Reliability : A step towards Business Excellence

1 Reliability : A step towards Business Excellence Using Minitab 16 Reliability : A step towards Business Excellence By Rajeev Chadha, Lean Six Sigma Master Black Belt Mosaic Potash ULC Canada In association with Minitab 16 Definition of Reliability Impact to the Business Data distribution characteristics Exercises Break Compressor Case Study Module Overview A - Minitab is a powerful tool for Reliability /survival analysis! How Do You Define Reliability ? The probability that an item (or system) will perform its intended function for a given period of time under a given set of conditions Or Achieving expected performance For consideration Pumps and Process Equipments Entire Manufacturing processes Customer invoicing & Services Month end Accounts closing Laptop computers Deep well blow out preventer Where can you use Reliability Analysis A - Reliability is not just for maintenance Reliability Why Do We Care ?

2 How Does Reliability Works? Most of the North American Blue chip Companies non capital spending for turnaround, R&M, CL, and internal labor are:- FY11 - $349 MM for Energy Sector FY11 $189MM for Mining Sector Would it be beneficial to save these $$ for your company and for your career? Bottom line dollars Reliability Why Do we Care? Reliability If You can Predict it You Can Prevent The Estimated MRO Inventory cost of an Average US Co. is 15-20% of the Worth just because our systems are not Reliable and Predictable. Creating a Community of Problem Solvers Reliability Management :- An Organizational Approach Probabilistic Measure Reliability of sub systems through failure Analysis Evaluate interactions between sub systems Predict Reliability of the entire system without full understanding of the details of what is causing the variation it s too complex Deterministic Religious vigilance and attention to detail constant and meticulous search for errors and defects through inspections and testing strict and thorough implementation of corrective actions Reliability Two approaches Reliability Bath Tub Curve Reliability Time Scale For Reliability , Time does not have to be measured in only time!

3 It can also be measured in:- Mileage Man Hours Cycles Start-ups Thermal Stress Mechanical Stress Reliability Hazard Rate The one formula to describe it all > 0 (scale or characteristic life) K > 0 (shape) H (x,k, ) = (k/ ) (x/ )k-1 A -Minitab calculates it for you! Hazard Function graph Defines how failure rater vary over life failure rate vs days Exercise To Calculate Reliability for Better Asset Utilization Let us take the example of Replacement of Hi-tension (Voltage) swing wires in Mines and Dredges:- Objective- To increase the Swing Wires life to improve the Asset utilization Reliability Parameters Reduce Employee Exposure to Hazards arising from changing Hi-tension wires, Asset failures and unsafe conditions. Increase Productivity of Assets by decreasing downtime in Repair of Wires.

4 Reduce Operating and Maintenance Expense arising from frequent wire failures. Asset Utilization (Cable Life) Exercise Asset Utilization (Cable Life) Exercise Asset Utilization (Cable Life) Exercise Asset Utilization (Cable Life) Exercise Asset Utilization (Cable Life) Exercise Asset Utilization (Cable Life) Exercise Reliability If You can Predict it You Can Prevent it ! Other Uses :- Predict Future Injury Rate Predict Future Warranty Costs and Customer Returns (Cost of Returned goods) Measure, Predict & Optimize Utilization of Your Assets Man Hours worked Between Recordable Injuries Demonstrate Improvements & Draw conclusions about your past performance. Additional Uses of Reliability analysis A Case Study On Reliability Analysis Compressor warranty claims: Case Study (2009-10) Suppose You work as a Service/ Reliability Manager for an appliance manufacturer and want to know the number of refrigerator compressor warranty claims you expect to see over the next five months.

5 Why Predictive Modeling? With Weibull Modeling you can Predict Future Compressor failures so that:- You can plan & Schedule after sales services ahead of time. Arrange stock of Replacement Parts (Compressors). Inform Manufacturing to improve process to reduce compressor failures and control warranty costs. Predict Future Compressor failures Use Weibull distribution to model time-to-failure The Past warranty claim data shows that Out of 12,000 units in the field, 69 compressors failed in year 2009-10 A -Remember Minitab is there to Help you! Steps in Minitab to Predict Future Compressor failures step 1. Arrange Data in a required format on Minitab worksheet Compressor Warranty Claims Data from past 12 months (2009-10) Steps in Minitab to Predict Future Compressor failures step 2.

6 Open worksheet and Choose Stat > Reliability /Survival > Warranty Analysis > Pre-Process Warranty Data. Input Pre-Process Warranty Data Steps in Minitab to Predict Future Compressor failures STEPS 3 AND 4:- In Shipment (sale) column, enter Ship. In Return (failure) columns, enter Month1-Month12. Click OK. Minitab Calculates Start Time, End Time & Frequencies of Warranty claims Steps in Minitab to Predict Future Compressor failures step 5 In Start time, enter 'Start time'. In End time, enter 'End time'. In Frequency (optional), enter Frequencies. step 6 Click Prediction. In Production quantity for each time period, enter 1000. Click OK in each dialog box. Open Warranty Predictions in Minitab Steps in Minitab to Predict Future Compressor failures step 7- Click Prediction. In Production quantity for each time period, enter 1000.

7 Click OK in each dialog box. OR Stat > Reliability /Survival > Warranty Analysis > Warranty Prediction > Graphs Minitab Output and Results Distribution: Weibull with shape = , scale = Estimation method:Least squares (failure time (X) on rank (Y)) Summary of Current Warranty Claims:- Total number of Compressor units shipped - 12000 Observed number of failures - 69 Expected number of failures - 95% Poisson CI - ( , ) Number of units at risk for future time periods 11931 Minitab Output and Results Production Schedule of Compressors :- Future time period (MONTHS) 1 2 3 4 5 Production quantity (UNITS) 1000 1000 1000 1000 1000 Table of Predicted Number of Failures:- Future Potential Predicted Time Number of Number of 95% Poisson CI Period Failures Failures Lower Upper 1 12931 2 13931 3 14931 4 15931 5 16931 Predicted No.

8 Of Compressor Failures Max Failure =104 Min Failure = 67 Max Failure in 1st month= 23 Interpreting results about the warranty claims Out of 12,000 compressor units Sold in 2009-10, approximately 71 units are expected to fail in next five months in 2011. OR From the failures plot, you can conclude with 95% confidence that between approximately 67 and 104 units are expected to fail in the next five months. So Now Your Life is Easy ! You Plan and Schedule Max. 13 replacement Compressors in 1st Month, Max. 30 in 2nd . CALCULATE the projected warranty costs. Carry (min. 20 max. 30) Replacement Compressors in stock. Advice Manufacturing to improve their process where 11,931 Out of 12,000 units are at risk. Thank You For your Interest in Reliability Questions on Reliability Analysis?

9 Questions on Minitab Software? *Apply for 30-Days Free Minitab trial online at The World Trusts MINITAB Minitab is the leading provider of software for statistics education, Lean Six Sigma, and quality improvement projects. For nearly 40 years we have been helping world-class organizations analyze problems, transform their Business , and train their students. Additional Resources For Reliability Analysis ASQ-Saskatchewan Acknowledgements I greatly appreciate my MOSAIC Corporate Six Sigma Team and ASQ-Saskatchewan team for helping me prepare this workshop presentation for ASQ-Canadian Quality Summit and Mining Gala 2012. Last but not the least I would like to thank Mr. Jeff Adams of Minitab for supporting this project for training and educating our global quality community.

10 With Regards Rajeev