Search

CN-122021026-A - Method for predicting working life of mining electric shovel by adopting high-torque load extrapolation model

CN122021026ACN 122021026 ACN122021026 ACN 122021026ACN-122021026-A

Abstract

The invention belongs to the technical field of life prediction and reliability analysis of electromechanical equipment, and particularly relates to a method for predicting the working life of a mining electric shovel by adopting a high-torque load extrapolation model, which comprises the following steps of S1, load data acquisition and preprocessing; the method comprises the steps of constructing a nuclear density estimation model, constructing a large-torque directional expansion model and expanding samples, constructing a load spectrum, extrapolating and compiling a load spectrum, and verifying an extrapolating result, wherein the problems of sparse large-torque sample number and low extrapolating precision in the extrapolating process caused by the defect of large-torque data in the software acquisition process in the extrapolating process of the motor torque of the electric shovel in the prior art are solved.

Inventors

  • ZHANG DONGGUANG
  • ZHAI QIANG
  • Yang Qining
  • WU YALI

Assignees

  • 太原理工大学

Dates

Publication Date
20260512
Application Date
20260131

Claims (10)

  1. 1. A method for predicting the working life of a mining electric shovel by adopting a high-torque load extrapolation model is characterized by comprising the following steps: S1, load data acquisition and preprocessing, namely acquiring actual measurement torque time sequence data of a transmission mechanism motor of a mining electric shovel, and preprocessing the actual measurement torque time sequence data to obtain a rain flow matrix of a torque load; S2, constructing a nuclear density estimation model, namely selecting an Epanechnikov kernel function based on the rain flow matrix and determining the optimal bandwidth of the nuclear density estimation by a ROT thumb rule insertion method to construct the nuclear density estimation model; S3, constructing a large-torque directional expansion model, namely constructing a sampling frame based on a Metropolis-Hastings algorithm by taking the nuclear density estimation model as target distribution, setting initial parameters containing a large-torque judgment standard, generating an expansion sample in a directional manner through iterative sampling, screening out a large-torque expansion sample conforming to the large-torque judgment standard from the expansion sample, and combining the large-torque expansion sample with an original sample in the rain flow matrix to form an expanded complete sample set; s4, extrapolation and programming of a load spectrum, namely, extrapolation is carried out through the kernel density estimation model based on the complete sample set to obtain an extended rain flow matrix covering the whole life cycle; S5, predicting the fatigue life, namely taking the program load spectrum as input, constructing a fatigue life prediction model of a motor transmission shaft for carrying out life simulation analysis, and predicting the working life of the mining electric shovel by combining the actual working parameters of the electric shovel; And S6, verifying an extrapolation result, namely extracting the statistical characteristic of the load based on the extended rain flow matrix, and comparing and verifying the statistical characteristic with the statistical characteristic of the actually measured torque time sequence data to ensure that the error of a large torque area meets the preset precision requirement.
  2. 2. The method for predicting the working life of the mining shovel by adopting the high-torque load extrapolation model according to claim 1 is characterized in that in the step S1, the transmission mechanism comprises at least one of a lifting mechanism, a pushing mechanism, a slewing mechanism and a traveling mechanism, the acquired actual measured torque time sequence data cover a range of 1.2 times to 2.0 times of rated torque of a motor of the transmission mechanism, the acquisition duration covers at least three typical working conditions, and the continuous acquisition duration is not less than 48 hours.
  3. 3. The method for predicting the service life of a mining shovel by adopting a high-torque load extrapolation model according to claim 1, wherein in step S1, the preprocessing comprises the following steps: the abnormal data screening comprises the steps of deleting two types of abnormal data, wherein one type is data with missing or discontinuous time stamps, and the other type is data with parameter values exceeding a preset physical range; Invalid loop filtering, namely judging and eliminating non-operation section data based on a rotating speed threshold value and a current threshold value of the motor of the transmission mechanism; And (3) data splicing and a rain flow matrix construction, namely splicing the data of each effective operation section according to time sequence, and constructing the rain flow matrix based on a rain flow counting method.
