Intelligent monitoring system for locomotive noise

Abstract

The present application provides a locomotive noise intelligent monitoring system, comprising: a data acquisition module for setting a sound sensor according to a locomotive track position to collect noise signals, form a noise difference vector dataset, and form a nonlinear model; a data analysis module for confirming noise samples according to track noise characteristics, and marking noise sample data after confirmation; a noise monitoring module for substituting noise sample data into the nonlinear model, identifying abnormal noise by judging the data distribution of track noise through Bayes, and discovering abnormal contact states of the track and the vehicle.

Description

Technical Field

[0001] This invention relates to the field of noise analysis, and more particularly to an intelligent locomotive noise monitoring system. Background Technology

[0002] As an indispensable choice for residents' travel, the safety and comfort of high-speed rail systems and rail transit are of paramount concern. During high-speed operation, train locomotives generate a large amount of noise due to the movement and collision of mechanical equipment, as well as the superposition and impact of airflow. This noise can also pose potential safety risks. For wheel hub vibration noise, overall vibration and noise analysis and monitoring are required, which necessitates the solution of relevant technical problems by those skilled in the art. Summary of the Invention

[0003] The present invention aims to at least solve the technical problems existing in the prior art, and in particular, innovatively proposes an intelligent locomotive noise monitoring system.

[0004] To achieve the above-mentioned objectives of the present invention, the present invention provides an intelligent locomotive noise monitoring system, comprising:

[0005] The data acquisition module is used to collect noise signals by setting up sound sensors according to the locomotive track position, forming a noise difference vector dataset, and forming a nonlinear model;

[0006] The data analysis module is used to confirm noise samples based on track noise characteristics, and then mark the noise sample data after confirmation.

[0007] The noise monitoring module is used to input noise sample data into a nonlinear model, use Bayesian analysis to determine the data distribution of track noise, identify abnormal noise, and detect abnormal contact conditions between the track and the vehicle.

[0008] In a preferred embodiment of the above technical solution, the data acquisition module includes:

[0009] In the n wheel noise acquisition data, noise data is collected by m sound sensors set on the track, and the noise difference vector dataset ΔN is obtained according to a certain time period t.

[0010] ΔN=(ΔN 11 ,...,ΔN n1 ,ΔN 1m ,...,ΔN nm ) T ;

[0011] in, Both indices n and m are positive integers, ΔN nm Let m be the noise difference obtained from the m-th sound sensor for the noise data of the nth wheel. Let N be the noise standard value of the nth wheel noise data at the mth sound sensor. nm This represents the actual noise value of the nth wheel noise data at the mth sound sensor.

[0012] In a preferred embodiment of the above technical solution, the data acquisition module includes:

[0013] Determining the locomotive-track high-frequency vibration nonlinear model U corresponding to different track and wheel contact points involves labeling each corresponding noise data with weighted noise sample data.

[0014] U=ΔN·u nm ·β;

[0015] Among them, u nm β represents the noise sample data of the nth wheel noise data at the mth sound sensor, and β is the noise sample adjustment parameter.

[0016] In a preferred embodiment of the above technical solution, the data analysis module includes:

[0017] The sample data u nm It is necessary to calculate the product of the location parameters and the abnormal noise location judgment model.

[0018] u nm = (L(s) - V(n,m))·S;

[0019] Where L(s) is the location judgment function for abnormal noise s, V(n,m) is the gain coefficient of the noise data, and S is the location judgment parameter;

[0020]

[0021] Where the constraint condition of the first formula of the judgment function is n > s, W s v represents the actual velocity of the noise data at the abnormal noise location s. s Let be the predicted velocity at the abnormal noise location s, a be the locomotive acceleration, R be the wind resistance, E be the average locomotive speed, and w be the speed at which the locomotive travels. s This indicates the location of the abnormal noise. ω is the locomotive vibration frequency, p is the locomotive operating power, and when the judgment function is 1, n≤s.

[0022] In the preferred embodiment of the above technical solution, the gain coefficient of the noise data is:

[0023] Among them, Q nm For the noise data of the nth wheel, a preset noise value is set for the mth sound sensor, N. nm This represents the actual noise value of the nth wheel noise data at the mth sound sensor.

