theses_web.jpg

Thesis for the Degree of Master of Science

 

Year 2015
Markus Sintonen

OPC UA based Multivariate Analysis and Data Acquisition System for Chemometric Applications


Chemical industries utilize a variety of different types of online analyzers, for example, in areas like quality monitoring and process control applications. Large production plants typically use several analyzer devices from multiple manufacturers which are employed to measure different target quantities. As manufacturers have their own proprietary protocols for accessing analyzer information it is usually only the target property estimates that are transferred to the higher level automation systems. Other analyzer data, for example spectra, are generally not in a suitable format for further action on the higher level systems.

This thesis outlines a solution to the analyzer data acquisition problem by utilizing OPC UA standard and OPC UA analyzer devices companion specification (ADI) to create a data acquisition system for analyzer information. The created system consists of an OPC UA server, which is used as a single access point for all analyzer related information. In addition, the analyzer data is collected and archived into a SQL database, which is accessible through the OPC UA server. The data acquisition system makes it convenient for the end users to access plant analyzer information using the standardized protocols. Furthermore, the OPC UA based data acquisition system can be used to integrate analyzer data with other types of process measurements. This thesis presents an example application where this type of data integration is utilized to increase the accuracy of the target quantity estimates of an analyzer. The same system also has a wide range of other potential applications, some of which are briefly examined in this work.

The literature part of the thesis mainly focuses on different aspects of chemometrics; pre-processing and multivariate modelling of the analyzer spectra. These techniques are used in the thesis' experimental part to process data obtained from Neste Oil Oy’s Porvoo refinery. The literature part also briefly examines different protocols used for the transfer of analyzer data through the automation networks.

The experimental part the thesis consists of three main parts: The first part is a case study where the refinery measurement data and the analyzer spectra are utilized to demonstrate how product quality estimates accuracies can be improved through data integration. In the second part, the analyzer data acquisition system is developed, including the provision of a separate OPC UA ADI wrapper server for ABB online analyzers. This wrapper was created in order to obtain the data from the production unit analyzers used in the case study. In the final part, a chemometric calculation platform is designed in order to implement the data processing sequence used to process the refinery data. This platform also utilizes the newly created data acquisition system through the OPC UA protocol.

Thesis electronical version can be downloaded from here


This info last modified 22 Nov 2017 by Jukka Kortela