CN115730516A - Contact net galloping monitoring method and system based on digital twin simulation model - Google Patents

Contact net galloping monitoring method and system based on digital twin simulation model Download PDF

Info

Publication number
CN115730516A
CN115730516A CN202211423680.9A CN202211423680A CN115730516A CN 115730516 A CN115730516 A CN 115730516A CN 202211423680 A CN202211423680 A CN 202211423680A CN 115730516 A CN115730516 A CN 115730516A
Authority
CN
China
Prior art keywords
galloping
simulation model
digital twin
amplitude
twin simulation
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202211423680.9A
Other languages
Chinese (zh)
Inventor
蒋锡健
肖晓晖
徐鸿燕
吕青松
朱久国
范卓艺
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
China Railway Siyuan Survey and Design Group Co Ltd
Original Assignee
China Railway Siyuan Survey and Design Group Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by China Railway Siyuan Survey and Design Group Co Ltd filed Critical China Railway Siyuan Survey and Design Group Co Ltd
Priority to CN202211423680.9A priority Critical patent/CN115730516A/en
Publication of CN115730516A publication Critical patent/CN115730516A/en
Pending legal-status Critical Current

Links

Images

Landscapes

  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The application discloses a method for monitoring galloping of a contact net based on a digital twin simulation model, which comprises the following steps: arranging a plurality of sensors and image acquisition devices on the contact network, and acquiring environmental data and video data of the conducting wire; acquiring the position of the central outline of the wire, and calculating the actual dominant frequency and amplitude of the wire waving; constructing a digital twin simulation model, simulating to obtain theoretical main frequency and amplitude of conductor galloping, and correcting parameter setting of the digital twin simulation model based on actual values of the main frequency and the amplitude; and inputting the amplitude into a digital twin simulation model to obtain an arrangement scheme of the anti-galloping device, and installing the anti-galloping device on the contact net. The application also discloses a system for monitoring the galloping of the contact net based on the digital twin body simulation model. The invention can monitor the galloping state in real time, predict the galloping state of the contact net according to the weather forecast, and give alarm information if the monitoring data and the prediction result exceed the warning value, so that processing measures are made in advance to avoid causing serious influence.

