【正文】 osts and necessary accuracy of the measurements Business process control Measurement System Analysis Goals – Recognize, understand and minimize sources of variation that could impact on measurement result. – Ensure a high measurement quality so that no false statements are made on the process capability. A reliable measurement system must satisfy the following requirements: ?Accuracy ?Repeatability ?Reproducibility ?Stability ?Linearity ?Granularity or Discrimination / Resolution Business process control Measurement System Analysis Repeatability Business process control Measurement System Analysis Business process control Measurement System Analysis Granularity or Discrimination / Resolution The measurement system should be able to recognize small changes or differences in the parts to be measured (a good measurement system for continuous data shoud be able to distinguish at least 5 different measurements / parts). Business process control Pareto Chart Steps – Depict the data in categories. – Add up the frequencies of the individual categories so as to select the correct standard on the Yaxis in the Pareto chart. – Depict the bars representing the categories true to scale: From left to right according to their frequency and in descending order of importance. – Depict the curve that shows the cumulative frequency. Goals – Depict the focal points and thus set priorities for discrete data. – Concentrate on those vital few causes whose optimization will generate the greatest impact (80:20 rule). Business process control Histogram Steps – Summarize continuous data in categories or classes. Rule of thumb: Number of categories (c) is c = √n. – Record the frequencies of actually occurring categories in bars true to scale. The relevant interval for the data to be depicted stands on the xaxis. The frequencies are on the yaxis (absolute or relative percentage values). Goals – Depict the distribution of continuous data. – With this aid, find out if a process is centered in relation to customer requirements and if the spread lies within the specification limits. Business process control Centering and Spread, Customer Requirements, Specification Limits Business process control Box Plot Steps 1. Record the maximum and minimum values true to scale and connect them with a vertical line. 2. Depict the median as a horizontal. 3. Plot a box between the 1st and 3rd quartile. Goals – Depict the location and spread of a continuous data set. – Undertake a quicker parison of different data sets (. pare suppliers, machinery). – Summarize a data set in graphs and charts, featuring the median, the 1st and 3rd quartile (which amount to 25 % and 75 % of the data, respectively), and the extreme values. Business process control Box Plot Goal Depict a (linear or nonlinear) relationship between two metric variables Business process control Tools that will help to analyze trends are run and control charts. A run chart is a tracking of data as it rises and falls over time A control chart is a run chart with mathematically determined limits added. These limits differentiate the type of everyday variation that is built into the process from truly unusual events Business process control Running chart Business process control Business process control Control chart (KPI is inside control limits) UCL – upper control limit LCL – lower control limit Business process control Control chart (KPI is outside control limits) Business process control Process Performance Business process control Process Performance Business process control Process Performance Rolled throughput yield: Determines the probability that a unit runs through the entire process without defects. The total yield is calculated by multiplying the single subprocess yields. Normalized yield: Determines the average yield per process step Business process control Process Performance Goal Determine the process performance when no specification limits are set (focus on attaining target value). Steps – Transform all values (x) into a new variable X: X = X current – X target. – Calculate the interval range between X current and X target without considering outliers. A span (90) is usually used. The smaller the interval range the more effective the span. Span (90) = Span / Percentile Distance Business process control Process Performance Goals – Identify the relationship between the customer specification limits (tolerance) and the natural range of the process (Cp value). – Ascertain the centering of the process (Cpk value). Cp and Cpk Values Business process control Analyze Goals – Support the team in brainstorming possible causes. – Visualize the possible causes. – Depict the relation between possible causes. – Focus the team on the possible causes for the problem (and not on symptoms). – Create a shared understanding in the team for the underlying problem. Cause Effect Diagram Business process control Spider chart Business process control Most anizations take a formal approach to strategic assessment to ensure they have considered all missioncritical issues Business process control Kano model of quality differs three levels of customer needs and satisfaction: 1) Funamentals (Dissatisfier, Must be) – the absence of a “must be” is sure to provoke an immediate negative customer reaction 2) Linear (Satisfier) – features that may not be absolutely necessary to the customers, but the more they see of them, the better they like it. 3) Exciters (Delighters) a special feature that the customer doesn’t expect, but when it is present it provokes an immediate positive reaction Business process