Measurement Errors
At some point in the manufacturing process you’ll find that differences in measurement results (measured values) will begin to show up due to one error or another. A variety of issues can contribute to a measurement error. As a result you need to have countermeasures in place so that you’re prepared. If you encounter a measurement error, it’s vital that you examine the cause of the error very carefully before using a countermeasure. Keep this in mind if you wish to maintain the quality of your product.
Types of Errors in a Measurement System
Gross Errors – gross errors occur when a mistake is made while recording data results, using a measurement instrument, or calculating measurement. Let’s take for example, a person who takes a reading from a pressure gauge. The person does not record the reading in any way shape or form. They commit it to memory and turn around and present it wrong later on. These blunders in the reported data will be taken as fact and may be used as a calculation to find the final results. Because the base data was wrong the final results will in turn be wrong.
Blunders – blunders, much like gross errors, tend to be caused by recording down the wrong information or forgetting a digit while reading a measurement. These types of blunders should stick out to anyone checking over the work of someone else, thus having a second person overlook the work of someone else should reduce the possibility that the wrong type of information will end up in the final analysis of the data.
Random errors – random errors can be caused by sudden changes in the work environment. These changes can include sudden noises or fatigue in the person working. An excellent example of a random error would be an unexpected change in temperature or a fluctuation in voltage. To reduce these errors take a large number of measurements and use the average of those results.
Measurement errors – a measurement error occurs when there is a variation detected of the true value of a measurement. In most cases, measurement errors are comprised of systematic and random errors. To get a better idea of what a measurement error is let’s look at an example: if an electronic scale is loaded with 1kg of standard weight and the reading is 10002 grams, then the measurement error is = (1002 grams – 1000 grams) = 2 grams.
Causes
Measurement errors can be broken down into three separate classifications:
Errors caused by the measurement system – an error that originates from within the system itself. This can be caused by worn components within the system.
Errors caused by the user – an issue that occurs due to human error. This can include mishandling of the measurement system, user-specific methods of reading the scale, and the skill of the user.
Errors caused by environmental conditions – an error caused by environmental factors. This can include deformation of the target due to a change in temperature and measuring in an area with unpredictable brightness levels.
To fix a measurement error caused by the measurement system or environmental conditions, you have to find the cause of the issue and implement the proper countermeasure. User related errors can be more difficult to deal with due to the human element. You may have to replace or retrain the employee who was at fault for the error.
Eliminating Errors
It’s very possible to eliminate measurement errors for items that are measured multiple times. Keep in mind, that the likelihood of catching measurement errors during the inspection process is highly unlikely. To achieve higher accuracy during the measuring process, it’s vital that you find the exact cause of the measurement error so that you can implement the proper countermeasure.
Here is a list of solutions to eliminating measurement errors:
Eliminating intrinsic measurement system errors – confirming accuracy before starting your work and checking once more upon completion, regular maintenance, replace components, and regular calibration.
Create a system that can compensate for individual differences – switch to inline measurement for items that require measurement at more than one location, standardize the measurement method, improving the skillset of employees through improved training methods, use more automated processes to eliminate human errors, and replacing contact-type measurement systems with non-contact-type measurement systems.
Change Judgment criteria – constrict the tolerance range for a more stringent acceptable judgment range.