Application of known and original statistical control methods in the process of production of medical devices for the respiratory support

Already known and original statistical process control methods were used to check the stability of the manufacturing process of one product. The aim of the study was to analyze four characteristics related to the product. Two of them (called Critical) were absolutely continuous and two (called Non-critical) - discrete. The analysis of the latter attributes required data transformations. The analysis itself was only indicative because these characteristics were not analyzed by manufacturer. The following known methods were used: Shewhart's control charts for the control of normal distribution mean, variance, range and for the probability of binomial distribution, regarding two-dimensional case - Hotelling's control charts. The main results of the analysis were as follows: the probability distribution of individual and also averaged measurements of the second critical characteristic were not normally distributed. Original control charts using left-truncated normal distribution for individual measurements and approximation by log-normal distribution of averaged measurements were created. Similarly, original two-dimensional control charts were proposed. The manufacturing process itself in Phase II was not completely stable regarding characteristics „Critical1“ and „Non-critical1“. Regarding other characteristics, the process was in control. All calculations were performed graphs were plotted using R environment