阅读说明：本技术 一种基于车轨状态监测的悬浮冗余控制系统及方法 (Suspension redundancy control system and method based on rail state monitoring ) 是由 林国斌 陈健 徐俊起 付善强 陈琛 郭海霞 荣立军 于 2021-08-26 设计创作，主要内容包括：本发明涉及一种基于车轨状态监测的悬浮冗余控制系统及方法,该系统用于在控制列车稳定悬浮的同时对车轨状态进行检测,并通过对悬浮间隙和电磁铁振动情况的分类和学习确定状态类型,所述的系统包括相互连接的电磁铁驱动主电路模块、用于正常情况悬浮控制的第一控制板和用于正常情况下车轨状态监测诊断和异常情况下替代第一控制板进行悬浮冗余控制的第二控制板；所述电磁铁驱动主电路模块包括用于控制切换第一控制板和第二控制板的第三FPGA模块和用于悬浮控制的驱动模块,所述的第三FPGA模块和驱动模块相互连接。与现有技术相比,本发明具有实时性好、容错性高等优点。(The invention relates to a suspension redundancy control system and a method based on rail state monitoring, the system is used for detecting the rail state while controlling the stable suspension of a train, and determining the state type by classifying and learning the suspension gap and the vibration condition of an electromagnet, the system comprises an electromagnet driving main circuit module, a first control board for suspension control under normal conditions and a second control board for performing suspension redundancy control under normal conditions by replacing the first control board under rail state monitoring diagnosis and abnormal conditions; the electromagnet driving main circuit module comprises a third FPGA module and a driving module, wherein the third FPGA module is used for controlling and switching the first control board and the second control board, the driving module is used for suspension control, and the third FPGA module is connected with the driving module. Compared with the prior art, the method has the advantages of good real-time performance, high fault tolerance and the like.)

1. A suspension redundancy control system based on rail state monitoring is characterized in that the system is used for detecting the rail state while controlling a train to stably suspend and determining the state type through classifying and learning the suspension gap and the vibration condition of an electromagnet, and comprises an electromagnet driving main circuit module, a first control board used for suspension control under normal conditions and a second control board used for replacing the first control board to perform suspension redundancy control under normal conditions and rail state monitoring diagnosis and abnormal conditions;

the electromagnet driving main circuit module comprises a third FPGA module and a driving module, wherein the third FPGA module is used for controlling and switching the first control board and the second control board, the driving module is used for suspension control, and the third FPGA module is connected with the driving module.

2. The levitation redundancy control system based on rail state monitoring as claimed in claim 1, wherein the first control board comprises a first FPGA module and a first DSP module which are connected with each other; the second control board comprises a second FPGA module, a second DSP module and a storage module which are connected with each other; the first DSP module is connected with the second DSP module; the output end of the driving module is respectively connected with the first DSP module and the second DSP module;

the first FPGA module and the second FPGA module are used for acquiring and processing sensor signals;

the first DSP module is used for calculating control quantity and carrying out Ethernet or CAN communication;

and the second DSP module is used for monitoring and diagnosing the rail state.

3. The levitation redundancy control system based on rail state monitoring as claimed in claim 2, wherein the storage module comprises a Flash memory for storing the sensor signal and the monitoring diagnosis signal sent by the second DSP module.

4. The system of claim 2, wherein the sensor signals comprise a 4-way clearance signal and a 2-way acceleration signal.

5. The suspension redundancy control system based on rail state monitoring as claimed in claim 1, wherein the first control board and the second control board are respectively connected with the electromagnet driving main circuit module through connectors.

6. The suspension redundancy control system based on the rail state monitoring is characterized in that the driving module comprises a driving plate and an IGBT module which are connected in sequence; and the IGBT module is used for receiving the PWM wave output by the drive plate and outputting control current to the electromagnet for suspension control.

