Multi-step Functional Process Adjustments to Reduce No-fault-found Product Failures in Service Caused by In-tolerance Faults

This paper proposes a methodology to eliminate or reduce No-Fault-Found (NFF) Product failure in field by adjusting manufacturing processes. No-Fault-Found (NFF) failures can be frequently characterized by in-tolerance faults, i.e., faults occur when design parameters represented by Key Product Characteristics (KPCs) and process variables represented by Key Control Characteristics (KCCs) are within the tolerance limits (in-spec) and therefore they cannot be eliminated by standard process control and adjustment methods. Literature in the area of statistical process and control are unable to monitor and adjust the NFF failures as these methodologies are based on design template itself. Also, all the adjustments strategies consider in process adjustment domain are single-step or adjustments are performed instantaneously. However, the single-step adjustment from process nominal (ud) to functional nominal (uf) is not always possible due to presence of system or resource constraints limiting the amount of adjustment that can be made in a given time such scenario in this paper is refer as multi-step adjustment scenario. This paper proposes a methodology for the multi-step functional process adjustment by identifying the adjustment paths that will minimize the total cost associated with product failure and adjustment cost. The proposed methodology includes: (i) discretization of space from ud to uf, orthogonally, with defined resolution; (ii) identification of warranty cost at each discretised node based on overlap between manufacturing distribution and in-tolerance NFF regions considering node as production nominal; and (iii) determination of the multi-step function process adjustment between ud and uf using recursive dynamic programming formulation. The methodology is illustrated through numerical simulation.