6.2 Simulation Model Details
167
The parameter setups, as well as delivery schedules, process times, etc., are
identical in all simulation model setups.
The data at the 15-minute resolution is available as a data export from the
Bosch energy portal but cannot be used for the simulation-based optimization
methodology. Due to the rather short cycle and machine state times in production, it is not possible to cleanly assign the data records to the machine states.
The example of the resolution comparison of the energy consumption profiles of
machine L1_load clearly shows that the number of measuring points for depicting the realistic energy consumption behavior separated according to the single
machine states is not possible (Figure 6.7). For the duration of the machine warmup of 19 minutes, e.g., only two data records are available when using such a
coarse data resolution; in the productive state of the machines, the period between
two measuring points covers the production of more than 40 produced parts. With
such strong data aggregation, the boundaries between the machine states become
blurred.
Figure 6.7 Comparison of data resolutions of L1_load over eight hours/one shift
Therefore, the data in a 15-minute resolution was not used to build a model
for the simulation-based optimization of energy efficiency in the Bosch production
scenario. The following section describes the optimization experiment runs using
the data with the resolutions of one data record per second and one data record
per minute.
167
The parameter setups, as well as delivery schedules, process times, etc., are
identical in all simulation model setups.
The data at the 15-minute resolution is available as a data export from the
Bosch energy portal but cannot be used for the simulation-based optimization
methodology. Due to the rather short cycle and machine state times in production, it is not possible to cleanly assign the data records to the machine states.
The example of the resolution comparison of the energy consumption profiles of
machine L1_load clearly shows that the number of measuring points for depicting the realistic energy consumption behavior separated according to the single
machine states is not possible (Figure 6.7). For the duration of the machine warmup of 19 minutes, e.g., only two data records are available when using such a
coarse data resolution; in the productive state of the machines, the period between
two measuring points covers the production of more than 40 produced parts. With
such strong data aggregation, the boundaries between the machine states become
blurred.
Figure 6.7 Comparison of data resolutions of L1_load over eight hours/one shift
Therefore, the data in a 15-minute resolution was not used to build a model
for the simulation-based optimization of energy efficiency in the Bosch production
scenario. The following section describes the optimization experiment runs using
the data with the resolutions of one data record per second and one data record
per minute.