7. A method of a levitation redundant control system based on rail condition monitoring according to claim 2, wherein the method comprises the steps of:

step S1: the sensor signals are transmitted to the first control board and the second control board at the same time, and the first FPGA module and the second FPGA module receive the sensor signals at the same time;

step S2: under normal conditions, the first control board performs suspension control, and the second control board performs monitoring and diagnosis of the rail state;

step S3: under the abnormal condition, when the first control board has a fault, the second DSP module cannot receive signals sent by the first DSP module, the second control board is controlled by the third FPGA module to take over the first control board for suspension control, the rail state monitoring and diagnosing work of the original second control board is suspended, and the classification and analysis results are stored in a Flash memory by taking the rail state monitoring and diagnosing work as a termination time point; after the first control board recovers, the second control board continues to take on the corresponding monitoring function.

8. The method according to claim 7, wherein the step S2 is specifically:

suspension control process: the first FPGA module transmits acquired sensor signals to the first DSP module after filtering processing to calculate control quantity and realize Ethernet or CAN communication, control signals are fed back to the first FPGA module, and the control signals are transmitted to the drive board through the third FPGA module to generate PWM waves to drive the IGBT module to generate control current to control the electromagnet to work in a suspension mode;

monitoring the diagnosis process: the second FPGA module processes the sensor signal and stores the processed sensor signal in a Flash memory; the second DSP module establishes a probabilistic neural network, analyzes the suspension condition according to different fault types, classifies the suspension condition according to the rail state, feeds the classification result back to the Flash memory for storage, transmits the analysis result to the first DSP module, and adjusts the output quantity for controlling suspension in real time, so that the control output can ensure the control current for stabilizing the suspension gap.

9. The method according to claim 8, wherein said monitoring a diagnostic procedure comprises in particular the steps of:

step S21: acquiring electromagnet vibration signals and suspension gap signals in different rail states in the suspension operation process, performing Fourier transform on the electromagnet vibration signals, and extracting time domain information and frequency domain information; calculating to obtain vibration mean value, variance, mean square value, peak value and frequency information;

step S22: establishing a probabilistic neural network, and classifying the vehicle rail states according to a Bayes minimum risk criterion;

input and output relationships Φ determined by a jth neuron defining a class i rail state_ij(x) Comprises the following steps:

wherein, i is 1,2, 6, d is the dimension of the sample, specifically represents the peak value, the mean value, the frequency of the electromagnet vibration and the key characteristics of the peak value and the mean value of the suspension gap, and x is_ijIs the jth center of the ith sample, and sigma is a smoothing factor;

weighted averaging is carried out on the outputs of the hidden neurons belonging to the same class in the hidden layer in the summation layer:

wherein v is_iOutputting the ith type of rail state, wherein L is the number of the ith type of neurons; the number of the neurons of the summation layer is the same as the number of the rail state categories;

the output layer takes the largest one of the summation layers as the class of output:

y＝arg max(v_i)

in actual calculation, the vector of the input layer is multiplied by a weighting coefficient, and then the vector is input into a radial basis function of the hidden layer for calculation:

Z_i＝xω_i

wherein Z_iComputing vector for the ith neuron radial basis, x being input vector of input layer, omega_iThe corresponding weight value of the ith neuron which is transmitted to the hidden layer for the input layer;

suppose x and ω_iAll are normalized to a unit length, and the results are subjected to radial basis calculationThereby obtaining a monitoring diagnosis result;

step S23: and transmitting the classification and analysis results to a first DSP module of the first control board to dynamically adjust the control quantity output, so as to realize the adaptive upgrade of the suspension system.

10. The method according to claim 9, wherein the rail state in step S21 includes: the suspension type track breaking device comprises a stable suspension working state, a track height difference exceeding +/-3 mm, a single-span beam end part generating a large arch angle state, an electromagnet instability track breaking state, a controller no-current output state and an electromagnet deadlocking state.

Technical Field

The invention relates to the field of suspension control of maglev trains, in particular to a suspension redundancy control system and method based on rail state monitoring.

