and Simulation Tool. Grounding on a simulation approach, operation costs can be calculated precisely considering the detailed logistics constraints, providing feedbacks to theReconfiguration Planning Tool that might change the reconfiguration sequence along the horizon. Hence, the optimal solution can change and a new optimal sequence of reconfigurations must be identified.
4.5 Industrial application case
The overall system design and management framework, as well as the individual tools were applied to an automotive case. The results of the case study are detailed in the following sections, highlighting the solutions provided by the Production Planning and Simulation Tool, and its application as an integrated part of the workflow.
4.5.1 Description of the application case
In the application case, a Tier-1 automotive supplier is selected, producing car body parts for OEMs. The external environment is characterized by fragmented orders, resulted by the ever changing product portfolio, and also by frequent changes in joining technologies that the company should follow according to the specifications created by the OEMs. Although total yearly volumes are relatively constant over time, new products are continuously added to the portfolio, therefore, the demands correspond to smaller batches. The company has limited shop-floor space, thus this high-mix-low-volume production requires efficient variety management strategy to keep the competitiveness and internal efficiency. Moreover, the market environment is uncertain, increasing the problem complexity.
In the case study, four different products (P1−P4) are selected, for which a modular cell is to be configured and managed over a time horizon of three periods (u∈U), with equal lengths of three months (480 working hours). Each product has its own assembly specification with the corresponding technologies that need to be applied. In the analysis, only joining technologies are considered, of which products require nut pressing, resistance spot welding, adhesive joining and riveting. These technologies are performed by the combination of fixed equipment (skeleton) and a set of modular devices j∈J. The equipment dimensions are known (only 2D dimensions are considered), as well as the investment costs of the devices, ranging between e10.000-e120.000.
The hourly labor costs are known (50e/h), and the total time consumption of performing major changes in the cell configuration (reconfiguration) is two working weeks. The results of assembly system configuration are presented as follows.
4.5.2 Assembly cell configuration results
First, multiple rough cell designs were created by the Assembly System Configuration Tool, relying on the available information about the expected market situations. By defining input data about the products and corresponding processes, candidate cell configurations were created that match the expected output rate. These configurations are built up of the equipment that was stored in the repository. For the same scenario (product mix and order volumes), multiple different cell alternatives were defined, of which designers can select the most promising one(s) for further, more detailed analysis. The created solutions differed in the total occupied area, total initial investment costs, and also in other predicted cost factors, e.g. the operation, logistics and storage costs.
4.5 Industrial application case 72
The Assembly Cell Configuration Tool was then applied to generate a cell layout by ar-ranging the set of equipment selected in the previous step. Besides, the set of possible execution modalities was identified, defining task sequences and resource-task assignments. Two different layouts were generated, of which the one with shorter total cycle time was selected to be the applied. Based on the investment costs, the operational costs (calculated as discussed in the following section) and the stochastic market environment represented by the scenario tree, the reconfiguration strategy could be determined by the Reconfiguration Planning Tool. Along the time horizon U, expected production volumes, as well as the set of products to be produced in the cell were changing. Therefore, the cell reconfiguration strategy was defined by stochastic optimization, identifying the pieces of equipment that need to be added (or removed) to the cell configuration in a given periodu, within a reconfiguration. As theReconfiguration Planning Tool planning tool strongly relies on the data about operation costs, prediction and refinement
—considering a system-wide production planning— of these parameters were performed with theProduction Planning and Simulation Tool.
4.5.3 Production planning and simulation results
Applying the Production Planning and Simulation Tool, one can analyze the future expected operation costs and production batch sizes, based on the contractual delivery volumes known already in the early design stage. Relying on the defined application case, inputs of the tool are system configurations for the subsequent time periods, as well as delivery volumes agreed with the customers. The main purpose of the planning is to refine estimation on the batch sizes:
whereas previous tools of the workflow considered average batch sizes, in this case, they are calculated by matching order stream with a detailed system structure. Executing these plans in the discrete-event simulation model of the system, realistic operation costs can be calculated that consider additional information compared to the previous module, as inventory, personnel and also backlog costs can be determined in this way. The refined operation costs are meaningful feedback information that can be applied by the Reconfiguration Planning Tool to select the cost-optimal reconfiguration strategy. Besides, batch sizes can be utilized by the Assembly Cell Configuration Tool to evaluate and/or refine the cell configuration.
