Method for classifying the quality of a product to be tested on a test stand

摘要

Data which are simple to measure (current, voltage, speed of rotation) are to be used for classifying products to be tested as "good" or "faulty" and with respect to their individual quality features. In this connection, the products classified as faulty should receive a note regarding type, location and possibly magnitude of the fault. The classification method should not permit any uncertainties about the quality features determined. For this purpose, m measurable features of the product to be tested are recorded which are related to a total of n parameters (P1 to P6) via mathematical equations describing various functions of the product. The parameters are variable in dependence on the quality features of the product to be tested and are calculated on fault-free products and on products with predetermined faults and are input into a memory. The values of the n parameters (P1 to P6) of the fault-free products (diagram 1) are allocated to a class of good products and those of the products with predetermined faults (diagram II) are allocated to classes of faulty products and sequentially stored in the memory. The values of the n parameters of a product to be classified passing through the test stand are calculated by a computer by means of the m features measured in the test stand, using the mathematical equations, and compared with the stored sequences. When the calculated sequence corresponds to a stored sequence ... Original abstract incomplete. IMAGE