Background

As a novel rail vehicle, the maglev train can realize non-contact operation on the premise of effectively meeting similar passenger carrying requirements of large-capacity transportation vehicles such as high-speed rails and subways, so that the problems of wheel-rail abrasion and noise of the conventional rail vehicle are solved, the operation and maintenance cost is effectively reduced, and the upper limit of the operation speed is improved. In addition, because the currently developed and operated magnetic suspension trains all adopt a rail holding form, the possibility of occurrence of serious accidents such as derailment and the like is effectively avoided, and the safety and reliability are further improved. The electromagnet is an actuating mechanism for ensuring the stable suspension of the magnetic-levitation train, and the suspension control system combines signals output by the suspension sensor and the acceleration sensor to change the internal current of the electromagnet, so that the magnetic-levitation train can stably suspend at an expected suspension gap (8-10 mm). In the development and research process of the last decades, the technology of the magnetic suspension train is basically mature and gradually goes to commercial production and operation.

Currently, commercially operated magnetic levitation trains (EMS type magnetic levitation trains) are all of electromagnetic levitation type magnetic levitation trains. Because the adjustable range of the suspension clearance of the train in the dynamic suspension running process is small, the influence of the train rail state on the suspension stability of the train is large.

The retrieved Chinese patent publication No. CN108382265A discloses a suspension redundancy control system of a low-speed maglev train, which comprises four suspension electromagnetic iron magnetic poles arranged on an electromagnet on one side of a suspension frame, two suspension sensors arranged on two sides of the electromagnet, four suspension choppers connected with the four suspension electromagnetic iron magnetic poles in a one-to-one correspondence manner, and suspension controllers respectively connected with the suspension choppers, wherein the suspension controllers are respectively connected with the two suspension sensors; however, based on the current technical means, the parameters of the levitation control algorithm are relatively fixed, the rail state cannot be analyzed in real time, and the rail state can only be determined by adopting a method of first acquisition and then analysis, which results in that the analysis efficiency is too low and the adjustment of the levitation gap of the train in the dynamic levitation process is not real-time.

Disclosure of Invention

The invention aims to overcome the defects of the prior art and provide a suspension redundancy control system and a suspension redundancy control method based on rail state monitoring, which have good real-time performance and high fault tolerance.

The purpose of the invention can be realized by the following technical scheme:

according to a first aspect of the invention, a suspension redundancy control system based on rail state monitoring is provided, the system is used for detecting the rail state while controlling a train to suspend stably, and determining the state type through classification and learning of suspension clearance and electromagnet vibration conditions, the system comprises an electromagnet driving main circuit module, a first control board used for suspension control under normal conditions and a second control board used for suspension redundancy control under normal conditions and rail state monitoring diagnosis and abnormal conditions instead of the first control board;

the electromagnet driving main circuit module comprises a third FPGA module and a driving module, wherein the third FPGA module is used for controlling and switching the first control board and the second control board, the driving module is used for suspension control, and the third FPGA module is connected with the driving module.

Preferably, the first control board comprises a first FPGA module and a first DSP module which are connected with each other; the second control board comprises a second FPGA module, a second DSP module and a storage module which are connected with each other; the first DSP module is connected with the second DSP module; the output end of the driving module is respectively connected with the first DSP module and the second DSP module;

the first FPGA module and the second FPGA module are used for acquiring and processing sensor signals;

the first DSP module is used for calculating control quantity and carrying out Ethernet or CAN communication;

and the second DSP module is used for monitoring and diagnosing the rail state.

Preferably, the storage module comprises a Flash memory for storing the sensor signal and the monitoring and diagnosing signal sent by the second DSP module.

Preferably, the sensor signals include 4 gap signals and 2 acceleration signals.

Preferably, the first control board and the second control board are respectively connected with the electromagnet driving main circuit module through connectors.

Preferably, the driving module comprises a driving board and an IGBT module which are connected in sequence; and the IGBT module is used for receiving the PWM wave output by the drive plate and outputting control current to the electromagnet for suspension control.

