Abstract:

The efficiency of solving crude scheduling optimization models is seriously affected by nonlinear constraints, which are produced by the property calculations of mixed crudes. In order to empower the model to cope with multiple crude pooling situations, a piecewise linear relaxation approach was employed to convert the nonlinear constraints into linear ones. Crude scheduling optimization of a real refinery was used as an example and the sulfur content calculations of mixed crudes were converted into linear forms. The effect of the piecewise linear relaxation approach was compared with another two frequently-used approaches. The results indicated that by use of the piecewise linear relaxation approach, the efficiency of solving the converted model met industrial application demand; while the sulfur contents of the processed crudes lied between 0.50% and 3.00%, the piecewise linear relaxation approach was more conducive to acquire optimal scheduling schemes; as the sulfur content difference of the processed crudes increased from 0.50% to 2.50%, the calculated sulfur contents of mixed crudes in charging tanks remained close to the actual value with the piecewise linear relaxation approach, the relative deviations were under 5% and therefore the industrial demand for model accuracy was satisfied.

Key words: crude scheduling, optimization model, piecewise linear relaxation, sulfur content