Description

Contact net galloping monitoring method and system based on digital twin simulation model
Technical Field
The application relates to the technical field of electrified railway contact net galloping monitoring, in particular to a method and a system for monitoring contact net galloping based on a digital twin simulation model.
Background
The contact net consists of contact wires, hanging strings, carrier cables, elastic suspension cables, struts, pantographs, insulators, cantilever supporting devices and additional wires. The waving means that the asymmetrical ice-coated power transmission conductor generates self-excited vibration with low frequency and large amplitude under the action of wind. After the overhead transmission line conducting wire is eccentrically coated with ice, the phenomenon of low-frequency and large-amplitude self-excited vibration is generated under the excitation of wind. Generally speaking, when wind blows on a wire with a non-circular cross section due to ice coating, certain aerodynamic force is generated, so that the wire is induced to generate self-excited oscillation with low frequency (about 0.1-3 Hz) and large amplitude, and the self-excited oscillation is called dancing because the wire is in a shape of flying up and down like dragon dance. The vibration amplitude of the wire rod is increased due to the galloping of the contact net, so that the mechanical damage of contact net equipment is aggravated, and great difficulty is brought to personnel maintenance; and often cause the bow net trouble, cause the electric locomotive pantograph to can not get the current normally, even cause the broken string short circuit tripping operation and interrupt the power supply, seriously threaten the safe operation of electric locomotive, thus cause huge economic loss and social influence.
At present, the main monitoring method for the waving of the overhead line system comprises the steps of manually and periodically carrying out routing inspection, building an observation station along a railway and the like. The methods have the problems of high investment cost, high labor intensity, incapability of timely finding accident hidden dangers, large difference between monitoring results and actual field data and the like.
The online monitoring system for the contact network galloping is researched and applied, the galloping state is monitored in real time, early warning information is sent out in time, disasters caused by the galloping of the contact network can be prevented, and the online monitoring system has important significance for improving the safe operation reliability of the electrified railway. The image method is a common online monitoring technology, and is characterized in that an online monitoring video device is installed on a contact net support, pictures of a contact line, a carrier cable, a main feeder line, a PW line or a return line are shot and collected, the pictures collected on site are transmitted to an online monitoring platform in a wired or wireless data transmission mode, a motion amplitude and a spectrogram of conductor galloping are obtained through certain processing, and finally alarm information is sent according to a galloping limit value. However, the camera is difficult to be arranged in a full line, and only the condition of a near wire can be observed, so that the information acquisition amount is limited.
Disclosure of Invention
Aiming at least one defect or improvement requirement in the prior art, the invention provides a method and a system for monitoring the galloping of a contact network based on a digital twin simulation model, which can monitor the galloping state in real time and send out early warning information in time, can prevent disasters caused by the galloping of the contact network and have important significance for improving the safe operation reliability of the electrified railway.
In order to achieve the above object, according to a first aspect of the present invention, there is provided a method for monitoring galloping of a catenary based on a digital twin simulation model, the method including:
arranging a plurality of sensors and a plurality of image acquisition devices on a contact network, acquiring environmental data and video data of a lead, and preprocessing the video data;
acquiring the position of the central outline of the wire according to the preprocessed video data, and calculating the actual dominant frequency and amplitude of the wire waving according to the position of the central outline of the wire;
constructing a digital twin simulation model, setting the environment data as boundary conditions of the digital twin simulation model, simulating to obtain theoretical main frequency and theoretical amplitude of conductor galloping, comparing actual values of the main frequency and the amplitude of the conductor galloping with the theoretical values, and correcting parameter setting of the digital twin simulation model based on a reinforcement learning algorithm according to a comparison result;
and adding an anti-galloping device in the corrected digital twin simulation model, inputting the amplitude of conductor galloping into the digital twin simulation model to obtain an arrangement scheme of the anti-galloping device, and installing the anti-galloping device on the contact net.
Further, the above method for monitoring galloping of the overhead contact line system based on the digital twin simulation model, wherein the preprocessing the video data specifically includes:
capturing each frame of image from the video data, and carrying out gray processing on each frame of image to obtain a gray image;
and smoothing the gray level image, and extracting the central outline of the wire in the gray level image through an edge detection algorithm to obtain a central outline image of the wire.
Further, the above method for monitoring the waving of the overhead contact system based on the digital twin simulation model, wherein the method includes the steps of obtaining the position of the central profile of the conductor according to the preprocessed video data, and calculating the dominant frequency and amplitude of the actual waving of the conductor according to the position of the central profile of the conductor, and specifically includes:
searching and positioning the position of the central outline of the wire in the central outline image of the wire to obtain the central position of the wire in each frame of image;
conducting wire center positions in each frame of image are processed in a gathering mode, and a time sequence of instantaneous displacement and instantaneous offset of the conducting wire center is generated;
and carrying out spectrum analysis on the instantaneous displacement and the instantaneous deviation in the direction vertical to the center of the wire to obtain the frequency and the amplitude of the time sequence, and calculating the actual main frequency and the amplitude of the actual waving of the wire.
