Manufacturing Process Optimization in the Process Industry

Literature Review

In this section we provide an overview of an assessment of current data analytics project modeling techniques and address the methods used by experts to model processes. KPIs were considered a basis for assessing the efficacy of process enhancement strategies. Baardman et al. (2023) proposed design of dynamic experiments, an innovative data-driven experimental design methodology for optimizing industrial crystallization processes (OICP). In this case, the technique was applied to the batch crystallization process to establish the optimum lowering improvement, which when integrated, yielded the optimal temperature profile. For the batch crystallization of potassium nitrate in water, we compared literature on the optimum temperature profile produced using a model-based optimization strategy with the data-driven optimum temperature profile to evaluate the efficacy of the OICP optimization method. This comparison revealed some degree of overlap between the distributions when only the parametric volatility was analyzed (Sun et al., 2023). There was significant overlap when the variation in the design variables and the seed capital distribution parameters were evaluated (Mandl & Minner, 2023).

Bernabei et al. (2023) introduced location-specific allocation problems over time intervals during a disease outbreak, factoring in state data collected at regular intervals from different locations and expanding understanding of transport protocol. The problem was formulated as a dynamic optimization model applied to a system of standard equations describing the transmission patterns of infection in various geographic regions (Liu et al., 2023). The data-driven optimization approach determined the optimal distribution of intervention funds across an indefinite number of communities and periods. The results showed that our data-driven optimization technique helps resolve inter-decision problems with unknown system dynamics.