Abstract
The information product designed and developed in this study primarily aims at "the qualitative analysis of production results" within a recipient company, using as the basis of analysis the steps required by the "statistical control of production quality, using the method of large selections" and, as a design support, the rules and concepts provided by object modeling. Specifying requirements: In designing the "quality control system for bearing production", the specified requirements are the following: ? the automatic calculation of technical-quality parameters; ? the graphic representation of the measured characteristics; ? determining the quality of a batch of products (bearings), by comparing the calculated parameters with parameters defined in a standardized manner in the Production Technical Norms or the Technical-Quality Parameter Charts; ? providing reports to substantiate conclusions and decisions. The main objectives are: 1. The quality control of production results; 2. The statistical analysis of experimental data (taken from measurements of the characteristics of the manufacturing process) by implementing the automated steps according to the procedures imposed by the method of large selections for the quality control of production results. Candidate objectives are: 1. Identifying the characteristics of experimental data in order to determine a "well-established distribution law" - in both the form and the parameters - for the measured characteristic; 2. The study of quality characteristics for data representation according to a well-established distribution law (the "normal distribution"); 3. determining the distribution form of the sampled data and identifying the salience characteristics thereof - by going through the following sequence of steps: a. an identification of the distribution form (normal) of the experimental data by the use of various forms of graphical representation: time graphic, statistic allotment card, diagram, histogram and the polygon; b. the implementation of tests checking the compliance of experimental distribution with the normal lawwe will use the mathematical and statistical support introduced with the "?2" test; c. a calculation of the normal distribution parameters (a type of distribution determined as a result of checking compliance through the "?2" test) in order to correctly notice the system characteristics and to determine the accuracy of data taken from the production process; we will calculate the "trendency parameters" - represented by "the average", "the median" and "the modulus" and "the diffusion parameters" - represented by "amplitude", "dispersion" and "square mean deviation"; 4. verifying the randomness of mesurement results, in order to identify the types of errors that impact the production process; we will propose to demonstrate (with a probability of 95%), the initial hypothesis according to which "the manufacturing technological process is affected only by random causes" - not by "production causes"; a. in order to determine the random distribution of the values noticed for the measured feature, we will implement the concepts proposed by the "Iterations test", taking into account the two criteria of analysis: the "number of iterations" and the "(maximum) length of an iteration"; 5. identifying and eliminating gross errors that negatively impact the production process, through the identification of values measured as "inconsistent" with the product manufacturing technical standards; a. in order to find aberrant measured values (and thus eliminate "scraps" - products with a non-compliant measurable characteristic), the "Romanovsky test" will be used for the study; b. optionally, for further verification of the results, we can also implement the statistical analysis support provided by the "Grubbs-Smirnov test"; 6. the rough estimate of parameters for the entire population, starting from the ones analyzed for a representative amount of measured results; a. we will complete the process of statistical control for the general population, by Hence, the analyst is more than a regular user of the automated control product, as, based on his/her expertise, (s)he establishes and updates the parameter tabular values issuing from statistical tests. These values are essential in the process of analysis, as, based on them, we determine the implementation result for each of the three analysis tests. Also, the analyst's primary role is to decide, after the execution of each iteration, on the direction in which the analysis will continue, or whether the current execution stage needs to be interrupted and the process resumed, whenever appropriate. |