Manufacturing System design: from basic deterministic analysis to real-world performance

Designing Manufacturing Systems (MS) to achieve strategic capabilities is a complex undertaking that demands substantial specialized knowledge, time, and resources. Modeling MS for this purpose faces substantial implementation challenges, particularly in the early design phase, where prior knowledge of management strategies and operational variability is limited.

This data suport the paper analisys, that propose to link fundamental deterministic and probabilistic analysis, enabling well-informed from design strategic decisions. Following the Design Science Research paradigm, a deductive approach was adopted to develop a prescriptive design artifact. The experimental analysis comprised five discrete-event simulation (DES) models. For each model, seven key parameters governing the MS workflow were systematically varied probabilistically [Cycle time (min), Setup time (min), Inventory Buffer (units), Transfer batch (units), Randomness in order sequencing (RS), Unvailable time (min) and Rework fraction (%)] across designed scenarios, ranging between 5 levels (1-5), from Worst-Case to Best-Case operational extremes.

Simulation output variables [Makespan (min), Work in process (units), Time within the system (min), Reworked items (units), Processing up time (%), Setup time (%), Break down time (%), Throughput rate (units/min)] and the expected dynamic for the studied MS validated the conceptual assumptions. This research established a formal connection between deterministic and probabilistic analysis, allowing for simple and accurate estimates of Makespan and Throughput for single-station MS, validated for both single- and multiple-product contexts.

Thus, the data presents the parameters used in the simulation model and the values ​​of the variables, which were obtained in 50 replications. The scenarios considered were the production of a single product (A) and multiple products (B). Additionally, the data provides 50 replications for the validation scenarios (V1 and V2) of the results, in which modifications were implemented to the parameters considered in the original simulation (A and B).