Unlock instant, AI-driven research and patent intelligence for your innovation.

Health state prediction method for industrial equipment in noisy environment

A technology for industrial equipment and health status, applied in design optimization/simulation, computer-aided design, calculation, etc., to achieve high prediction accuracy

Active Publication Date: 2021-10-22
BEIHANG UNIV
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The present invention aims at the defects in the above-mentioned prior art, and proposes a method for predicting the health status of industrial equipment in a noisy environment. The sequence data samples of the model are removed to build a more accurate prediction model

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Health state prediction method for industrial equipment in noisy environment
  • Health state prediction method for industrial equipment in noisy environment
  • Health state prediction method for industrial equipment in noisy environment

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment 1

[0099] The simulation experiment is carried out with a turbofan engine. This experiment uses a data set generated by a turbofan engine simulation program, Commercial Modular Aviation Propulsion System Simulation (C-MAPSS), to verify the effectiveness of the method of the present invention.

[0100] This dataset contains four subsets, denoted by FD001, FD002, FD003 and FD004. Each subset contains a training set and a test set. The training set contains the life cycle monitoring data of 21 sensors and 3 operating condition sensors of multiple engines of the same type. In subsets FD001 and FD003, the operating conditions experienced by each engine remained constant, while in FD002 and FD004, the operating conditions were constantly changing. Therefore, the input sequences of subsets FD002 and FD004 contain operating condition sensor data, but FD001 and FD003 do not. So we merged FD001 and FD003 into one dataset, and FD002 and FD004 into another dataset, denoted by FD013 and FD...

specific Embodiment 2

[0120] The present invention uses real monitoring data of a milling machine to verify the effectiveness of the proposed method. In this embodiment, the monitoring data obtained by six sensors are used to predict the wear amount of the milling cutter.

[0121] This data set contains 16 knives, each of which has undergone different working times, and six sensors have recorded 9000 data points in each work, and only use the last 5000 data points in the steady working state to predict the current milling Knife wear. In this embodiment, only the data of No. 7, No. 13, No. 3 and No. 11 knives are selected, because these knives have the largest amount of data. Table 3 shows the experimental conditions of these four knives, which belong to two different experimental conditions, so two sets of experiments are designed to analyze the data of these two experimental conditions respectively. Knife No. 13 and No. 11 are used as training sets to build the model, and No. 7 and No. 3 knives ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention provides a method for predicting the health state of industrial equipment in a noisy environment. The method is based on deep learning and unsupervised noise sample detection, including: using sensors installed on industrial equipment to obtain monitoring data containing noise; using window size for s tw The sliding time window generates multiple sequence data samples based on noisy monitoring data; uses the conversion model with attention mechanism to generate attention weight vector α for each sequence data sample T ; The bottom-up hierarchical clustering algorithm based on the average connection algorithm will α T Perform clustering and remove the α in the abnormal class T For the corresponding sequence data samples, use the remaining sequence data samples to train the LSTM prediction model; use the trained LSTM prediction model to predict the health status of industrial equipment; the present invention converts the multivariate input sequence into an attention weight vector related to prediction, Hierarchical clustering is performed on it to detect and remove noise samples to make the prediction model more accurate.

Description

technical field [0001] The invention relates to the technical field of industrial equipment health prediction, and relates to a method for predicting the health state of industrial equipment in a noisy environment, in particular to an unsupervised noise sample detection method for industrial equipment health state prediction. Background technique [0002] With the development of sensor technology, a large amount of monitoring data of equipment can be obtained to predict its health status. Compared with traditional forecasting methods based on signal processing, methods based on deep learning can process large amounts of data without requiring too much manual operation and professional domain knowledge. Many deep learning models have been used to predict the remaining lifetime of devices, such as multi-layer perceptron models, convolutional neural networks, recurrent neural networks, and long short-term memory networks. Sliding time window methods are widely used to generate...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06F30/20G06F119/04
Inventor 林焱辉常亮
Owner BEIHANG UNIV