KR20200036599A - Apparatus and method for diagnosing machine fault - Google Patents

Abstract

The present invention relates to a machine defect diagnosis device capable of increasing accuracy of machine defect diagnosis and a method thereof. More specifically, the machine defect diagnosis device comprises: a collection part collecting sound data based on at least one sensor; an extraction part acquiring a feature image based on the collected data and selecting features by reducing dimensions by applying main component analysis to the feature image; and a diagnosis part making a final diagnosis of defects based on the selected features by using an artificial neural network based nonlinear classifier.

Description

Machine defect diagnosis apparatus and method {APPARATUS AND METHOD FOR DIAGNOSING MACHINE FAULT}

The present invention relates to an apparatus and method for diagnosing machine defects, and more particularly, to an apparatus and method for diagnosing machine defects by clearly distinguishing internal characteristics from collected data to accurately diagnose a machine defect in real time.

The 4th Industrial Revolution is a strategy to graftly respond to changes in the social, technological, economic, ecological, and political sectors faced by the German manufacturing industry, and to respond in a comprehensive manner. The Internet of Things (IoT), enterprise software, and location information It aims to be a smart factory that actively utilizes ICT-related technologies such as security, cloud, big data, and virtual reality.

In this smart factory, smart manufacturing analysis is mainly focused on business intelligence solutions, and data analysis using structured data or numerical data (BI) is the main focus, and some research institutes are attempting to develop tools to support expert predictions, but foreign technology Compared to this, it is in the beginning stage and highly dependent on manual work. Predominantly, predictive analysis using numerical data is predominant in Korea, and predictive research in which information is extracted from unstructured data is in its infancy.

2. Description of the Related Art With the development of predictive analysis technology, a predictive production system that provides self-recognition capability to production machines to provide workers with status information of machines clearly and improve productivity, efficiency, and safety has been actively studied.

In particular, the management of production facilities in the manufacturing industry is a very important factor in production management, and the failure of production machines or errors leads to problems such as a decrease in production rate and an increase in defect rate in the field. It is a trend that introduces predictive maintenance technologies that require and take necessary measures before problems arise.

Therefore, it is necessary to clearly classify internal characteristics from the collected data to diagnose a machine defect, thereby increasing the diagnostic accuracy of the machine defect.

Therefore, the present invention has been proposed to solve the above-described problems, and machine defect diagnosis to clearly improve the accuracy of diagnosis for machine defects by clearly distinguishing internal characteristics from collected data for diagnosing machine defects. It provides an apparatus and method.

In addition, the present invention provides a machine defect diagnosis apparatus and method for diagnosing machine defects in real time, thereby improving safety and reliability of a system while satisfying a minimum maintenance cost.

The objects of the present invention are not limited to those mentioned above, and other objects not mentioned will be clearly understood by those skilled in the art from the following description.

The machine defect diagnosis apparatus according to the present invention for achieving the above object is a collection unit for collecting sound data based on at least one sensor; acquires a feature image based on the collected data, and the feature image An extraction unit for selecting a feature by reducing a dimension by applying principal component analysis; And a diagnostic unit for finally diagnosing a defect based on the selected feature using an artificial neural network based nonlinear classifier.

In addition, in the method for diagnosing a machine defect according to the present invention for achieving the above object, the collecting unit collects sound data based on at least one sensor; the extracting unit acquires a feature image based on the collected data, , Selecting a feature by reducing a dimension by applying principal component analysis to the feature image; And a diagnostic unit for finally diagnosing a defect based on the selected feature using an artificial neural network-based nonlinear classifier.

According to the present invention, it is possible to increase the diagnostic accuracy for machine defects by clearly distinguishing internal characteristics from data collected for diagnosing machine defects.

Further, according to the present invention, it is possible to improve machine safety and reliability while satisfying a minimum maintenance cost by diagnosing machine defects in real time.

The effects of the present invention are not limited to those mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the following description.

1 is a block diagram showing the configuration of a machine defect diagnosis apparatus according to an embodiment of the present invention,
2 is a flow chart showing a method for diagnosing machine defects according to an embodiment of the present invention;
3A and 3B are diagrams showing data clustering results according to the prior art,
4A and 4B are diagrams showing data clustering results according to the present invention.

Objects and effects of the present invention, and technical configurations for achieving them will be clarified with reference to embodiments described below in detail together with the accompanying drawings. In the description of the present invention, when it is determined that a detailed description of known functions or configurations may unnecessarily obscure the subject matter of the present invention, the detailed description will be omitted. And terms to be described later are terms defined in consideration of donation in the present invention, which may vary depending on the intention or custom of the user or operator.

However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various different forms. Only the present examples are provided to make the disclosure of the present invention complete, and to fully inform the person of ordinary skill in the art to which the present invention pertains, the scope of the present invention being defined by the scope of the claims. It just works. Therefore, the definition should be made based on the contents throughout this specification.

