Published: Jan 1994
| ||Format||Pages||Price|| |
|PDF (280K)||20||$25||  ADD TO CART|
|Complete Source PDF (2.9M)||20||$55||  ADD TO CART|
In the petroleum refining industry, use of Continuous Process Analyzer Systems to perform product quality measurement functions has been common practice for over at least two decades. Designed to operate unattended, these systems provide analytical information ranging from physical properties such as viscosity, density to total and trace chemical composition, on samples continuously extracted from the main process streams. While this information has primarily been used for process monitoring, control, and real time optimization, its usage for on-line product quality certification has been limited. In spite of the capability of these systems to equal or outperform their labour-intensive laboratory counterparts both in precision and accuracy, most organizations are reluctant to use them for product certification purposes. This is primarily due to the lack of a structured, objective, data-driven approach taken by both internal and external users (customers) of these systems to measure, compare, and demonstrate their capabilities. With the increasing cost of crude oil, and more stringent product specifications demanded by customers and environmental legislation, the business incentive capturable by using a more capable measurement system for product quality certification has become increasingly significant.
This paper begins with an overview discussion of the measurement process, its uses, and performance metrics. The main body deals with the application of basic statistical quality assurance concepts, principles and practices to identify the appropriate (statistical) model, quantify, compare, and continuously demonstrate in-control status of measurement systems. Control charts referenced include X-bar, I, R, MR2, and EWMA. Statistical tools referenced include t, F tests of significance, normal probability plot, ANOVA. This paper concludes with examples of how these tools and practices are successfully applied to the on-line certification of gasoline vapour pressure quality using a continuous process analyzer system.
Detailed presentation on mathematical formulae and computational mechanics for statistical tools and control charts are beyond the intended scope of this paper. Readers interested in these details are referred to statistics and statistical quality assurance textbooks. Several are listed in the reference section.
measurement process, continuous process analyzer system, specification, product quality certification, statistical quality assurance, precision, bias, true value, point estimate, interval estimate, statistical control, control charts
Fuels Blending and Quality Assurance Specialist, Imperial Oil, Toronto, Ontario