Paper Authors

Abstract

NOTE: The first page of text has been automatically extracted and included below in lieu of an abstract

Session 2248

Teaching Measurement Uncertainty in a Mechanical Engineering Technology Laboratory Maurice Bluestein Indiana University-Purdue University at Indianapolis

Abstract

In preparing students for careers in mechanical engineering technology, we have recognized that many of our graduates are hired into industrial positions involving performance and test. These positions involve setting up experiments and making mechanical measurements. Modem methods of data collection utilize measurement uncertainty analysis. We include concepts of measurement uncertainty in our course entitled “Instrumentation” which includes a laboratory segment. One laboratory experiment is utilized to help teach measurement uncertainty and is described in this paper. The experiment utilizes ordinary materials and industrial measurement devices. It shows the students typical sources of uncertainty in measurements. Statistical methods are utilized for both small and large samples. Propagation of measurement uncertainty is also shown as students make length and diameter measurements, determine the uncertainty of those dimensions and utilize the data to determine the uncertainty of a volume calculation. Students exchange data sheets and so have an opportunity for understanding human variation and to work cooperatively on a team project. Lab reports m required and graded by the instructor.

Introduction

In order to characterize the state or performance of a system, device, or process, engineers and technologists are taught to take measurements. ‘l%e data from such measurements are often used to make critical decisions and/or set manufacturing tolerances concerning that which is under test. All meamuements are subject to error so it is important to know how good is the data collected. This is described by the term “measurement uncertainty.” This is really a better term than “error” since the latter implies a true value of the measurement and them is usually no way to determine the “true value” of a measured parameter. Since the true value is not known, uncertainty values are always determined with a certain level of cotildence using the principles of statistics. Thus a measurement should typically be given as a mean value k its uncertainty for a given conildence level.

There are two types of measurement errors, aside from gross blunders due to either poor construction of the test or errors in observation or computation. Fixed or bias errors are inherent in the testing technique; examples are hysteresis, miscalibration and poor resolution. Random or precision errors are due to such things as human variation and electrical fluctuation; it usually follows a statistical pattern. Such errors can occur anywhere along the chain of measuremen~ from the sensor through to the recording of the data. The total uncertainty of a measurement combines both bias and precision errors in a root-mean-squm sense asl:

u = (B + P)””s 2 (1)

{hxd~ 1996 ASEE Annual Conference Proceedings ‘.,,,IZI13J

• Citation

• Format
  • APA
  • APA - LaTeX bibitem
  • MLA
  • MLA - LaTeX bibitem
  • Bibtex
  • EndNote - RIS

Bluestein, M. (1996, June), Teaching Measurement Uncertainty In A Mechanical Engineering Technology Laboratory Paper presented at 1996 Annual Conference, Washington, District of Columbia. 10.18260/1-2--6326