On the other hand, in the specification, when a part is "connected" with another part, it is not only "directly connected", but also "indirectly connected" with another member in between. Includes. Also, when a part is said to “include” a certain component, this means that other components may be further provided instead of excluding the other component unless otherwise stated.

The present invention analyzes a signal generated by a machine, converts a one-dimensional signal into two-dimensional, and applies a signal processing technology to reinforce a healthy pattern, reduces the dimension by principal component analysis, and then uses a non-linear classifier for final discrimination. The fault is automatically diagnosed.

Hereinafter, an apparatus and method for diagnosing a machine defect according to an exemplary embodiment of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram showing the configuration of a machine defect diagnosis apparatus according to an embodiment of the present invention.

Referring to FIG. 1, the machine defect diagnosis apparatus 100 largely includes a collection unit 11, an extraction unit 13, and a diagnosis unit 15.

The collection unit 11 collects data based on at least one sensor. At this time, the data may be sound data. For example, when collecting sound data of a drill, a sound generated by a drill during an active time in an industrial site may be used, and a signal of a normal state and a signal of an abnormal state where the blade is worn may be configured.

The extracting unit 13 acquires a feature image to extract differential features from the data collected by the collecting unit 11 and selects a feature by acquiring projection data by applying principal component analysis (PCA), that is, principal component analysis do.

Here, PCA is a statistical procedure that uses an orthogonal transform to transform a correlated data set into a set of values that do not have a linear correlation. These values are called main components. The number of principal components obtained is less than or equal to the dimension of the original data. PCA is defined so that the first principal component has the largest possible variance, making it sensitive to the variability of the data set. Each successive component should have the highest possible variance under the constraint that it must be orthogonal to the preceding component. To obtain a new expression, the original data is projected to a new standard called PCA space, and the acquired vector is used as the main component. This is part of the PCA's dimensionality reduction capabilities.

Specifically, the extraction unit 13 includes a calculation unit 131, a conversion unit 133, an acquisition unit 135 and an analysis unit 135.

First, the calculation unit 131 calculates the spectral size of each data using Fourier analysis on the data collected by the collection unit 11.

The converter 133 divides the 1-dimensional vector containing the spectrum again and converts it into a 2-dimensional vector by setting the maximum amplitude of the matrix to 255 and setting the minimum value to 0 to perform an 8-bit gray scale image through a linear normalization process. Convert to

The acquisition unit 135 is characterized by enhancing contrast using histogram smoothing to spread and reinforce a healthy pattern, that is, a useful pattern, which may exist in the data of the image converted into the 8-bit gray scale image by the conversion unit 133 Acquire an image.

The analysis unit 137 acquires projection data by applying principal component analysis on the feature image to reduce the order and obtain the first projection data, and then selects some projection data. That is, after projecting the feature image for the main component, the feature of the system is selected.

On the other hand, the diagnostic unit 15 uses a nonlinear classifier based on an artificial neural network, and diagnoses a machine defect based on a feature selected by the analysis unit 137.

Here, artificial neural networks (ANNs) consist of interconnected groups of artificial neurons whose information flows from the starting neuron to the neurons at the very end of the network. Information flows using nonlinear functions. The purpose of this function is to help you understand the data by discovering the nonlinearities contained within the data. ANN is an adaptive system that changes the structure based on the received information and acquires the ability to generalize the learned comprehension after the learning process. The network learns by back propagating errors between the actual output and the desired output.

That is, in the present invention, the projected value of the feature image is given to the neural network-based classifier, and the output represents the final diagnosis result of the machine fault diagnosis (MDF) system for diagnosing machine defects.

2 is a flowchart illustrating a method for diagnosing machine defects according to an embodiment of the present invention.

First, when the collection unit 11 collects data on the machine (S201), the calculation unit 131 calculates the spectral size of each data using Fourier analysis on the collected data (S203), and then converts In step 133, the 1-dimensional vector including the spectrum is divided and converted into an 8-bit gray scale image in order to convert it into a 2-dimensional vector (S205).

Thereafter, the acquisition unit 135 acquires a feature image by enhancing contrast by using histogram smoothing in order to supply and reinforce useful patterns that may exist in the data of the converted image (S207).

The analysis unit 137 extracts a principal component by using the histogram-smoothed image as the original data of the feature image (S209), applies the extracted principal component to the feature image, that is, acquires projection data by projection, and selects a feature (S211).

The diagnosis unit 15 finally diagnoses a machine defect based on the feature selected in step S211 using a nonlinear classifier based on an artificial neural network.

The series of processes are briefly described below.

To make a one-dimensional vector into a two-dimensional vector, 2205 = 49 × 45 is calculated, and accordingly, the vector length of the spectral size is divided into 49 × 45 dimensions. We have 45 49-dimensional vectors as columns of the 2D matrix to be created, and we provide a 2-dimensional matrix of size 49 × 45. After the linear normalization process, an 8-bit gray scale image of 49 x 45 is obtained. The 2183 data sounds are at the origin, and the feature image to be processed forms a matrix with a size of 49 x 45 x 2183. Finally, histogram smoothing is applied to the generated image to adjust the contrast and disperse the good pattern throughout the image. The PCA base is computed using the histogram smoothed image as the original data of the feature image. After projecting the feature image into PCA space, some projected data is selected for use as input to the nonlinear classifier. Nonlinear classifier Final diagnosis of machine defects based on the selected projected data.

