Page 382 - Design of Simple and Robust Process Plants
P. 382

9.5 Closed Loop Steady-state Optimization  369
                Model  Input data reconciliation is based on mass or mass and energy balance
                reconciliation. These balances have been selected because they do require any valida-
                tion while based on conservation laws. The overall mass balance of the system is
                used to provide the simulation/optimization of the process with good input data.
                Feed quantity and quality are particularly important. The accuracy of the reconcilia-
                tion step improves if more redundant nodes are built into the flowsheet (Figure 9.9).
                It might be beneficial for the increase in redundancy of the flow-sheet for the plant
                mass balance to include the storage area. Due to its inventory measurement, an
                additional redundancy is added by averaging the mass measurement over a time
                period. Large-sized vessels in the process can also be equipped with a hold-up meas-
                urement and taken into account for the mass balance reconciliation.
                  The accuracy for the reconciled data depends on: the accuracy of the individual
                measurements; the redundancy built into the flowsheet; and process stability. The
                method of mass balance reconciliation over the process is also applied to the perfor-
                mance measurement, as discussed under performance meter (see Section 9.4).





                  Redundant
                  nodes one


                  Redundant
                  nodes three



                                                           L


                                                               L
                 Redundant
                 nodes seven
                                                          L
                                                                L


                Fig. 9.9. Illustration of adding redundancy for flow measurements.




                Data reconciliation technique (Britt and Luecke, 1973; Albers, 1999) This technique
                is based on minimization of the deviations of the square of the difference between
                the reconciled value of the measurement and the measured value. In formulae
                form:
   377   378   379   380   381   382   383   384   385   386   387