City Research Online

Bulk sampling: Some strategies for improving quality control in chemical industries

Girardi, B. A. (1993). Bulk sampling: Some strategies for improving quality control in chemical industries. (Unpublished Doctoral thesis, City, University of London)

Abstract

An increasing number of industries are concerned about variability in the quality of chemicals. This thesis is devoted to these concerns, particularly to three underlying, but overlapping, strands: (i) Unfolding a Bulk Sampling Scheme; (ii) Sampling of Heterogeneous and Dynamic Material Systems; (iii) Designing of Experiments with Divisible Materials.

In Unfolding a Bulk Sampling Scheme we detail a sampling protocol to determine the sample size, the minimum amount of material, and an acceptance criteria centred on the characteristic variability of the particulate material to be assayed. Segregation, heterogeneity, particle size, randomization of solid-solid mixtures and all properties regarding every sort of lot — zero, one, two or three-dimensional — have a thorough examination not in conceptual terms but, indeed, within a mathematical model that allows for materialization of errors with predictable risks.

In Sampling of Heterogeneous and Dynamic Material Systems we con-sider the problem of serial measurements within the scope of Matheron’s Regionalized Variables. The dependency of two neighbouring samples from the same one-dimensional lot whether moving or stationary is studied to refine precision statements, minimize quality fluctuations and reduce heterogeneity of consignments. Moments of continuous selection errors and their computation using nonparametric methods are presented through Gy’s bulk sampling approach. Simplified methods to assess variability of continuous materials are also considered.

In Design of Experiments with Divisible Materials we deal with the problem of isolating variance components associated with sampling — primary, secondary and tertiary increments — and measurement systems. The method of conducting multifactor experiments, particularly Nested Experimental Design, to identify where in the process the quality improvement effort needs to be focused mostly is studied in a practical viewpoint according to the required needs. Balanced, staggered and Nested-factorial Design are analysed accordingly. The insight allows optimization of sample size, amount of replication and reduction of variability and cost.

In the last chapter Designing a Bulk Sampling Test Station some suggestions based on the ideas of Unit Operations are presented. Fundamental rules for a sound design and choice of equipment for guarding against estimation bias are also considered together with a full layout for a Sampling Test Station.

Publication Type: Thesis (Doctoral)
Subjects: H Social Sciences > HA Statistics
Departments: Bayes Business School > Actuarial Science & Insurance
Bayes Business School > Bayes Business School Doctoral Theses
Doctoral Theses
[img]
Preview
Text - Accepted Version
Download (7MB) | Preview

Export

Downloads

Downloads per month over past year

View more statistics

Actions (login required)

Admin Login Admin Login