[0024] In a preferred embodiment of the above technical solution, the noise monitoring module includes:

[0025] After calculating the nonlinear model, a nonlinear abnormal noise location model vector (U1, U2, ..., U) is formed. k ), where k is a positive integer, and when the actual noise dataset N nm and nonlinear model set Uk

[0026]

[0027] Where Θ is the preset parameter for abnormal noise, P(N) nm |Θ,U k P(Θ|U) is the likelihood function of the preset parameter Θ for abnormal noise. k P(Θ|N) represents the prior probability of the abnormal noise preset parameter Θ. nm U k P(N) represents the posterior probability of the abnormal noise with preset parameter Θ. nm |U k ) represents the marginal likelihood value of the nonlinear model Uk.

[0028] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0029] The noise vector is acquired by the data acquisition module, and the abnormal noise is screened and analyzed by the data analysis module. The distribution of the abnormal noise is verified by Bayesian method. The distribution characteristics and frequency change characteristics of the abnormal noise are recorded in real time at a periodic interval of a certain time t, so as to evaluate the overall performance of the locomotive.

[0030] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0031] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which:

[0032] Figure 1 This is a schematic diagram of the overall invention;

[0033] Figure 2 This is a flowchart of the workflow of the present invention. Detailed Implementation

[0034] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0035] like Figure 1 and 2 As shown, this invention discloses an intelligent locomotive noise monitoring system, comprising:

[0036] The data acquisition module is used to collect noise signals by setting up sound sensors according to the locomotive track position, forming a noise difference vector dataset, and forming a nonlinear model;

[0037] The data analysis module is used to confirm noise samples based on track noise characteristics, and then mark the noise sample data after confirmation.

[0038] The noise monitoring module is used to input noise sample data into a nonlinear model, use Bayesian analysis to determine the data distribution of track noise, identify abnormal noise, and detect abnormal contact conditions between the track and the vehicle.

[0039] The sound sensor is set on the side edge of the locomotive track to acquire instantaneous noise data. The acquired data is then subjected to vector analysis to identify and judge any abnormal contact between the locomotive wheels and the track, thereby preventing the risk of locomotive derailment and damage, which could cause safety hazards during vehicle transportation.

[0040] Due to the high-frequency vibration environment of the track, the velocity relationship changes during the collision response of a fixed-length wheel and track, which is superimposed on the noise acquisition data. Based on the data collection mechanism, the data is processed using a nonlinear signal offset model.

[0041] In a preferred embodiment of the above technical solution, the data acquisition module includes:

[0042] In the n wheel noise acquisition data, noise data is collected by m sound sensors set on the track, and the noise difference vector dataset ΔN is obtained according to a certain time period t.

[0043] ΔN=(ΔN 11 ,...,ΔN n1 ,ΔN 1m ,...,ΔN nm ) T ;

[0044] in, Both indices n and m are positive integers, ΔN nm Let m be the noise difference obtained from the m-th sound sensor for the noise data of the nth wheel. Let N be the noise standard value of the nth wheel noise data at the mth sound sensor. nm This represents the actual noise value of the nth wheel noise data at the mth sound sensor.

[0045] Determining the locomotive-track high-frequency vibration nonlinear model U corresponding to different track and wheel contact points involves labeling each corresponding noise data with weighted noise sample data.

[0046] U=ΔN·u nm ·β;

[0047] Among them, u nm β represents the noise sample data of the nth wheel noise data at the mth sound sensor, and β is the noise sample adjustment parameter;

[0048] Because the locomotive's track and wheels rub against each other, corresponding noise data is inevitably generated. By weighting the noise data and forming sample data of the locomotive's noise data, it is easier to carry out subsequent noise analysis operations, thereby improving the accuracy of locomotive noise data extraction.

[0049] In a preferred embodiment of the above technical solution, the data analysis module includes:

[0050] Based on the noise sample data generated by the nonlinear contact model, noise sample data analysis is performed to determine the location of abnormal noise.

[0051] The sample data u nm It is necessary to calculate the product of the location parameters and the abnormal noise location judgment model.

[0052] u nm = (L(s)-V(n,m))·S; where L(s) is the location judgment function for abnormal noise s, V(n,m) is the gain coefficient of the noise data, and S is the location judgment parameter;

[0053]

[0054] Where the constraint condition of the first formula of the judgment function is n > s, W s v represents the actual velocity of the noise data at the abnormal noise location s. s Let be the predicted velocity at the abnormal noise location s, a be the locomotive acceleration, R be the wind resistance, E be the average locomotive speed, and w be the speed at which the locomotive travels. s This indicates the location of the abnormal noise. ω is the locomotive vibration frequency, p is the locomotive operating power, and when the judgment function is 1, n≤s;

[0055] In a preferred embodiment of the above technical solution, the noise monitoring module includes:

[0056] Gain coefficient of noise data Among them, Q nm For the noise data of the nth wheel, a preset noise value is set for the mth sound sensor, N. nm This represents the actual noise value of the nth wheel noise data at the mth sound sensor.

