| dc.contributor.author | Ketata, Chefi. | en_US |
| dc.date.accessioned | 2014-10-21T12:33:19Z | |
| dc.date.available | 1998 | |
| dc.date.issued | 1998 | en_US |
| dc.identifier.other | AAINQ39321 | en_US |
| dc.identifier.uri | http://hdl.handle.net/10222/55614 | |
| dc.description | The analysis of the data obtained from stream sampling is a crucial step in understanding the performance of a mineral processing plant. To control the sampling process efficiently, it is very important to minimize sampling errors and estimate them. The factors influencing these errors are divided into two categories: material properties and cutter features. This thesis includes four innovations dealing with knowledge-assisted stochastic evaluation of sampling errors in mineral processing streams. The sampling error is the difference between the sample composition and the composition of the total material that flows during the sampling period. | en_US |
| dc.description | First, the influence of the data variation on sampling errors is studied. A new expression for sampling error variances is developed. It is essential for the application of data reconciliation techniques and for the estimation of process performance indicator reliability. For some low sampling periods compared to the correlation length of the component flowrate signal, the approximation by Gy's expression is inaccurate. The difference between the formula developed in this thesis and Gy's expression increases with the number of sample increments. | en_US |
| dc.description | Secondly, the influence of the data variation on sampling errors throughout a two-stage flotation circuit is analyzed. The material balance technique is used to upgrade the measured variables. The weighted least-squares method is applied to minimize the estimation errors. The sampling errors are evaluated before and after material balancing for comparison. In one case, the covariance terms between the different components in a stream and between the different streams in the flotation circuit are included. In the other case, these terms are not included. This proves that, by including the covariance terms in the calculation, the variances of the sampling errors are reduced and therefore the reliability of the material balancing is improved. Consequently, a Sampling Error Filter is constructed. | en_US |
| dc.description | Thirdly, fuzzy logic can be employed to assess the sampling performance index that is the sample reliability. The latter is influenced by the sampler features describing the sampling conditions. Fuzzy sets lead to the sampling performance index value for each sample increment. It can be used as a weight in further calculations of average stream compositions or correct composition values. Hence, a Sampling Performance Indexer is proposed. | en_US |
| dc.description | Finally, two expert systems, Sampling Correctness Inspector and Sampling Error Evaluator, are developed. The knowledge is collected from experts publications in sampling of mineral processing streams in addition to the author's expertise in the considered domain. These expert systems take into account the stream properties, the cutter features, and the sampling manner. (Abstract shortened by UMI.) | en_US |
| dc.description | Thesis (Ph.D.)--DalTech - Dalhousie University (Canada), 1998. | en_US |
| dc.language | eng | en_US |
| dc.publisher | Dalhousie University | en_US |
| dc.publisher | | en_US |
| dc.subject | Engineering, Mining. | en_US |
| dc.title | Knowledge-assisted stochastic evaluation of sampling errors in mineral processing streams. | en_US |
| dc.type | text | en_US |
| dc.contributor.degree | Ph.D. | en_US |