 3A and 3B are diagrams showing data clustering results according to the prior art.

Specifically, FIG. 3A is a view showing data projected using a time series according to the prior art, and FIG. 3B is a view showing data projections represented by statistical characteristics of the time series according to the prior art.

First, referring to FIGS. 3A and 3B, it can be seen how two types of data are clustered together. This means that the time series of sound data does not represent useful information for diagnostic analysis. 3B, it can be seen that in the case of time series, the data are very similar, and all are clustered together.

4A and 4B are diagrams showing data clustering results according to the present invention.

Specifically, FIG. 4A is a diagram showing data by a spectral size based feature image according to the present invention, and FIG. 4B is a diagram showing data by histogram smoothing of a spectral size based feature image according to the present invention.

4A and 4B, it can be seen that a difference between the two data appears, and further, a clearer difference appears when a histogram smoothed image is used. In particular, Figure 4b shows a projection using an image with improved contrast used as the final feature, and you can see how the data is properly separated. Two distinct clusters are created.

In other words, after applying PCA using statistical features of time series as the original data, the classifier is having difficulty in recognizing abnormal data, and many of them are misclassified. Discrete wavelet transform (DWT) coefficients and continuous wavelet transform (CWT) coefficients represent time-frequency domain based analysis. In Table 1 below, the DWT coefficient shows an accuracy of 95.7% in total, and the DWT coefficient is still excellent, while the CWT coefficient does not recognize abnormal data well.

Recall that the present invention is based on the frequency domain in a method used by converting the spectral size of sound into an image. Therefore, even in comparison with the existing technology, the feature image coefficient showed excellent performance by capturing the main characteristics of the data.

P.K. Kankar et al. K.M. Lee et al. P.K. Kankar et al. Example Feature extraction method Time series + statistical features DWT coefficients + PCA CWT components + statistical features Feature images Classification method Neural Network Neural Network Neural Network Neural Network accuracy 86.1% 95.7% 75.8% 99.5%

Therefore, the present invention makes it possible to increase the diagnostic accuracy of machine defects by clearly distinguishing internal characteristics from data collected to diagnose machine defects.

In the present specification and drawings, preferred embodiments of the present invention have been disclosed, and although specific terms are used, they are merely used in a general sense to easily describe the technical contents of the present invention and to understand the present invention. It is not intended to limit the scope. It is apparent to those skilled in the art to which the present invention pertains that other modifications based on the technical spirit of the present invention can be implemented in addition to the embodiments disclosed herein.

10: machine fault diagnosis device 11: collection unit
13: extraction unit 15: diagnostic unit
131: calculation unit 133: conversion unit
135: acquisition unit 137: analysis unit

Claims (4)

In the machine fault diagnosis apparatus,
A collection unit that collects sound data based on at least one sensor;
An extracting unit for acquiring a feature image based on the collected data, and selecting a feature by reducing a dimension by applying principal component analysis to the feature image; And
And a diagnostic unit for finally diagnosing a defect based on the selected feature using an artificial neural network-based nonlinear classifier.
According to claim 1,
The extraction unit,
A calculation unit that calculates a spectral size of each data using Fourier analysis on the collected data;
A conversion unit that converts an 8-bit gray scale image based on the calculated spectral size to convert a 1-dimensional vector including a spectrum into a 2-dimensional vector;
An acquiring unit that acquires a feature image by enhancing the contrast by smoothing the histogram of the converted image; And
A machine defect characterized by including an analysis unit that selects a feature by selecting a portion of projection data after obtaining projection data by applying a principal component base extracted using the histogram smoothed image as the original data of the feature image to the feature image. Diagnostic device.
In the method of diagnosing machine defects,
Collecting step of collecting sound data based on at least one sensor;
The extracting unit acquires a feature image based on the collected data, and applies a principal component analysis to the feature image to select a feature by reducing a dimension; And
Method for diagnosing a machine defect, characterized in that the diagnostic unit includes a diagnostic unit for finally diagnosing a defect based on the selected feature using an artificial neural network-based nonlinear classifier.
According to claim 3,
The step of selecting the feature,
Calculating the spectral size of each data using Fourier analysis on the collected data;
Converting a 1-dimensional vector including a spectrum into a 8-bit gray scale image based on the calculated spectral size in order to convert the converted 1-dimensional vector into a 2-dimensional vector;
An acquiring unit acquiring a feature image by enhancing the contrast by smoothing a histogram of the converted image; And
An analysis unit extracting a principal component base using the histogram smoothed image as original data of a feature image;
Obtaining, by the analysis unit, projection data by applying the extracted principal component base to the feature image; And
And selecting a feature by selecting, by the analysis unit, some projection data from the obtained projection data.

Images