Further, the method for monitoring the galloping of the overhead contact system based on the digital twin simulation model further comprises the following steps:
and comparing the actual values of the main frequency and the amplitude of the conductor galloping with the theoretical values, and correcting the digital twin body simulation model when the error between the actual values and the theoretical values of the main frequency and the amplitude of the conductor galloping exceeds a preset error threshold value.
Further, the above method for monitoring the galloping of the overhead contact system based on the digital twin simulation model, wherein the constructing of the digital twin simulation model, the setting of the environmental data as the boundary condition of the digital twin simulation model, the simulation to obtain the theoretical dominant frequency and the theoretical amplitude of the galloping of the conductor, the comparison of the actual values of the dominant frequency and the amplitude of the galloping of the conductor with the theoretical values, and the correction of the parameter setting of the digital twin simulation model based on the reinforcement learning algorithm according to the comparison result specifically includes:
setting the rigidity k of the lead in all directions along with environmental changes l And damping b l And, during the movement time T, the load vector f (T, a) of the wire varying with the wind direction and the wind force 1 ,a 2 ,…,a n ) Wherein t is ∈ [0,T],a 1 ,a 2 ,…,a n Parameters of a curve equation for describing the load vector;
under a certain weather condition, a parameter vector x = (k) is formed according to the parameters l ,b l F), setting each parameter search range V = { x } according to the simulation result of the digital twin simulation model, and discretizing V = { x }, wherein V is a discrete value 1 ,x 2 ,…,x m }。
Further, the above method for monitoring the galloping of the catenary based on the digital twin simulation model, where the digital twin simulation model is constructed, the environment data is set as the boundary condition of the digital twin simulation model, a theoretical dominant frequency and a theoretical amplitude of the galloping of the wire are obtained through simulation, actual values of the dominant frequency and the amplitude of the galloping of the wire are compared with the theoretical values, and the parameter setting of the digital twin simulation model is corrected based on a reinforcement learning algorithm according to the comparison result, further includes:
clustering the environment data according to the weather state, inputting the clustered environment data into a digital twin simulation model, and training the digital twin simulation model to correct errors;
determining that the digital twin simulation model has no Markov property, searching in the parameter search range V, and selecting one group of parameter vectors x in the parameter vectors under a certain weather state when the j +1 th search is performed i Carrying out simulation calculation according to the root mean square error e between the output result and the actual measurement result of the digital twin body simulation model in the time T rmse Calculating the current parameter vector x for the j +1 search i Is awarded r i j+1
According to the reward r i j+1 Updating value function q of digital twin simulation model i For the value function at j +1 search
Figure BDA0003943953900000041
Wherein α is the learning rate; after a number of training, the value function q i And converging, wherein the parameter vector with the maximum function is the finite element model parameter which best accords with the actual situation.
Further, the method for monitoring the galloping of the overhead contact system based on the digital twin simulation model further comprises the following steps:
setting greedy factor epsilon [0,1 ∈ ]]Generating a random number rand between 0 and 1 during each step of search, and if rand is greater than greedy factor epsilon, performing function q according to value of each parameter vector i To select the parameter vector x of the current round i And if rand is smaller than the greedy factor epsilon, selecting the parameter vector randomly.
Further, the method for monitoring the galloping of the overhead contact system based on the digital twin simulation model further comprises the following steps:
when parameter x in the parameter vector i When the corresponding root mean square error is the minimum root mean square error in all the searched parameters, the forward reward is given to 1, otherwise, the penalty-1 is given, namely
Figure BDA0003943953900000042
Wherein x is best Refers to the historical optimal solution in the same environment.
Further, the above method for monitoring the galloping of the overhead contact system based on the digital twin simulation model, wherein the anti-galloping device is added to the corrected digital twin simulation model, and the amplitude of the conductor galloping is input to the digital twin simulation model to obtain the arrangement scheme of the anti-galloping device, specifically comprises:
the anti-galloping device mounting position p and the number k are set as target parameters x ' = (p, k) of reinforcement learning, each parameter search range V ' = { x ' } is set empirically, and is discretized into V ' = { x ' 1 ,x' 2 ,…,x' m };
Determining that no Markov property exists in the digital twin simulation model, searching in the parameter search range V ', and selecting a parameter vector x ' in a certain weather state when the j +1 th search is carried out ' i Carrying out simulation calculation, and calculating a current parameter vector x 'according to the obtained galloping amplitude h of the cable' i Is awarded
Figure BDA0003943953900000051
According to the reward
Figure BDA0003943953900000052
Updating value function q' i For the value function at the j +1 th search
Figure BDA0003943953900000053
Wherein alpha ' is a learning rate, and after a plurality of training, a value function q ' is obtained ' i And converging, wherein the parameter vector with the maximum function is the optimal scheme for arranging the anti-galloping device.
According to the second aspect of the invention, the invention also provides a digital twin simulation model-based overhead line system galloping monitoring system, which comprises a physical layer, a sensing layer, a calculation layer and an application layer which are sequentially connected;
the physical layer comprises a plurality of sensors and a plurality of image acquisition devices which are arranged on a contact net and used for acquiring environmental data and video data of a lead;
