3.3 Energy Costs
57
Often, production specific data is not available due to missing manufacturing
execution system (MES) data. Teiwes et al. developed a load profile clustering (LPC) algorithm using only electrical load profiles and processing times to
identify different load levels such as processing and waiting times of machines
[Te+2018, p. 273]. The five-step approach is based upon a k-means clustering
algorithm and aims at a simplification of energy allocation (Figure 3.13). First,
the electrical consumption profile of the machine is measured. Thereafter, the single data entries are clustered in different clusters according to the machine states
using the k-means algorithm in several iterations. The clustering thus allows the
distinction of different load levels. After having defined the clusters, the process
time intervals are assigned, followed by product identification for the case that
several different products are processed at the machine under consideration. In a
fifth step, the results are subject to a plausibility check [Te+2018, p. 274].
Thus, the LPC methods allows the extraction of machine state-based energy
consumption information from available load profile data even though further process data is not available. However, the interpretation of energy data
is still challenging but measurability and transparency are the most important
requirements towards the optimization of energy efficiency in manufacturing
systems.
3.3
Energy Costs
The calculation of energy costs in producing companies is highly complex. As
the share of energy costs in total production costs differs depending on the industry sector, the products, as well as the processes, general statements regarding
the height of the energy cost share for industrial customers can hardly be made
[We2010, p. 21]. Additionally, it has to be considered, that main price elements
of electricity are taxes and duties, as well as regulated price components that
can neither be influenced nor optimized 10 . Therefore, the management of energy
consumption and every unused kilowatt hour (kWh) make up the best and costeffective, near-term way of energy efficiency optimization [Ge+2015, p. 82 f.;
Jo+2012, p. xvi; DS2008, p. 42].
The time frame of energy consumption can also influence the height of energy
costs. A high flexibility regarding the energy supply and the energy consumption
10 A detailed explanation of the cost structure and its components for the German electricity
prices as well as an example calculation for an electricity bill for a German industrial company
can be found in [DI2017, p. 7 ff.].
57
Often, production specific data is not available due to missing manufacturing
execution system (MES) data. Teiwes et al. developed a load profile clustering (LPC) algorithm using only electrical load profiles and processing times to
identify different load levels such as processing and waiting times of machines
[Te+2018, p. 273]. The five-step approach is based upon a k-means clustering
algorithm and aims at a simplification of energy allocation (Figure 3.13). First,
the electrical consumption profile of the machine is measured. Thereafter, the single data entries are clustered in different clusters according to the machine states
using the k-means algorithm in several iterations. The clustering thus allows the
distinction of different load levels. After having defined the clusters, the process
time intervals are assigned, followed by product identification for the case that
several different products are processed at the machine under consideration. In a
fifth step, the results are subject to a plausibility check [Te+2018, p. 274].
Thus, the LPC methods allows the extraction of machine state-based energy
consumption information from available load profile data even though further process data is not available. However, the interpretation of energy data
is still challenging but measurability and transparency are the most important
requirements towards the optimization of energy efficiency in manufacturing
systems.
3.3
Energy Costs
The calculation of energy costs in producing companies is highly complex. As
the share of energy costs in total production costs differs depending on the industry sector, the products, as well as the processes, general statements regarding
the height of the energy cost share for industrial customers can hardly be made
[We2010, p. 21]. Additionally, it has to be considered, that main price elements
of electricity are taxes and duties, as well as regulated price components that
can neither be influenced nor optimized 10 . Therefore, the management of energy
consumption and every unused kilowatt hour (kWh) make up the best and costeffective, near-term way of energy efficiency optimization [Ge+2015, p. 82 f.;
Jo+2012, p. xvi; DS2008, p. 42].
The time frame of energy consumption can also influence the height of energy
costs. A high flexibility regarding the energy supply and the energy consumption
10 A detailed explanation of the cost structure and its components for the German electricity
prices as well as an example calculation for an electricity bill for a German industrial company
can be found in [DI2017, p. 7 ff.].