[0057] By applying Bayes' theorem and calculating the nonlinear model, a nonlinear anomaly noise location model vector (U1, U2, ..., U...) is formed. k ), where k is a positive integer, and when the actual noise dataset N nm and nonlinear model set Uk

[0058]

[0059] Among them, the preset parameters Θ, P(N) for the abnormal noise after quantization nm |Θ,U k P(Θ|U) is the likelihood function of the preset parameter Θ for abnormal noise. k P(Θ|N) represents the prior probability of the abnormal noise preset parameter Θ. nm U k P(N) represents the posterior probability of the abnormal noise with preset parameter Θ. nm |U k ) represents the marginal likelihood value of the nonlinear model Uk.

[0060] Based on the calculated distribution, the relative position of abnormal noise on the locomotive is determined, and the corresponding probability value of locomotive abnormality is provided. The health status of the wheels is also identified. The distribution characteristics and frequency change characteristics of abnormal noise are recorded in real time at a certain time interval t, thereby evaluating the overall performance of the locomotive.

[0061] The working method of this invention is as follows:

[0062] S1, Sound sensors are set up according to the locomotive track position to collect noise signals, forming a noise difference vector dataset, and forming a nonlinear model;

[0063] S2, perform noise sample confirmation operation based on track noise characteristics, and mark the noise sample data after confirmation;

[0064] S3 substitutes the noise sample data into the nonlinear model, uses Bayesian analysis to determine the data distribution of track noise, identifies abnormal noise, and discovers abnormal contact states between the track and the vehicle.

[0065] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A locomotive noise intelligent monitoring system, characterized in that, include: The data acquisition module is used to collect noise signals by setting up sound sensors according to the locomotive track position, forming a noise difference vector dataset, and forming a nonlinear model; The data acquisition module includes: In collecting noise data from n wheel noise samples, noise data is collected using m sound sensors placed on the track, and a noise difference vector dataset is obtained based on a periodic interval of a certain time t. ; ； in, The indices n and m are both positive integers. Let m be the noise difference obtained from the m-th sound sensor for the noise data of the nth wheel. For the nth wheel noise data, the noise standard value of the mth sound sensor is given. The actual noise value of the nth wheel noise data at the mth sound sensor; The nonlinear model U for high-frequency locomotive-track vibration corresponding to different track and wheel contact points is determined by weighted noise sample data for each corresponding noise data label. ； in, For the noise data of the nth wheel, the noise sample data of the mth sound sensor is... Adjust parameters for noise samples; The data analysis module is used to confirm noise samples based on track noise characteristics, and then mark the noise sample data after confirmation. The data analysis module includes: This noise sample data It is necessary to calculate the product of the location parameters and the abnormal noise location judgment model. ； in, This is a function for determining the location of abnormal noise s. S is the gain coefficient for the noise data, and S is the position determination parameter; Among them, when the first formula of the judgment function has the following constraint condition: , The actual velocity of the noise data at the abnormal noise location s. Let be the predicted velocity at the abnormal noise location s, a be the locomotive acceleration, and R be the wind resistance. The average speed of the locomotive. This indicates the location of the abnormal noise. , Let p be the locomotive vibration frequency and p be the locomotive operating power. When the judgment function is 1, ; The noise monitoring module is used to input noise sample data into a nonlinear model, use Bayesian analysis to determine the data distribution of track noise, identify abnormal noise, and detect abnormal contact conditions between the track and the vehicle.

2. The intelligent locomotive noise monitoring system according to claim 1, characterized in that, The gain coefficient of the noise data is: ,in, Preset the noise value for the nth wheel noise data at the mth sound sensor. This represents the actual noise value of the nth wheel noise data at the mth sound sensor.

3. The intelligent locomotive noise monitoring system according to claim 1, characterized in that, The noise monitoring module includes: After calculating the nonlinear model, a nonlinear abnormal noise location model vector is formed. k is a positive integer, when the actual noise dataset and nonlinear model set Uk in, Preset parameters for abnormal noise. Preset parameters for abnormal noise The likelihood function, Preset parameters for abnormal noise The prior probability, Preset parameters for abnormal noise The posterior probability, is the marginal likelihood value of the nonlinear model Uk.

Images