# Statistical Tools for Maintenance Engineering

on May 12, 2013

Measurement and analysis drive initiative, innovation, development, and lateral thinking. Statistical tools help in analyzing measurements in many ways.

Maintenance of relevant Statistical Tools

• Failure Rate Percent Per Year (FRPCPY)
• Bench Marking
• ABC Analysis
• Pareto chart
• Bath Tub
• Statistical Analysis ( Mean, Mode, Median and Standard Deviation )

# Failure Rate Percent Per Year (FRPCPY)

This is a compilation of data on a uniform base of 100 and for 12 months. This is commonly used in Indian Railways to compare the performance of the same item maintained at many locations and over a period. Para 2.8 of the book titled ‘Electrical Fire and Failure’ written by Mr. A A Hattangadi describes it as follows

• FRPCPY = (No. Of failures * 100*12) /Population*Period in months
• This formula provides uniformity in comparing the data of different units, in a different application, of different make etc.
• RFPCPY data are compiled by RDSO of the different work units over the same period. This data is very important tool for identifying the problem that can be solved.
• FRPCPY data of different units shall be examined to formulate strategies for the future.

Example

 Sl No. No. of failures Population Period(months) FRPCPY 1 5 15 16 25 2 60 1200 15 4 3 1030 1500 80 10.3

# Benchmarking

Benchmarking is the process of measuring performance parameters, cost, services, and processes against those of organizations known to be leaders in one or more aspects of their operations. The efficiency and Research Directorate of the Railway Board have issued benchmarking data on manpower per unit output for all the units of Indian Railways.

• Benchmarking provides necessary insights to help you understand how your organization compares to similar organizations, even if they are in a different business or have a different group of customers.
• Reasons for excellent and average performance is generally not supported by the analytical tool, and are more reasoned subjectively.
• Better than last year is commonly followed as benchmarking tools

# ABC Analysis

Initially, this tool was designed for material management techniques categorized as:

• A Items: Very tight control
• B items: Medium tight Control
• C Items: Normal control

This analysis is very helpful in prioritization of problems concerning cost, failure implication, downtime of assets and can be done in many ways, such as

• Defects for which the Cost of repair and down time high: The cost of repair and downtime is high when it involves lifting the rolling stock, coaching stock marked sick requiring withdrawal from service. Such defects shall be identified with a valuation of each defect for its role in downtime and cost of repair.
• The implication of defect on Line failure
• Cost of maintenance material: Listing the cost of material

# Pareto Analysis

This is a statistical technique in decision-making.

• It uses the Pareto Principle (80%-20% rule), the idea that  20% of the work generates 80% of the benefit of doing the whole job. Or in terms of quality improvement, a large majority of problems (80%) is produced by a few key causes (20%).
• This is also known as VITAL FEW TRIVIAL MANY OR PENNY WISE POUND FOOLISH.
• The challenge is to find out the vital 20%
•  It is done by statistical analysis

# Bath Tub

The technique is used for reliability study and performance dependence with time.

A typical time-dependence performance curve looks like

The most important part is to identify the interval of the three different stages of the failure patterns. The interval will depend on the type of equipment. The wear-out failure will establish the life of the equipment or component. There are possibilities when a sudden jump of failure is observed during constant random failure due to manufacturing/maintenance or operating defect and taken as a wear-out failure, this will result in an evaluation of wrong life, and replacement will start prematurely. The corrective action for this increase in failure, which can be addressed by replacing one of the components, is necessary to analyze otherwise, the maintenance costs will be very high unnecessarily. It may be of interest that wear-out failure shall be performed on a sample locomotive/asset and condition monitoring in very controlled conditions at frequent intervals.

Top