the perception layer is used for receiving the video data and preprocessing the video data; acquiring the position of the central outline of the wire according to the preprocessed video data, and calculating the actual dominant frequency and amplitude of the wire waving according to the position of the central outline of the wire;
the calculation layer comprises a digital twin simulation model and is used for setting the environment data as boundary conditions of the digital twin simulation model, simulating to obtain theoretical main frequency and theoretical amplitude of conductor galloping, comparing actual values of the main frequency and the amplitude of the conductor galloping with the theoretical values, and correcting parameter setting of the digital twin simulation model based on a reinforcement learning algorithm according to a comparison result;
the application layer is used for adding an anti-galloping device in the corrected digital twin simulation model, and inputting the amplitude of conductor galloping into the digital twin simulation model to obtain the arrangement scheme of the anti-galloping device.
In general, compared with the prior art, the above technical solution contemplated by the present invention can achieve the following beneficial effects:
(1) According to the method and the system for monitoring the galloping of the contact net based on the digital twin simulation model, the digital twin simulation model is constructed to obtain the theoretical main frequency and the theoretical amplitude of the galloping of the wire, the digital twin simulation model is trained through reinforcement learning, and the parameter setting of the digital twin simulation model is corrected, so that the theoretical value obtained through simulation is infinitely close to the actual value, the model precision is improved, the galloping state can be monitored in real time, early warning information can be sent out in time, disasters caused by the galloping of the contact net can be prevented, and the method and the system have important significance for improving the safe operation reliability of the electrified railway. The galloping state can be monitored in real time, the galloping state of the overhead line system is predicted and deduced according to weather forecast, if the monitoring data and the prediction result exceed a warning value, alarm information is given, processing measures are made in advance, and serious influence is avoided.
(2) By adopting the method and the system for monitoring the galloping of the contact net based on the digital twin simulation model, which are provided by the invention, the problems of limited and possibly invalid information acquired by an image method are solved by adopting a digital twin technology based on reinforcement learning, and the visual monitoring can be carried out on the galloping of the contact net in the whole section.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
Fig. 1 is a schematic flow diagram of a method for monitoring galloping of a catenary based on a digital twin simulation model according to an embodiment of the present application;
fig. 2 is a schematic structural principle diagram of a system for monitoring galloping of a catenary based on a digital twin simulation model provided by the embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
The terms "first," "second," "third," and the like in the description and claims of this application and in the above-described drawings are used for distinguishing between different objects and not for describing a particular order. Furthermore, the terms "include" and "have," as well as any variations thereof, are intended to cover a non-exclusive inclusion. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those steps or elements listed, but may alternatively include other steps or elements not listed, or inherent to such process, method, article, or apparatus.
On one hand, the application provides a contact net galloping detection method based on reinforcement learning, and fig. 1 is a schematic flow diagram of the contact net galloping monitoring method based on a digital twin simulation model provided by the embodiment of the application. Referring to fig. 1, the method includes the following steps:
(1) Arranging a plurality of sensors and a plurality of image acquisition devices on a contact network, acquiring environmental data and video data of a lead, and preprocessing the video data;
the contact net includes: contact wire, dropper, carrier cable, elastic sling, support, pantograph, insulator, cantilever support device and additional wire. Wherein, the sensor includes: wind speed and direction sensor, temperature sensor, humidity sensor, tension sensor, etc. and the image acquiring device may be a camera, etc. The environmental data at least comprises wind speed, wind direction, temperature, humidity, tension on the wire and the like.
Capturing each frame of image from the video data, and carrying out gray scale processing on each frame of image (originally, color) to obtain a gray scale image; and smoothing the gray level image, and extracting the central outline of the wire in the gray level image through an edge detection algorithm to obtain a central outline image of the wire.
(2) Acquiring the position of the central outline of the wire according to the preprocessed video data, and calculating the actual dominant frequency and amplitude of the wire waving according to the position of the central outline of the wire;
specifically, searching and positioning the position of the central outline of the wire in the central outline image of the wire to obtain the central position of the wire in each frame of image; conducting wire center positions in each frame of image are processed in a gathering mode, and a time sequence of instantaneous displacement and instantaneous offset of the conducting wire center is generated; and carrying out spectrum analysis on the instantaneous displacement and the instantaneous deviation in the direction vertical to the center of the wire, transforming the time domain signal into a frequency domain to obtain the frequency and the amplitude of a time sequence, and calculating the actual main frequency and the amplitude of the actual waving of the wire.
(3) Constructing a digital twin simulation model, setting the environment data as boundary conditions of the digital twin simulation model, simulating to obtain theoretical main frequency and theoretical amplitude of conductor galloping, comparing actual values of the main frequency and the amplitude of the conductor galloping with the theoretical values, and correcting parameter setting of the digital twin simulation model based on a reinforcement learning algorithm according to a comparison result;
