Knowledge base

Measurement system analysis (MSA)

Understanding Measurement System Analysis: Ensuring Data Reliability for Process

Reliability of collected data is a critical need in the domain of process control and improvement. The need to be able to measure processes, a key component of quality improvement efforts, has taken center stage in recent years. Ideally, a perfect measurement system should produce the same results every time.

Given that all measuring processes are subject to uncertainty and variation, there are two bases of this variation:

  1. Process Variation: It is the result of inherent variations within the process.
  2. Measurement Process Variation: The variation that arises from inherent or from the detection method such as the measurement equipment or the operators involved in measurement processes.

Discrepancies in measurement can be due to both humans and the measurement system. The system should be tested for the essence of MSA, which is to examine how the measurement outcomes express process dynamics accurately.

In simple and valid terms, the term “representative” has two primary categories often termed as measurement error:

  1. Accuracy: It also might just refer to as correctness in output. No difference between the measured values and the real value of errors is referred to as the lack of accuracy or bias. This can be controlled by several measurements from a known standard pattern.
  2. Precision (variability): The ability to reflect the right standard deviation choice is a mechanism of errors. There is a chance of introducing systemic error, and a precise measurement judgment is how the output measurement occurs in any line.

The errors of measurement of precision are further subdivided as:

Repeatability: It is observed as the variation when a single tester performs the measurements under exactly the same conditions and with the same instrument; fluctuations of the same person in the same situation are assessed; and

Reproducibility: This is the variation in measurements when different testers conduct the measurements using different conditions or other measuring instruments. It incorporates long-lasting experienced differences, even in routine conditions and with personnel.

Key criteria for a reliable measurement system include:

Accuracy/Relevance: Minimal deviation from the standard reference.

Repeatability: Similarities in measurements when measurements are done under the same conditions.

Reproducibility: Similarities when measurements are done under different conditions by different people.

Stability: when there are similarities with time included.

Linearity: Minimal bias variation with changing measurement scales.

Distinction: It is the ability when you can differentiate small differences in the same.

Although the approach to ensure data reliability is a significant step in data analysis, it is important to give a clear comparison of the received data to the actual data. Thus, when analyzing completion times from systems, it is essential to warm them to on-site observations at the Gemba.

Wrapping Up

After knowing and assessing the characteristics of measurement systems, the organization can create reliable data to enable effective process improvements.

Online Lean courses
100% Lean, at your own pace

Most popular article