17 A Framework for Dynamic Simulation of Interconnected …
621
Table 3 List of solvers available in the Dyssol library
Method
Description
Agglomeration
Cell average
Based on the work of Kumar et al. [31]
Fixed pivot
Based on the work of Kumar and Ramkrishna [32]
FFT-based
Based on the work of Skorych et al. [54]
Table 4 List of additional modules available in Dyssol
Description
Dustiness tester This module evaluates the dust tendency of each material stream in the
flowsheet. It applies prognosis functions comparable to the testing methods
such as single drop, rotating drum and UNC dustiness tester [35]
Fig. 20 Genetic optimization algorithm for parameter adjustment
To demonstrate the work of the developed optimization algorithm, it was applied
to a simple generic flowsheet model of a grinding process, shown in Fig. 21. The units
were taken from the general unit library (see Sect. 8). The model and optimization
parameters are listed in Table 5. The model of the crusher unit uses Bond’s law with
the power input as a user parameter to adjust the degree of comminution. The splitter
recycles 50% of the milled material independent of the particle size, which is fed
back and mixed with the feed after applying a time delay of one second. The system
response to a stepwise change of the inlet mass flow is shown in Fig. 22.
The goal of the optimization case is to reach a constant value of the Sauter mean
diameter (SMD) at the outlet. The temporal progressions of the SMD and the power
input for the best solution after 30 and 65 iterations are shown in Fig. 23, in which
621
Table 3 List of solvers available in the Dyssol library
Method
Description
Agglomeration
Cell average
Based on the work of Kumar et al. [31]
Fixed pivot
Based on the work of Kumar and Ramkrishna [32]
FFT-based
Based on the work of Skorych et al. [54]
Table 4 List of additional modules available in Dyssol
Description
Dustiness tester This module evaluates the dust tendency of each material stream in the
flowsheet. It applies prognosis functions comparable to the testing methods
such as single drop, rotating drum and UNC dustiness tester [35]
Fig. 20 Genetic optimization algorithm for parameter adjustment
To demonstrate the work of the developed optimization algorithm, it was applied
to a simple generic flowsheet model of a grinding process, shown in Fig. 21. The units
were taken from the general unit library (see Sect. 8). The model and optimization
parameters are listed in Table 5. The model of the crusher unit uses Bond’s law with
the power input as a user parameter to adjust the degree of comminution. The splitter
recycles 50% of the milled material independent of the particle size, which is fed
back and mixed with the feed after applying a time delay of one second. The system
response to a stepwise change of the inlet mass flow is shown in Fig. 22.
The goal of the optimization case is to reach a constant value of the Sauter mean
diameter (SMD) at the outlet. The temporal progressions of the SMD and the power
input for the best solution after 30 and 65 iterations are shown in Fig. 23, in which