specifically, a 3D digital twin simulation model is created on the ANSYS platform. The conductive cable is simulated by using an Euler beam unit, six degrees of freedom need to be set for the conductive cable, wherein 2 bending degrees of freedom are released, namely 2 bending degrees of freedom are not constrained, and 3 translation degrees of freedom are reserved, namely 3 translation degrees of freedom respectively correspond to the directions of x, y and z in a three-dimensional coordinate system and 1 torsion degree of freedom; the structural size of the conductive cable is determined according to design drawings. According to the environmental data: setting boundary conditions of a digital twin body simulation model according to wind speed, wind direction, temperature, humidity, ice coating type, ice coating thickness, tension on the wire and the like, and performing simulation to obtain theoretical dominant frequency and theoretical amplitude of wire galloping.
And comparing the actual values of the main frequency and the amplitude of the conductor galloping with the theoretical values, and correcting the digital twin body simulation model when the error between the actual values and the theoretical values of the main frequency and the amplitude of the conductor galloping exceeds a preset error threshold value.
Specifically, the rigidity k of the wire in each direction according to environmental changes (temperature, icing, type of icing, thickness of icing, and the like) is set l And damping b l And a load vector f (T, a) of the wire changing with wind direction and wind force during a period of time T (T can be obtained by Fourier transform of cable displacement or load data) capable of representing the cable motion state 1 ,a 2 ,…,a n ) Wherein t is ∈ [0,T],a 1 ,a 2 ,…,a n Parameters of a curve equation for describing the load vector;
under a certain weather condition, a parameter vector x = (k) is formed according to the parameters l ,b l F), setting each parameter searching range V = { x } according to the simulation result of the digital twin simulation model, and discretizing V = { x } according to the parameter searching range V = { x }, wherein the parameters are set in the digital twin simulation model 1 ,x 2 ,…,x m };
Because the data volume of the sensor is large, when a reinforcement learning data set is manufactured, the environmental data is firstly clustered according to the weather state, then the same amount of data is randomly sampled in each class for model training, the digital twin body simulation model is trained to correct errors, and the data utilization rate can be improved (K-means clustering is used, and the clustering number is determined by Bayesian information criterion BIC).
The training process of the digital twin body simulation model specifically comprises the following steps:
the meter assumes that the digital twin simulation model does not have markov properties, considering only the reward of a single step search. Searching in the parameter searching range V, and when the (j + 1) th time of searching is carried out, selecting one group of parameter vectors x in the parameter vectors under a certain weather state i Carrying out simulation calculation according to the root mean square error e between the output result and the actual measurement result of the digital twin body simulation model in the time T rmse Determining the current parameter vector x for the j +1 search i Is awarded r i j+1
According to the reward r i j+1 Updating value function q of digital twin simulation model i For the value function at the j +1 th search
Figure BDA0003943953900000091
Wherein alpha is a learning rate and reflects the speed of accumulated experience of a value function from the current round of search; value function q i The coincidence degree of the current parameters and the real parameters of the model is reflected. After a number of training, the value function q i And converging, wherein the parameter vector with the maximum function is the finite element model parameter which best accords with the actual situation.
Wherein the parameter vector x i The selection method comprises the following steps:
value function q i The coincidence degree of the current parameters and the real parameters of the digital twin simulation model can be reflected. However, only the value function q is relied on in the early stage of training i To select a parameter vector x i It is possible to get into local optima. Therefore, the parameter vector is selected by referring to the function q i There is also a certain randomnessTo prevent falling into local optima. Setting greedy factor epsilon [0,1 ]]Generating a random number rand between 0 and 1 during each step of search, and if rand is greater than greedy factor epsilon, performing function q according to value of each parameter vector i To select the parameter vector x of the current round i Otherwise, the parameter vector is randomly selected.
Wherein the prize r i j+1 The calculation method of (2) is as follows: when a certain parameter x i When all the parameters searched in the past have better performance, i.e. when a certain parameter x i And if the corresponding root mean square error is the minimum root mean square error in all the searched parameters, the searched parameters are rewarded positively, otherwise, a penalty is given.
Figure BDA0003943953900000092
In the formula, x best Refers to the historical optimal solution in the same environment.
(4) And adding an anti-galloping device in the corrected digital twin simulation model, inputting the amplitude of conductor galloping into the digital twin simulation model to obtain an arrangement scheme of the anti-galloping device, and installing the anti-galloping device on the contact net.
Wherein, anti-galloping device includes: the utility model provides a conductor spacer, spoiler, prevent waving whip, integral eccentric weight and two pendulum anti-galloping ware etc. its form has the mixed use of single circuit anti-galloping device, the anti-galloping device of specific circuit and multiple anti-galloping device.
Select certain types of anti-galloping devices, such as weight anti-galloping devices, turbulence anti-galloping devices, double-pendulum anti-galloping devices, and the like.
The anti-galloping device mounting position p and the number k are set as target parameters x ' = (p, k) of reinforcement learning, each parameter search range V ' = { x ' } is set empirically, and is discretized into V ' = { x ' 1 ,x' 2 ,…,x' m };
Assuming that the model does not have a Markov nature, only the reward of a single step search is considered. Searching in the parameter searching range V ', and when the (j + 1) th search is carried out, selecting a parameter vector x ' under a certain weather state ' i Carrying out simulation calculation, and calculating a current parameter vector x 'according to the obtained galloping amplitude h of the cable' i Is awarded
Figure BDA0003943953900000101
According to the reward
Figure BDA0003943953900000102