In the experiments, four different scenarios were analyzed with the planning and simula-tion models. In the first scenario(contractual), the contractual delivery volumes and frequency were applied (represented by variablesdpt), evaluating the solutions calculated by theAssembly Cell Configuration Tool considering ideal order stream. In the other three scenarios (Sc #1-3), delivery frequencies were increased by splitting the total volumes in smaller parts. In these sce-narios, the total volumes were the same, while delivery frequency was increased by 10−20−30%
subsequently. This resulted in smaller production batch sizes, more changeovers and thus higher operation costs, which might occur in real life. All experimental results are reported in Table 4.1. The results show that even in the contractual case, operation costs are higher than those considered by the previous modules. This refined information can be applied by the Reconfig-uration Planning Tool, if one assumes that contractual volumes will not change in the future.
A more conservative solution is applying the operation costs resulted by (Sc #1-3) scenarios, where smaller batch sizes and higher costs are resulted.
Based on the above results, a robust cell reconfiguration strategy could be identified that minimizes the overall lifecycle costs of the cell, including investment, operation and
reconfigura-73 4.5 Industrial application case
Period KPI Ideal Contractual Sc #1 Sc #2 Sc #3 copc [e] 10 863 13 714 14 030 16 028 17 184
u0
Batch P1 40 124 42 42 33
Batch P2 0 0 0 0 0
Batch P3 30 50 40 30 30
Batch P4 0 0 0 0 0
copc [e] 11 478 15 456 16 627 18 663 20 677
u1
Batch P1 0 0 0 0 0
Batch P2 0 0 0 0 0
Batch P3 30 53 53 40 40
Batch P4 35 42 33 33 25
copc [e] 14 637 17 779 19 406 22 452 21 772
u2
Batch P1 35 127 124 124 124
Batch P2 40 47 40 33 27
Batch P3 35 50 50 33 33
Batch P4 35 42 33 33 33
Table 4.1.Feedback on the resulted operation costs and batch sizes provided by theProduction Planning and Simulation Tool. TheIdeal includes the costs and batch sizes considered by the previous
tools, whereasContractual refines these costs. Scenarios Sc #1-3 assume that contractual delivery volume might change in the future resulting in more frequent deliveries.
tion costs. As discussed by Colledani et al. (2016), this robust reconfiguration strategy resulted in better solution than the so-called single path optimum that takes into account a single scenario of the tree, and looks for the best configuration in each time period. The robust solution, how-ever, considers all possible scenarios with their probabilities, and determines the reconfiguration strategy accordingly. The solution (cell configuration) selected for the case study is illustrated in Figure 4.5. This cell configuration results in the lowest overall lifecycle costs along the horizon, while meeting the requirements of all possible market scenarios without major changes in its configuration (reconfiguration), but it is enough to exchange the assembly modules when a setup takes place.
Important elements of the workflow are the feedback loops, implemented to refine a given configuration if requested after an evaluation with a subsequent tool, applying more detailed input data. Focusing on production planning, the results are applied to refine the system con-figuration with the Assembly Cell Configuration Tool if batch sizes differ from the previously considered ones. As reported in Table 4.1, the need to consider all details and constraints at the planning level could entail different feasible lot sizes compared to those used in the Reconfigura-tion Planning Tool. This could affect operation costs, moreover, might have impact also on the cell’s performance, if the actual lot sizes are smaller. This information can be exploited in the overall approach in two ways:
using the new estimated operational costs to identify a possible new optimal solution through theReconfiguration Planning Tool;
using the new estimated batch sizes to search for alternative configurations using Recon-figuration Planning Tool.