Client Services  /  Statistical Engineering  /  Case Studies

Client Services

» Statistical Engineering Case Studies


Statistical Problem Solving A Department of Defense Manufacturing Company Abstract The turret assembly when assembled into the M-ATV vehicle was not functioning as intended. The problem as presented to the OEM was that the system level TIR was excessive at the stop ring in the vehicle, and that if controlled to less than .060" at the system level, the unit would function properly. We were asked to converge down to the source of the problem and statistically prove root cause to a confidence level of 95%. Challenge Inspect and segregate fully assembled BOB and WOW vehicles so we could swap components between the units. The delivery schedule for M-ATV was aggressive and holding vehicles was difficult. Working at the extremes of the distribution of interest allowed for small sample sizes to be used. Execution

  • Disassemble and reassemble BOB and WOW units while maintaining rotational origin to determine if the assembly process was significant (Torque and tightening sequence).
    It was not.
  • Disassemble and reassemble BOB and WOW units while rotating the internal stop ring 90°. This action caused BOB to degrade and WOW to improve they became equal in run out values. This shows the main effect lives within the "parts" and not process.
  • A two level (BOB and WOW) three factor test was designed with 16 replicates that contrasted the best and worst suspected values of the parts. The roof was run in a flat state as opposed to being significantly bowed, the bearing was run at high and low levels of torque, and the stop ring was included with less than .030" run out at the BOB end and with over .100" run out at the WOW end.

Result The test showed at 95% confidence that the stop ring was the main effect and neither the roof or bearing at either end of their respective distributions was significant. All effort at that time was focused on improving the dimensional capability of the stop ring.

Download the PDF version.

SE Training Program A North American Powertrain Supplier Abstract An engine and transmission component supplier in North America was interested in increasing their technical problem solving competencies within their product manufacturing and engineering group. Challenge The client's objective was to develop 16 certified Statistical Engineering Apprentices. The TPS training model combines our interactive training class with full time coaching support as we work with the Apprentice candidates to solve impact issues affecting the clients bottom line. Execution 16 engineers were selected by management to attend the TPS Apprentice level training class. At the end of the 2.5 day class session, TPS assisted the candidates with project selection. This ensured that each project opportunity selected would generate significant savings or improvement in the companies business metrics. Eight impact projects were selected and completed as we trained 16 candidates. Result All candidates that completed a Statistical Engineering project within the anticipated program timeframe earned our accredited SE Apprentice Certification. The overall cost of the training program was less than 8% of the total savings and cost avoidance realized from the projects selected.

Download the PDF version.

Technical Problem Solutions LLC
Achieving Results © 2007 -  |  Site Map  |  Follow Technical Problem Solutions on LinkedIn