According to a second aspect of the present invention, there is provided a method of the suspension redundancy control system based on the rail condition monitoring, the method comprising the following steps:

step S1: the sensor signals are transmitted to the first control board and the second control board at the same time, and the first FPGA module and the second FPGA module receive the sensor signals at the same time;

step S2: under normal conditions, the first control board performs suspension control, and the second control board performs monitoring and diagnosis of the rail state;

step S3: under the abnormal condition, when the first control board has a fault, the second DSP module cannot receive signals sent by the first DSP module, the second control board is controlled by the third FPGA module to take over the first control board for suspension control, the rail state monitoring and diagnosing work of the original second control board is suspended, and the classification and analysis results are stored in a Flash memory by taking the rail state monitoring and diagnosing work as a termination time point; after the first control board recovers, the second control board continues to take on the corresponding monitoring function.

Preferably, the step S2 is specifically:

suspension control process: the first FPGA module transmits acquired sensor signals to the first DSP module after filtering processing to calculate control quantity and realize Ethernet or CAN communication, control signals are fed back to the first FPGA module, and the control signals are transmitted to the drive board through the third FPGA module to generate PWM waves to drive the IGBT module to generate control current to control the electromagnet to work in a suspension mode;

monitoring the diagnosis process: the second FPGA module processes the sensor signal and stores the processed sensor signal in a Flash memory; the second DSP module establishes a probabilistic neural network, analyzes the suspension condition according to different fault types, classifies the suspension condition according to the rail state, feeds the classification result back to the Flash memory for storage, transmits the analysis result to the first DSP module, and adjusts the output quantity for controlling suspension in real time, so that the control output can ensure the control current for stabilizing the suspension gap.

Preferably, the monitoring and diagnosing process specifically comprises the steps of:

step S21: acquiring electromagnet vibration signals and suspension gap signals in different rail states in the suspension operation process, performing Fourier transform on the electromagnet vibration signals, and extracting time domain information and frequency domain information; calculating to obtain vibration mean value, variance, mean square value, peak value and frequency information;

step S22: establishing a probabilistic neural network, and classifying the vehicle rail states according to a Bayes minimum risk criterion;

input and output relationships Φ determined by a jth neuron defining a class i rail state_ij(x) Comprises the following steps:

wherein, i is 1,2, 6, d is the dimension of the sample, specifically represents the peak value, the mean value, the frequency of the electromagnet vibration and the key characteristics of the peak value and the mean value of the suspension gap, and x is_ijIs the jth center of the ith sample, sigma is a smoothing factor, and is generally 0.1;

weighted averaging is carried out on the outputs of the hidden neurons belonging to the same class in the hidden layer in the summation layer:

wherein v is_iOutputting the ith type of rail state, wherein L is the number of the ith type of neurons; the number of the neurons of the summation layer is the same as the number of the rail state categories;

the output layer takes the largest one of the summation layers as the class of output:

y＝argmax(v_i)

in actual calculation, the vector of the input layer is multiplied by a weighting coefficient, and then the vector is input into a radial basis function of the hidden layer for calculation:

Z_i＝xω_i

wherein Z_iComputing vector for the ith neuron radial basis, x being input vector of input layer, omega_iThe corresponding weight value of the ith neuron which is transmitted to the hidden layer for the input layer;

suppose x and ω_iAll are normalized to a unit length, and the results are subjected to radial basis calculationThereby obtaining a monitoring diagnosis result;

step S23: and transmitting the classification and analysis results to a first DSP module of the first control board to dynamically adjust the control quantity output, so as to realize the adaptive upgrade of the suspension system.

Preferably, the rail state in step S21 includes: the suspension type track breaking device comprises a stable suspension working state, a track height difference exceeding +/-3 mm, a single-span beam end part generating a large arch angle state, an electromagnet instability track breaking state, a controller no-current output state and an electromagnet deadlocking state.

