Analysis of the quality control problems using the WinQSB software product.

Hamat, Codruta Oana ; Amariei, Olga Ioana ; Dumitrescu, Constantin Dan 等

1. INTRODUCTION

The industrial development has generated the necessity that, in the manufacturing systems, the information and data collected about quality should be accessible to the decision makers, with the possibility of co-ordinating the actions of different compartments of the company, so that the information could contribute to the identification and elimination of the causes provoking defects.

These conditions imposed, as appropriate methodology for obtaining, processing and analysing quality information from the industrial processes, the grounding of the quality control on methods of mathematical statistics.

A cause of the more difficult affirmation of the statistic control in the industrial practice was represented by the fact that the abundance of the statistic methods presented in the literature accessible to the industrial management was not accompanied by examples of applications (Hamat, 2003).

Consequently, the purpose of this paper is to present the possibility of using a software product, i.e. WinQSB, which allows the solving of the quality control problems simulation.

2. PRESENTING THE PROBLEM

A factory produces pistons. The Quality Control Service intends to introduce the statistic control, and in this respect one removed from the manufacturing process, every 4 hours, samples made of n = 5 pieces. The quality characteristic is the diameter of the pistons, corresponding to the value of 76.12 [+ or -] 0.05 mm. The data collected are introduced in a table (figure 1).

For each problem of quality control, the module predefines a set of 14 rules, which can be modified using the options of the Edit menu.

The Quality Control Chart module of the WinQSB software provides 21 types of graphs for variable data, 15 types of graphs for attributive data, histograms, Pareto analysis, curves of the operational characteristics etc (Mihalca et al., 2003).

For this type of problem we may apply the statistic control, using the arithmetic mean and amplitude method. The arithmetic mean measures the stability in time of the process adjustment, while the amplitude measures the stability precision in time of the manufacturing process (Olaru, 1999).

We first perform a statistic analysis (average values, median, dispersion, maximum and minimum values within the group) of the recorded data, by sub-groups and on the whole. The results obtained are shown in figure 2.

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

We continue the analysis with the elaboration of certain representative graphs, available in the Gallery menu, first the control charts for the data obtained "through measurement" which can by of several types: control charts for means (X charts--figure 3), control charts for amplitude (R charts- figure 5) and control charts for the sample standard (Axinte, 2007).

[FIGURE 3 OMITTED]

We remark in figure 3 that sample 4 is below the inferior limit of 76.1032. We can obtain more information about this sample by making appeal to the Analysis option--Individual Point, in a window of the type shown figure 4. We notice that the average / mean of this subgroup is not under control, because it violates the rules number 2 and 14 in the rules table offered by the software.

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

The CUSUM graph for average / mean values--CUSUM Chart for Mean, presenting the dispersion of the average dispersion from the central average / mean and we verify if from this viewpoint the values fall within the normal limits. The graph is presented in figure 6, where we remark that all the points are represented in green, and consequently there are no uncontrolled values.

This type of graph is one of the several different modes provided by the product for the verification of the compliance of values with the normal limits (Amariei et al., 2008).

[FIGURE 6 OMITTED]

3. CONCLUSION

From the analysis of the graph corresponding to the mean variation it results that the sample number 4 recorded on 03.07.2008, 16:00 hours, is not under control, and thus we consider that the manufacturing process has not been adjusted appropriately. Moreover, all the other points are situated between the superior and the inferior limit, but we remark a rather large dispersion of the process.

Analysing the graph corresponding to the amplitude variation results that all the points are situated below the superior control limit, indicating the fact that the process provides the necessary precision.

In the present situation, the manufacturing process is considered inappropriate. We must analyse the causes having triggered deviations of the adjustment, using the cause-effect diagram. In order to bring the manufacturing process to the normal operation condition, corrective actions are initiated and implemented, actions meant to eliminate the causes generating defects, after which the statistic control is continued.

4. REFERENCES

Amariei, O.I., Dumitrescu, C. D., Gillich N., Hamat, C., Malos, C.R.(2008). Optimizing a manufacturing system after analysis of Extend[TM] and WinQSB simulation environment, 2008 IEEE International Conference on Automation, Quality and Testing, Robotics AQTR 2008--THETA 16th edition--May 22-25, ISBN 1-4244-2577-8, 2008 Cluj-Napoca, Romania,

Axinte, E. (2007). Elements of quality insurance in industrial engineering, Demiurg Editorial House, Iasi

Hamat, C.O. (2003). Quality Management, Orizonturi Universitare Publishing House, Timisoara, ISBN 973-638-047-5

Mihalca, R; Fabian, C (2003). The use of the software products--Word, Excel, PMT, WinQSB, Systat, ASE Publishing House, Bucharest ISBN 973-594-320-4

Olaru, M. (1999). Quality Management, ASE Publishing House, Bucharest, ISBN 973-946-202-2