A Clustering Method for Industrial Process Data Based on Density Peak Clustering

An industrial process and data clustering technology, applied in the fields of instrument, calculation, character and pattern recognition, etc., can solve the problem that the number of cluster centers cannot be automatically determined, the optimal cluster center cannot be determined, and the accuracy of clustering results is low. problem, to achieve the effect of good applicability, reduced calculation amount, and high classification accuracy

CN108537249BInactive Publication Date: 2020-05-19HUAZHONG UNIV OF SCI & TECH
3 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Publication Date
2020-05-19
Estimated Expiration
Not applicable · inactive patent

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention discloses an industrial process data clustering method of density peak clustering, comprising: obtaining industrial process data to form a data set; combining the Euclidean distance between the data in the data set with a time factor to obtain the distance between the data ;According to the distance between the data and the adjustment parameters, the cut-off distance is obtained, and then the local density of each data is obtained, and the minimum distance between each data and the data larger than its local density is calculated; for the local of each data in the data set The product of the density and the minimum distance is sorted, and the first H data with a larger product is taken as the cluster center, and the data closest to the cluster center among the data with a higher local density than the cluster center belongs to the same class as the cluster center; for the data The clustered data without class attributes judges their class attributes according to the order of local density from large to small, and then obtains the clustering results of industrial process data. The cluster center of the present invention is more reasonable, automatically determines the number of cluster centers and has less time complexity.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention belongs to the technical field of normal data, fault data and different mode classification of industrial processes, and more specifically relates to a clustering method for industrial process data of density peak clustering. Background technique

[0002] For a large-scale industrial system, due to the change of production strategy and production environment, the industrial process often presents the characteristics of multi-mode and multi-fault. For different modes, we need to establish different sub-models, so that the whole model can have better performance to monitor industrial processes and predict quality indicators. Therefore, it is of great significance to conduct mode identification and fault classification for multi-mode industrial processes before modeling.

[0003] Currently the most widely used methods for modal identification and fault classification are based on data-driven methods. There are two main methods, one is to u...

Examples

Embodiment 1

[0052] Embodiment 1 adopts the industrial process data clustering method based on the improved Density Peaks Clustering (DPC: Density Peaks Clustering) provided by the present invention, and verifies through the industrial process of semiconductors. Table 1 shows 16 different modalities and Correspondence table for industrial process data.

[0053] Table 1

[0054] modal Data points corresponding to the modality 1 1-24 2 25-49 3 50-73 4 74-82 5 83-107 6 108-132 7 133-153 8 154-178 9 179-203 10 204-223 11 224-248 12 249-269 13 270-294 14 295-318 15 319-340 16 341-364

[0055] The verification data used in Embodiment 1 of the present invention comes from the data of the semiconductor industry process, and the modern semiconductor production line is composed of hundreds of continuous batch processing stages. Each stage includes many steps carried out by expensive tools monitored by nu...