  4. 4. The method for predicting the service life of the mining shovel by adopting the high-torque load extrapolation model according to claim 1, wherein in the step S2, the formula of the nuclear density estimation model is as follows: ; wherein f (x, y) represents a joint probability density estimated value at a position where an initial value of torque load is x and a final value is y, n is an effective torque circulation group number in the rain flow matrix , ) For the initial value and the termination value of the ith group of loops in the rain flow matrix, i=1, 2..n, h is the optimal bandwidth, K (·) is the Epanechnikov kernel function, lambda i is the adaptive correction coefficient, and the calculation formula is Wherein alpha is a sensitivity coefficient, epsilon is a minimum value for avoiding zero of denominator, and the value range of epsilon is 10 -6 ~10 -8 .
  5. 5. The method for predicting the service life of the mining shovel by adopting the high-torque load extrapolation model according to claim 4, wherein in the step S2, the optimal bandwidth h is calculated by a ROT thumb rule insertion method, and a calculation formula is as follows: ; Wherein A is the two-dimensional minimum of the Epanechnikov kernel function, And (3) as the standard deviation of the torque data in the rain flow matrix, IQR is the standard quartile range of the torque data in the rain flow matrix, and n is the number of effective torque circulation groups in the rain flow matrix.
  6. 6. The method for predicting the service life of a mining shovel by adopting a high-torque load extrapolation model according to claim 1, wherein in step S3, the initial parameters comprise: an initial sample x 0 is randomly selected from the rain flow matrix, and the mean deviation of the initial sample x 0 and the measured torque time sequence data is not more than 5%; Proposal distribution adopts the current iteration sample as the center and covariance matrix as the center Wherein h is the optimal bandwidth; The high torque judgment standard is that the torque value exceeds the rated torque of the motor and does not exceed the maximum allowable torque of the motor; iteration parameters: the total number of iterations is 10000 to 50000, wherein the first 10% to 30% of iterations are discarded as burn-in periods.
  7. 7. The method for predicting the service life of a mining shovel by adopting a high-torque load extrapolation model according to claim 6, wherein in the step S3, the iterative sampling process comprises the following steps: (a) Based on the proposal distribution Extracting an alternative sample Where t is the index of the number of iterations, Representing the samples after the t-1 th iteration, And (3) with The two-dimensional vectors are respectively corresponding to the initial value and the termination value of the load cycle; (b) Calculating the probability of acceptance Wherein f (·) is a probability density function of the kernel density estimation model; (c) Generating a random number uniformly distributed in the [0,1] interval If (if) Accept the alternative sample Order-making If not, reject, let ; Repeating steps (a) - (c) until the total number of iterations is reached, and screening samples meeting the high-torque judgment standard from the effective iteration samples as the high-torque expansion samples.
  8. 8. The method for predicting the service life of a mining shovel by adopting a high-torque load extrapolation model according to claim 1, wherein in step S4, the programming of the program load spectrum comprises the following steps: Extracting torque amplitude and corresponding accumulated frequency data from the extended rain flow matrix; Fitting the torque amplitude and the accumulated frequency data by using a power function to obtain a fitting curve; Determining a torque amplitude corresponding to the accumulated frequency which is a preset extremely low cycle number from the fitting curve, and taking the torque amplitude as a maximum-level amplitude reference of a program load spectrum; Calculating the amplitude of each stage according to the 8-stage amplitude ratio coefficient of Conover criterion and the maximum-stage amplitude standard; And calculating the accumulated frequency corresponding to each level of amplitude according to the fitting curve, and generating a program load spectrum.
  9. 9. The method for predicting the service life of the mining shovel by adopting the high-torque load extrapolation model according to claim 8, wherein the power function is a bivariate power function, and the expression is as follows: ; Wherein y is the cumulative frequency, x is the torque amplitude, and y 0 、A 1 、t 1 、A 2 、t 2 is the fitting parameter.
  10. 10. The method for predicting the service life of the mining shovel by adopting the high-torque load extrapolation model according to claim 6, wherein in the step S5, the preset precision requirement is that the statistical parameter error of a high-torque area is less than or equal to 5%, the statistical parameter error comprises a mean value error, a standard deviation error and a maximum amplitude error, and the statistical parameter error of a medium-torque area is less than or equal to 10%.

