An Online Monitoring Method of Batch Process Based on Tensor Global-Local Preserving Projection
A technology that maintains projection locally and globally. It is applied in program control, computer control, general control systems, etc., and can solve problems such as information loss and poor accuracy.
Active Publication Date: 2015-12-02
ZHEJIANG UNIV OF TECH
View PDF3 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0006] In order to overcome the shortcomings of the existing intermittent process monitoring methods that only focus on the global or local characteristics of the data, information loss and poor accuracy when the data is unfolded, the present invention provides a technology that integrates tensor global-local projection and moving data windows An intermittent process online monitoring method based on tensor global-local projections that can simultaneously mine global and local features of data, effectively avoid information loss caused by data expansion, and have good accuracy
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
example
[0080] Example: Take the monitoring of the penicillin fermentation process as an example below to illustrate the effectiveness of the present invention. The sample set of the penicillin fermentation process was generated using the penicillin fermentation simulation platform Pensim2.0 developed by Professor AliCinar of the Illinois Institute of Technology. The sample set consists of 40 normal batches of fermentation data, and each batch contains 17 monitoring variables ( As shown in Table 1), the fermentation period is 400 hours, and the sampling interval is 1 hour, then the sample set is .
[0081]
[0082] Table 1
[0083] In addition, a failure batch was generated for online testing, the failure of this batch was caused by a step change in the acceleration rate of the medium flow (a sudden decrease of 10% from the original basis), which was introduced at 100h and lasted until 200h. Choose window length l to be 5, confidence level 1-α is 99%, utilize the present inventi...
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
Login to View More Abstract
An intermittent process online monitoring method based on tensor global-local projection, including the following steps: 1) select normal batch data, establish a training sample set; establish a moving data window, and obtain the data in the k time window to be standardized as Xk ; Execute the tensor global-local projection algorithm on Xk to obtain the projection matrices U and V; establish a monitoring model Yk=UTXkV, calculate the SPE statistics of the sample batch at time k; calculate the control limit of the SPE statistics; 2) record monitoring The data of the batch in the k time window is standardized to map to the monitoring model, calculate the SPE statistics; check whether the SPE exceeds the limit; if it exceeds, perform fault diagnosis, otherwise continue to monitor the k+1 time. The present invention integrates tensor global-local projection and moving data window technology, can mine global and local features of data at the same time, effectively avoids information loss caused by data expansion, and has good accuracy.
Description
technical field [0001] The invention relates to the field of process monitoring, in particular to a batch process monitoring method. Background technique [0002] At present, batch processes are widely used in industries such as biopharmaceuticals, food processing, fine chemicals, metallurgy, and papermaking. Due to the influence of many factors such as multi-process, multi-variable, complex reaction, uncertain process running time, frequent product replacement, etc., the intermittent process has a high probability of abnormal working conditions. Once an abnormal working condition occurs, a chain reaction may occur, leading to a decline in production capacity and fluctuations in product quality. In severe cases, equipment may even be damaged, causing casualties and environmental pollution, affecting the sustainable development of enterprises. How to ensure the safe and reliable operation of the batch process and the output of products with stable quality has become the focu...
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
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G05B19/048
Inventor 罗利佳包士毅高增梁
Owner ZHEJIANG UNIV OF TECH