STATISTICAL ASPECTS OF MIXING

THE AUTHOR GIVES VARIOUS METHODS FOR ASSESSING THE MIXING OF SOLIDS. FIRST HE DEFINES IDEAL MIXING, MAINLY IN THE CASE OF TWO COMPONENTS OF THE SAME PARTICLE SIZE. THIS DEFINITION IS VALID FOR MORE THAN TWO COMPONENTS, BUT WHEN THE COMPONENTS HAVE DIFFERENT PARTICLE SIZE OR THE PARTICLE SIZE OF ONE COMPONENT VARIES THIS DEFINITION IS NOT VALID. BUT THE DISTRIBUTION OF THE NUMBER OF PARTICLES OF A COMPONENT IN A SAMPLE TAKEN AT RANDOM CAN STILL BE TAKEN AS A BASE. IT IS POSSIBLE, BY MEANS OF THE CHI-SQUARED METHOD, TO ESTIMATE THE DEGREE OF MIXING BY COUNTING THE NUMBER OF PARTICLES OF ONE COMPONENT IN A NUMBER OF SAMPLES TAKEN AT RANDOM. IT IS OFTEN IMPOSSIBLE IN PRACTICE TO COUNT PARTICLES AND THE MIXING HAS TO BE TESTED BY WEIGHING AND CHEMICAL ANALYSIS. NOW MIXING CAN BE TESTED BY TAKING SAMPLES AND MIXING EACH SAMPLE THROUGH AND THEN SPLITTING IN TWO. THEN ONE OBTAINS PAIRS OF RESULTS FROM WHICH, WITH AN F-TEST, THE MIXING CAN BE TESTED. FINALLY THE AUTHOR DISCUSSES A CASE OF INCOMPLETE MIXING, NAMELY THE SPECIAL CASE OF SEGREGATION BY GRAVITY. IN THIS CASE SAMPLES SHOULD BE TAKEN FROM HORIZONTAL LAYERS OF THE MIXTURE AND THE METHOD OF SPEARMAN OR KENDALL MUST BE USED. /SWOV/RRL/

  • Supplemental Notes:
    • Vol 13, No 1, PP 4-6, 5 REF
  • Corporate Authors:

    Stichting Kwaliteitsdienst Indus /Neth

    ,    
  • Authors:
    • Doornbos, R
  • Publication Date: 1967-2

Subject/Index Terms

Filing Info

  • Accession Number: 00214899
  • Record Type: Publication
  • Source Agency: Road Research Laboratory /UK
  • Files: TRIS
  • Created Date: Aug 12 1970 12:00AM