Use of the computer package STATISTICA for data handling and analysis

Among more important parameters used for hard coal quality assessment should be ranked such ones, on the basis of which the type of coal can be determined. From among these parameters the deciding significance have: volatile matter content, caking properties – RI, swelling index – SI, and dilatometric properties. On all these parameters high impact exert factors, resulting from the presence in coal of the socalled ballast, among which ash and water are counted. Under great differentiation of parameters and their simultaneous mutual connections for drawing reliable conclusions necessary is the use of suitable statistical techniques. One of such contemporary statistical tools, helpful in investigations of this type, is the computer package STATISTICA. In this package different kinds of statistical tests are used for the characterisation of properties of the tested coal, on the basis of relationships between the analysed data. In investigations have been used data collected in the Data Bank about the Quality of Polish Hard Coals, conducted since many years in the Department of Solid Fuels Quality Assessment of the Central Mining Institute. Those were power coals produced in five selected hard coal mines from the Upper Silesian Coal Basin region. The computer package STATISTICA was used to carry out the following statistical tests : t-Student test, multiple regression, analysis of variance, analysis of main components. In order to ascertain if the average ash content in coals produced in the selected mines in 2009 differs from the average ash content in coals produced in the period 2005–2008, the t-Student test was carried out. Then the multiple regression was carried out in order to ascertain if there exist relationships between three parameters: caking properties according to Roga, content of element C and volatile matter content. It has pointed out that the relationships existing between each of the variables are small. The analysis of variance was used for the estimation of relationships between the ash content in coals from five different mines. On this basis the lack of variance equality in individual groups was ascertained, what argued for the lack of relationaships between coal contents in the analysed mines. The analysis of main components allowed to ascertain, which from the investigated parameters were statistically essential. In the analysed PCA model only three main components were essential (Wa, Aa, and Vdaf), of which the third component had inconsiderable significance.