Description

Method for predicting working life of mining electric shovel by adopting high-torque load extrapolation model Technical Field The invention belongs to the technical field of life prediction and reliability analysis of electromechanical equipment, and particularly relates to a method for predicting the working life of a mining electric shovel by adopting a high-torque load extrapolation model. Background In the surface mine exploitation operation, a large-scale mining electric shovel is used as core exploitation equipment, and the working efficiency and the reliability of the large-scale mining electric shovel directly determine the mine production scale and the economic benefit. The four large transmission mechanism motors for lifting, turning, pushing and walking of the electric shovel are required to bear heavy alternating torque loads for a long time, wherein the lifting mechanism motors are required to overcome the gravity and inertia force of materials in the lifting stage of the materials, instantaneous impact large torque is easy to generate, the pushing mechanism motors are distributed in a multimodal manner due to the hardness difference of the shoveled materials (from soft rock to hard ore), the peak torque is random, the turning mechanism motors are influenced by turning start and stop inertia, the torque fluctuation is severe, and the walking mechanism motors are relatively stable in load but still generate impact torque due to uneven pavement when walking for a long distance. Meanwhile, the electric shovel working environment is accompanied with severe conditions such as high dust, strong vibration, abrupt temperature change and the like, a motor transmission shaft is used as a core component for torque transmission, fatigue failure is easy to occur due to long-term bearing of alternating load, particularly high torque load, equipment is stopped and maintained when the equipment is light, safety accidents are caused when the equipment is heavy, and huge economic loss is caused. The definition of large torque refers to a torque value which exceeds rated torque of a motor and does not exceed maximum torque of the motor, and the contribution ratio of the torque to fatigue damage of a transmission shaft of the electric shovel motor can reach more than 60%, so that the electric shovel motor is a load type needing to be focused in working life prediction. The medium and small torque is a torque value not exceeding the rated torque of the motor. The number of load samples has a direct correlation to the analysis of the working life and the fatigue resistance design. For the transmission shaft of the electric shovel motor, the working life prediction of the whole life cycle needs to completely cover the whole working condition torque change from starting, working to stopping, and particularly needs to accurately represent the distribution characteristics of the large torque load. However, the key problem of the existing electric shovel motor torque extrapolation technology is that the large torque sample size is small, and the essence is that in the software acquisition process, the occurrence frequency of the large torque load is extremely low (the large torque load can only occur for a plurality of times or even zero times in a single actual measurement period), so that the large torque data is lost, and the number of extrapolated samples is insufficient; the problem of insufficient sample size caused by the loss of large torque data of an electric shovel is not solved pertinently, and finally, an extrapolated load spectrum cannot accurately reflect the influence of large torque on fatigue damage of a transmission shaft, and the reliability of fatigue life prediction and anti-fatigue design is affected. Disclosure of Invention The invention provides a method for predicting the working life of a mining electric shovel by adopting a large-torque load extrapolation model, aiming at the problems of sparse large-torque sample number and low extrapolation precision in the extrapolation process caused by the defect of large-torque data in the software acquisition process during the extrapolation of the motor torque of the electric shovel in the prior art. Under the condition of ensuring that the overall distribution of samples is unchanged, the large-torque samples are directionally expanded by a directional expansion technology aiming at large torque, the number of samples in a large-torque area is increased, an Epanechnikov kernel density estimation model is combined, the accurate extrapolation of the torque of the electric shovel motor in the whole life cycle is realized, meanwhile, the extrapolation result is ensured to be actually matched with engineering through multi-dimensional verification, and finally, the fatigue test and life prediction are realized. The invention is realized by adopting the following technical scheme: A method for predicting the working life of a mining electric shovel by adopt