DESIGN OF A REPRESENTATIVE AND COST-EFFECTIVE SAMPLING PROGRAMS FOR INDUSTRIAL WASTE-WATER WITH EXAMPLES FROM BLEACHED KRAFT PULP AND PAPER-MILLS

The diagram constructed for selection of sampling methods indicates th at, for a systematic error, E less than or equal to 13%, grab sampling (GS) map be used to characterize effluents with variation coefficient of flow less than or equal to 120% and of contaminant concentrations less than or equal to 10%. For the whole studied range of variation co efficient of contaminant concentrations (2-82%), time-proportional com positing (TC) method may be applied with E less than or equal to 10% f or effluents characterized by variation coefficients in flow < 90%. Th e more complicated flow-proportional compositing (FC) method is requir ed only for effluents with larger variation coefficients or to produce more precise results. The diagram constructed for selection of sampli ng frequencies indicates that sequential sampling at intervals of appr oximately 60 min may be applied with E less than or equal to 10% for e ffluents characterized by variation coefficients less than or equal to 30%. Practical application of the diagrams, constructed using normal series, was checked against monitoring data from two pulp and paper mi lls in Vietnam. The two diagrams provided results on sampling methods and frequencies in good agreement with those obtained from actual moni toring data with percentages of agreement cases of 80 and 75%, respect ively. The approach was applied in design of a monitoring program at t he Bai Bang integrated pulp and paper mill in Vietnam. Copyright (C) 1 996 Elsevier Science B.V.