Updating value function q' i For the value function at the j +1 th search
Figure BDA0003943953900000103
Wherein, alpha' is a learning rate and reflects the speed of accumulated experience of a value function from the current round of search; value function q' i The coincidence degree of the current parameters and the real parameters of the model is reflected. After a plurality of training times, the function q 'is valued' i And converging, namely, the parameter vector with the maximum function is the optimal anti-galloping device arrangement scheme, and installing the anti-galloping device on the contact network according to the anti-galloping device arrangement scheme.
(5) And predicting and deducing the waving state of the overhead line system according to the weather forecast. And setting a main frequency warning value and an amplitude warning value of conductor waving, and giving warning information if the monitoring data and the prediction result exceed the warning values to prompt workers to take treatment measures in advance so as to avoid causing serious influence.
On the other hand, the application provides a monitoring system for monitoring galloping of a catenary based on a digital twin simulation model, and fig. 2 is a schematic diagram of a structural principle of the monitoring system for monitoring galloping of a catenary based on the digital twin simulation model provided by the embodiment of the application.
The physical layer is a physical object of digital twin modeling, namely a real contact network system, and comprises the actual composition of the contact network, various sensing acquisition devices, image acquisition devices and anti-galloping devices installed in the contact network system. The actual composition of the contact line comprises: contact wire, dropper, carrier cable, elastic sling, support, pantograph, insulator, cantilever support device and additional wire. The sensing acquisition device includes: wind speed and direction sensor, temperature sensor, humidity sensor, tension sensor, camera, etc. In order to prevent the transmission line from galloping, the anti-galloping device which has the function of inhibiting the line galloping comprises: spacer, spoiler, anti-galloping whip, integral eccentric weight and double pendulum anti-galloping ware etc. its form has the anti-galloping device of single circuit, specific circuit and the mixed use of multiple anti-galloping device.
The sensing layer is based on sensing and data acquisition technologies, and environmental data and video data of a lead are acquired through a plurality of sensors and a plurality of image acquisition devices. The environmental data includes: wind speed, wind direction, temperature, humidity in the contact net system, tension to which the wire is subjected, type of ice coating of the wire, and thickness of the ice coating. The sensing layer is used for receiving video data and preprocessing the video data; and acquiring the position of the central outline of the wire according to the preprocessed video data, and calculating the actual dominant frequency and amplitude of the wire waving according to the position of the central outline of the wire.
The transmission layer is an Ethernet formed on the basis of the switch and the optical fiber and used for uploading the data collected by the sensing layer to the data layer at a high speed.
The data layer stores the acquired data in the server, realizes integration, fusion and storage of multi-source data and information through the data processing center, and distributes the integrated and fused multi-source data and information to the calculation layer and the application layer. The multi-source data and information includes: the real-time state data of the physical layer collected by the sensing layer, the simulation model result of the calculation layer, and the prediction deduction result, the state analysis result and the decision operation information of the application layer.
The computing layer integrates core contents such as a digital twin simulation model, a reinforcement learning algorithm, simulation calculation and the like, and a hybrid driving mode of data driving and model driving is adopted. The digital twin body simulation model sets the environmental data as the boundary conditions of the digital twin body simulation model, the theoretical dominant frequency and the theoretical amplitude of conductor galloping are obtained through simulation, the actual values of the dominant frequency and the amplitude of the conductor galloping are compared with the theoretical values, and the parameter setting of the digital twin body simulation model is automatically corrected based on a reinforcement learning algorithm according to the comparison result. The computing layer has the capability of sensing the system in real time, retains the physical characteristics of the model and improves the precision of the model.
The application layer carries out prediction deduction, state analysis and decision operation according to the simulation result and the measured data of the calculation layer, the prediction deduction result, the state analysis result and the decision operation information are stored in the data layer, and the prediction deduction result, the state analysis result and the decision operation information can directly act on the physical layer to carry out decision guidance.
And the prediction deduction completes the wire state prediction in a future period of time according to the simulation model and the meteorological forecast. The state analysis includes: the method comprises the steps of conducting wire waving amplitude and mode analysis, icing thickness development trend analysis and conducting wire tension analysis.
It should be noted that, for simplicity of description, the above-mentioned method embodiments are described as a series of acts or combination of acts, but those skilled in the art will recognize that the present application is not limited by the order of acts described, as some steps may occur in other orders or concurrently depending on the application. Further, those skilled in the art should also appreciate that the embodiments described in the specification are preferred embodiments and that the acts and modules referred to are not necessarily required in this application.
In the foregoing embodiments, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.
The above description is only an exemplary embodiment of the present disclosure, and the scope of the present disclosure should not be limited thereby. That is, all equivalent changes and modifications made in accordance with the teachings of the present disclosure are intended to be included within the scope of the present disclosure. Embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and that any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the present invention.