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

1) the invention realizes the monitoring and diagnosis of the rail state while carrying out suspension control; the method has the advantages that suspension gap data and electromagnet vibration data are obtained in real time in the suspension operation process, Fourier transform is carried out on the vibration data, time domain information and frequency domain information are extracted, and the peak value, the average value, the inherent frequency and other key characteristics are calculated according to the results, so that the rail state can be diagnosed in real time, the rail state type is monitored and diagnosed by combining a probabilistic neural network, the problems of low analysis efficiency and poor real-time performance after the prior acquisition are effectively solved, and convenience is brought to system overhaul and maintenance;

2) the electromagnet vibration signal and the suspension gap data acquired by the second control board are transmitted to the first DSP module of the first control board in real time, and the control quantity output of the first control board is adjusted to ensure that the train can reach the optimal stable state, so that the real-time adaptability optimization of the suspension control system is realized;

3) the electromagnet for redundancy control drives a third FPGA in a main circuit module to control switching; when the system is in a normal working state, the first control board is used for calculating electromagnet control current and CAN/Ethernet communication; the second control board stores the received and processed sensor signals in a Flash memory, and after calculation such as Fourier transform and the like is carried out on the data, the data are used as a data base for building and training the probabilistic neural network. And classifying the rail states based on the probabilistic neural network. And transmitting the analysis result to the first control board for optimization and adjustment after further processing. When the first control board is in fault, the second control board takes over the first control board for suspension control and carries out Ethernet or CAN communication; at the moment, the rail state monitoring and diagnosing work is suspended, and the classification and analysis results are stored in a Flash memory by taking the rail state monitoring and diagnosing work as a termination time point. After the first control board recovers, the second control board continues to take on the corresponding monitoring function.

4) The invention integrates the original signal board and control board, and can add the rail diagnosis function on the basis of not changing the original case structure; the functions of the control panel are expanded, on one hand, the real-time optimization and adjustment of suspension control can be realized, and on the other hand, the electrical redundancy of the control panel is realized.

Drawings

FIG. 1 is a block diagram of a redundant levitation system of a magnetic levitation train with rail condition monitoring and diagnosis;

fig. 2 is a structure of a rail condition monitoring fault diagnosis algorithm proposed in this patent.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, shall fall within the scope of protection of the present invention.

An embodiment of the system of the present invention is given below. A suspension redundancy control system based on rail state monitoring mainly comprises an electromagnet driving main circuit module, a first control board and a second control board, wherein the electromagnet driving main circuit module is connected with the electromagnet driving main circuit module, the first control board is used for suspension control under normal conditions, and the second control board is used for performing suspension redundancy control under normal conditions, rail state monitoring and diagnosis and abnormal conditions by replacing the first control board; the electromagnet driving main circuit module comprises a third FPGA module and a driving module, wherein the third FPGA module is used for controlling and switching the first control board and the second control board, and the driving module is used for outputting control current and performing suspension control; the first control board comprises a first FPGA module and a first DSP module which are connected with each other; the second control board comprises a second FPGA module, a second DSP module and a storage module which are connected with each other; the first control board and the second control board are connected through a first DSP module and a second DSP module; the output end of the driving module is respectively connected with the first DSP module and the second DSP module; and a Flash memory is additionally arranged on the second control board and used for storing basic data of probabilistic neural network learning. The first control board and the second control board are respectively connected with the electromagnet driving main circuit module through connectors; the driving module comprises a driving plate and an IGBT module which are sequentially connected; and the IGBT module is used for receiving the PWM wave output by the drive plate and outputting control current to the electromagnet for suspension control. The system also comprises supporting circuit devices such as a shared supporting capacitor and a charging and discharging circuit.

The controller comprises an external power interface for receiving a DC-440V power supply and a DC-110V power supply, wherein the DC-110V power supply is connected with the power module and used for receiving, processing and transmitting signals after voltage reduction, and the controller comprises a control module, a monitoring and diagnosing module and a sensor unit. The DC-440V power supply is used for driving the main circuit electromagnet to drive the main circuit module.

