driving cycle and road grade on-board predictions for the optimal energy management in EV-PHEVs | |
| CN114492919A (en) | Electric automobile charging load prediction system and method based on traffic flow | |
| CN112734094B (en) | Intelligent city intelligent rail vehicle fault gene prediction method and system | |
| CN114510778A (en) | Track irregularity prediction method based on hybrid intelligent optimization LSTM | |
| CN115805840A (en) | Energy consumption control method and system for range-extending type electric loader | |
| CN115907122A (en) | Regional electric vehicle charging load prediction method | |
| CN111369053A (en) | Method for predicting tracks of pedestrians and vehicles in human-vehicle interaction scene | |
| Grossard et al. | An optimal energy-based approach for driving guidance of full electric vehicles | |
| Amrani et al. | Train speed profiles optimization using a genetic algorithm based on a random-forest model to estimate energy consumption | |
| Gu et al. | A review of intelligent road preview methods for energy management of hybrid vehicles | |
| Ohde et al. | Statistical analysis of real-world urban driving cycles for modelling energy consumption of electric vehicles | |
| CN114495036A (en) | Vehicle track prediction method based on three-stage attention mechanism |
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 |