Claims (10)

1. A method for monitoring the galloping of a contact net based on a digital twin simulation model is characterized by comprising the following steps:
arranging a plurality of sensors and a plurality of image acquisition devices on a contact network, acquiring environmental data and video data of a lead, and preprocessing the video data;
acquiring the position of the central outline of the wire according to the preprocessed video data, and calculating the actual dominant frequency and amplitude of the wire waving according to the position of the central outline of the wire;
constructing a digital twin simulation model, setting the environment data as boundary conditions of the digital twin simulation model, simulating to obtain theoretical main frequency and theoretical amplitude of conductor galloping, comparing actual values of the main frequency and the amplitude of the conductor galloping with the theoretical values, and correcting parameter setting of the digital twin simulation model based on a reinforcement learning algorithm according to a comparison result;
and adding an anti-galloping device in the corrected digital twin simulation model, inputting the amplitude of conductor galloping into the digital twin simulation model to obtain an arrangement scheme of the anti-galloping device, and installing the anti-galloping device on the contact net.
2. The method for monitoring the galloping of the catenary based on the digital twin simulation model according to claim 1, wherein the preprocessing the video data specifically comprises:
capturing each frame of image from the video data, and carrying out gray processing on each frame of image to obtain a gray image;
and smoothing the gray level image, and extracting the central outline of the wire in the gray level image through an edge detection algorithm to obtain a central outline image of the wire.
3. The method for monitoring the contact net galloping based on the digital twin simulation model as claimed in claim 2, wherein the step of obtaining the position of the central profile of the conductor according to the preprocessed video data and calculating the dominant frequency and the amplitude of the actual galloping of the conductor according to the position of the central profile of the conductor comprises the following steps:
searching and positioning the position of the central outline of the wire in the central outline image of the wire to obtain the central position of the wire in each frame of image;
conducting wire center positions in each frame of image are processed in a gathering mode, and a time sequence of instantaneous displacement and instantaneous offset of the conducting wire center is generated;
and carrying out spectrum analysis on the instantaneous displacement and the instantaneous deviation in the direction vertical to the center of the wire to obtain the frequency and the amplitude of the time sequence, and calculating the actual main frequency and the amplitude of the actual waving of the wire.
4. The method for monitoring the galloping of the overhead line system based on the digital twin simulation model as claimed in claim 1, further comprising:
and comparing the actual values of the main frequency and the amplitude of the conductor galloping with the theoretical values, and correcting the digital twin body simulation model when the error between the actual values and the theoretical values of the main frequency and the amplitude of the conductor galloping exceeds a preset error threshold value.
5. The method for monitoring the galloping of the overhead contact system based on the digital twin simulation model according to claim 4, wherein the digital twin simulation model is constructed, the environment data is set as the boundary condition of the digital twin simulation model, the theoretical dominant frequency and the theoretical amplitude of the galloping of the lead are obtained through simulation, the actual values of the dominant frequency and the actual values of the amplitude of the galloping of the lead are compared with the theoretical values, and the parameter setting of the digital twin simulation model is corrected based on a reinforcement learning algorithm according to the comparison result, and the method specifically comprises the following steps:
setting the rigidity k of the lead in all directions along with the change of the environment l And damping b l And, during the movement time T, the load vector f (T, a) of the wire varying with the wind direction and the wind force 1 ,a 2 ,…,a n ) Wherein t is ∈ [0,T],a 1 ,a 2 ,…,a n Parameters of a curve equation for describing the load vector;
under a certain weather condition, a parameter vector x = (k) is formed according to the parameters l ,b l F), setting each parameter searching range V = { x } according to the simulation result of the digital twin simulation model, and discretizing V = { x } according to the parameter searching range V = { x }, wherein the parameters are set in the digital twin simulation model 1 ,x 2 ,…,x m }。
6. The method for monitoring the galloping of the overhead contact system based on the digital twin simulation model as claimed in claim 5, wherein the constructing of the digital twin simulation model, the setting of the environmental data as the boundary condition of the digital twin simulation model, the simulation of the environmental data to obtain the theoretical dominant frequency and the theoretical amplitude of the galloping of the conductors, the comparison of the actual values of the dominant frequency and the amplitude of the galloping of the conductors with the theoretical values, and the correction of the parameter setting of the digital twin simulation model based on the reinforcement learning algorithm according to the comparison result further comprises:
clustering the environment data according to the weather state, inputting the clustered environment data into a digital twin simulation model, and training the digital twin simulation model to correct errors;
determining that the digital twin simulation model has no Markov property, searching in the parameter search range V, and selecting one group of parameter vectors x in the parameter vectors under a certain weather state when searching for the (j + 1) th time i Carrying out simulation calculation according to the root mean square error e between the output result and the actual measurement result of the digital twin body simulation model in the time T rmse Calculating the current parameter vector x for the j +1 th search i Is awarded r i j+1
According to the reward r i j+1 Updating value function q of digital twin simulation model i For the value function at the j +1 th search
Figure FDA0003943953890000031
Wherein α is the learning rate; after a number of training, the value function q i And converging, wherein the parameter vector with the maximum function is the finite element model parameter which best accords with the actual situation.
7. The method for monitoring the galloping of the overhead line system based on the digital twin simulation model as claimed in claim 6, further comprising:
setting greedy factor epsilon [0,1 ∈ ]]Generating a random number rand between 0 and 1 during each step of search, and if rand is greater than a greedy factor epsilon, performing a function q according to the value of each parameter vector i To select the parameter vector x of the current round i And if rand is smaller than the greedy factor epsilon, selecting the parameter vector randomly.
8. The method for monitoring galloping of the catenary based on the digital twin simulation model as claimed in claim 7, further comprising:
when parameter x in the parameter vector i When the corresponding root mean square error is the minimum root mean square error in all the searched parameters, the forward reward is given to 1, otherwise, the penalty-1 is given, namely
Figure FDA0003943953890000032
Wherein x is best Refers to the historical optimal solution in the same environment.
9. The method for monitoring the galloping of the overhead line system based on the digital twin simulation model as claimed in claim 8, wherein the anti-galloping device is added to the corrected digital twin simulation model, and the amplitude of the conductor galloping is input into the digital twin simulation model to obtain the arrangement scheme of the anti-galloping device, and specifically comprises the following steps:
taking the installation position p and the number k of the anti-galloping devices asTarget parameters x ' = (p, k) for reinforcement learning, and each parameter search range V ' = { x ' } is empirically set and discretized into V ' = { x ' 1 ,x' 2 ,…,x' m };
Determining that no Markov property exists in the digital twin simulation model, searching in the parameter search range V ', and selecting a parameter vector x ' in a certain weather state when the j +1 th search is carried out ' i Carrying out simulation calculation, and calculating a current parameter vector x 'according to the obtained galloping amplitude h of the cable' i Is awarded
Figure FDA0003943953890000041
According to the reward
Figure FDA0003943953890000042
Updating value function q' i For the value function at j +1 search
Figure FDA0003943953890000043
Wherein alpha ' is a learning rate, and after a plurality of training, a value function q ' is obtained ' i And converging, wherein the parameter vector with the maximum function is the optimal scheme for arranging the anti-galloping device.
10. A catenary galloping monitoring system based on a digital twin simulation model is characterized by comprising a physical layer, a sensing layer, a calculation layer and an application layer which are sequentially connected;
the physical layer comprises a plurality of sensors and a plurality of image acquisition devices which are arranged on a contact net and used for acquiring environmental data and video data of a lead;
the perception layer is used for receiving the video data and preprocessing the video data; acquiring the position of the central outline of the wire according to the preprocessed video data, and calculating the actual dominant frequency and amplitude of the wire waving according to the position of the central outline of the wire;
the calculation layer comprises a digital twin simulation model and is used for setting the environment data as boundary conditions of the digital twin simulation model, simulating to obtain theoretical main frequency and theoretical amplitude of conductor galloping, comparing actual values of the main frequency and the amplitude of the conductor galloping with the theoretical values, and correcting parameter setting of the digital twin simulation model based on a reinforcement learning algorithm according to a comparison result;
the application layer is used for adding an anti-galloping device in the corrected digital twin simulation model, inputting the amplitude of conductor galloping into the digital twin simulation model to obtain the arrangement scheme of the anti-galloping device, and installing and implementing the anti-galloping device on the contact network.
CN202211423680.9A 2022-11-15 2022-11-15 Contact net galloping monitoring method and system based on digital twin simulation model Pending CN115730516A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211423680.9A CN115730516A (en) 2022-11-15 2022-11-15 Contact net galloping monitoring method and system based on digital twin simulation model