When the electromagnetic valve normally works, the electromagnet outputs 4 paths of clearance signals and 2 paths of acceleration signals to the first control board and the second control board respectively, and the FPGA is adopted to receive, process and transmit the signals. A first DSP module in the first control board receives data processed by the first FPGA to perform control quantity calculation and CAN/Ethernet communication, the calculation result is obtained and then transmitted to the first FPGA, a signal is sent to a drive board through a third FPGA module, and a PWM wave is output to drive an IGBT module to send control current.

Meanwhile, data sent by the electromagnet are simultaneously transmitted to a second FPGA module in a second control board, and the sent data are stored in a Flash memory and used as a data base for constructing and training a probabilistic neural network.

And carrying out Fourier transform on the data in a second DSP module in a second control board based on the data stored in the Flash memory, and extracting the vibration frequency. By time domain information and frequency domain information of the collected vibration data. And 6 rail state types are defined: 1) stabilizing the suspension working state; 2) the track height difference exceeds +/-3 mm; 3) the end part of the single span beam generates a larger arch angle state; 4) the electromagnet is in a destabilization rail-smashing state; 5) the controller is in a no-current output state; 6) and the electromagnet is in a dead state, and a large amount of data of each rail state is used as a network input layer and is transmitted to a network hidden layer for training a probabilistic neural network.

Therefore, the input/output relationship determined by the jth neuron of the ith type of rail state can be defined as:

i is 1,2, 6, d is the dimension of the sample, in this case, the vibration peak value, the mean value, the frequency of the electromagnet, the suspension gap peak value, the mean value and other key features, and x is_ijIs the jth center of the ith type sample. Weighted averaging is carried out on the outputs of the hidden neurons belonging to the same class in the hidden layer in the summation layer:

v_iand L is the output of the ith vehicle rail state, and the number of the ith neurons is L. The number of the neurons of the summation layer is the same as the number of the rail state categories.

The output layer takes the largest one of the summation layers as the class of output:

y＝argmax(v_i)

in actual calculation, the vector of the input layer is multiplied by a weighting coefficient, and then the vector is input into a radial basis function of the hidden layer for calculation:

Z_i＝xω_i

suppose x and ω_iAll are normalized to a unit length, and the results are subjected to radial basis calculationThereby obtaining the monitoring diagnosis result.

After the monitoring and diagnosis result is obtained, on one hand, the monitoring and diagnosis result is stored in a Flash memory to facilitate further analysis, and on the other hand, the monitoring and diagnosis result is sent to a first DSP module in a first control board to adjust the suspension output control quantity, so that real-time adaptive upgrading of the suspension system is realized.

When the first control board breaks down, the vital signal between the first DSP module and the second DSP module is interrupted, and the second control board takes over the first control board to perform suspension control calculation and CAN/Ethernet communication work. And simultaneously suspending the track state monitoring and diagnosing function of the second control board. The data interception point is the current moment when the second control panel takes over the work of the control panel. In addition, a third FPGA module on the motherboard is switched, and signals sent by the second FPGA module are output to the drive board to take over the whole work of the first control board.

In conclusion, the invention adopts the redundancy and monitoring diagnosis integrated design, the two control boards are mutually independent and have the functions of signal receiving, processing and transmitting, the DSP is adopted as the main control chip, the normal work of the suspension controller cannot be influenced when any one control board fails, and the reliability is high. And the second control panel can monitor and diagnose the rail state when normally working, and the diagnostic result can be used for further optimization analysis on the one hand, transmits on the other hand to the first control panel and is used for real-time adjustment of the controlled variable, further improves the stability and the control accuracy of the suspension controller.

While the invention has been described with reference to specific embodiments, the invention is not limited thereto, and various equivalent modifications and substitutions can be easily made by those skilled in the art within the technical scope of the invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.