Form error evaluation using L1-approximation Original Research Article

In general, systematic and accidental non-conformities contribute to the non-deterministic nature of the manufacturing processes. The measurement data generated during production and inspection should be further processed to get the desired information contained in it. The measurement data can be considered as the sum of the desired information and certain random errors and variations. Processing requires a method that reduces the effect of wild non-conformities giving due importance to the desired information present in the original measurement data. In this paper an efficient L1-approximation is presented and used for form error evaluation of engineering surfaces to show its effectiveness when the measurement data contains wild points. The result is compared with the least-squares solution. The use of L1-approximation for form error evaluation will also help to identify the wild points present on the surface of the part that can be compensated by further operation(machining) to produce a better quality part.