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211423680.9A CN115730516A (en) 2022-11-15 2022-11-15 Contact net galloping monitoring method and system based on digital twin simulation model

Publications (1)

Publication Number Publication Date
CN115730516A true CN115730516A (en) 2023-03-03

Family

ID=85296523

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211423680.9A Pending CN115730516A (en) 2022-11-15 2022-11-15 Contact net galloping monitoring method and system based on digital twin simulation model

Country Status (1)

Country Link
CN (1) CN115730516A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116244975A (en) * 2023-05-11 2023-06-09 众芯汉创(北京)科技有限公司 Transmission line wire state simulation system based on digital twin technology

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116244975A (en) * 2023-05-11 2023-06-09 众芯汉创(北京)科技有限公司 Transmission line wire state simulation system based on digital twin technology
CN116244975B (en) * 2023-05-11 2023-07-25 众芯汉创(北京)科技有限公司 Transmission line wire state simulation system based on digital twin technology

Similar Documents

Publication Publication Date Title
Balluff et al. Short term wind and energy prediction for offshore wind farms using neural networks
CN110009037B (en) Short-term engineering wind speed prediction method and system based on physical information coupling
CN112362274B (en) Method and system for monitoring, early warning and evaluating vortex-induced vibration of long-span bridge in operation period
CN105139274A (en) Power transmission line icing prediction method based on quantum particle swarm and wavelet nerve network
WO2023103278A1 (en) Online monitoring method for position information of conducting wire of power transmission line based on electromagnetic signal of ground wire
CN110319990B (en) Bridge dynamic deflection monitoring method based on inclinometer optimized arrangement
CN112398230B (en) Power transmission line galloping monitoring system and method
CN115730516A (en) Contact net galloping monitoring method and system based on digital twin simulation model
CN108320052A (en) Transmission line status prediction technique under meteorological disaster and device
CN109102101A (en) Wind speed prediction method and system for wind power plant
CN112598883B (en) Power transmission line lightning-fall probability early warning method and early warning system based on Bayesian network
CN107817423B (en) Power transmission line motion early warning method
CN113320445B (en) Online monitoring and intelligent hidden danger and fault distinguishing and early warning system for contact network
CN116324639A (en) Monitoring and predicting method and device for traction power supply system of rail transit
CN105117538A (en) Method for warning waving of power transmission channels
CN109100703A (en) A kind of transmission line of electricity dangerous point detection method and device
Molinar et al. From data points to ampacity forecasting: Gated Recurrent Unit networks
CN114912355A (en) Method and device for predicting short-term icing of power transmission line and storage medium
CN117313537B (en) Data-driven icing prediction method and system
CN113674512A (en) Online monitoring and early warning system and method for live cross-over construction site
Zhao et al. Full scale experiment for vibration analysis of ice-coated bundled-conductor transmission lines
CN110645939A (en) System and method for dynamically predicting relative vertical distance between two wires
CN116128307A (en) Transmission line galloping early warning system based on multi-metadata transmission network
CN116046076A (en) Online detection system for power transmission line galloping based on machine vision technology
CN109508808A (en) A kind of arc sag prediction technique based on PSO-BP neural network

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