Setting up a control chart

To set up a control chart, several replicate analyses of a standard reference material are made periodically, such as once a week. If possible, the true value of the reference material and the standard deviation of the analytical method should be known. If not already known, the correct values can be estimated accurately by repeated analysis. The reference material must be very stable and should approximate as closely as possible to the composition of actual samples to be analyzed. 

The idea of monitoring accuracy and precision was developed by Walter Shewhart [18] and the target value here was the known concentration of analyte in a control standard. The range graph monitors the precision, and the target value is the capability that it is necessary to establish in order to set up the control chart. Process capability will be limited by the random errors involved in measurements rather than error in preparing the standards. 

The concept of a confidence interval may be used to set up a statistical control chart on the mean. Let us consider the reactor from Example 1.14. Suppose we want to use the results of the 11 runs to establish a procedure for operation of the reactor in future runs. 

When all major sources of errors are eliminated from all steps of the procedure and only small random variations remain, the laboratory produces results that can be considered as being under statistical control. This means that it can be predicted that results produced on the same reference sample in reproducibility conditions will remain within a certain limit of precision. Having reached this stage, the analyst can set up a statistical control scheme, which verifies that these conditions are maintained. Such systems, which have been developed for the control of production lines, have been adapted to analytical chemistry they are known as control charts (see section 2.4). 

RMs may be used for the verification of the longterm reproducibility of a method by setting up control charts. A control chart is a graphical representation of how results of RM analyses vary in time it is used to detect possible systematic fluctuations (e.g., drift) in a method. The current practice is that one RM should be analyzed with 10-20 unknown samples and the results plotted on a... 

The zone control chart (also known as the J-chart) is a control chart for the mean that combines features of the Shewhart chart and the cusum chart. It is simple to use, but effective. First it is necessary to establish a value for a, as was done in Example 4.7.1. Then the chart is set up with horizontal lines at the target value, /r,... 

The control chart is set up to answer the question of whether the data are in statistical control, that is, whether the data may be retarded as random samples from a single population of data. Because of this feature of testing for randomness, the control chart may be useful in searching out systematic sources of error in laboratory research data as well as in evaluating plant-production or control-analysis data. ... 

The first group of tests is carried out on specimens generally fabricated into a dumb-bell shape, with forces applied uniaxially. The usual apparatus consists of a machine with a pair of jaws, which during the test are moved relative to each other, either together or apart, in a controlled manner. A chart recorder is employed to give a permanent record of the results obtained, so that the force at fracture can be determined. Whether this kind of set up measures tensile, compressive, or flexural strength depends on how the sample is oriented between the jaws, and on the direction that the jaws are set to travel relative to one another. 

The analysis of quality control samples is an important activity for laboratories and to make the most of the data, control charts should be used. This chapter has discussed a number of common types of control chart and described how they are set up and interpreted. 

For case (1), a group control chart could be maintained on all streams but it suffices to plot only the two streams corresponding to the highest and lowest values. The chart limits are set up for a single stream and run rules should not be used. If the output of the heads is highly correlated, then a single chart on one stream may be used as a surrogate for all streams. 

A key aspect of statistical quality control is evaluating the ability of a production process to meet or exceed preset specifications. This is called process capability. Simply setting up control charts to monitor whether a process is in control does not guarantee a high process capability. To produce an acceptable product, the process must be capable and in control before production begins. The process capability refers to the ability of the process to produce an output characterized by key product quality features that are within the imposed specification limits. This determines the level of nonconformities produced by the process. 

An entirely distinct approach to estimating uses control chart principles (see Chapter 4). We have seen that these charts can be used to monitor the quality of laboratory methods used repeatedly over a period of time, and this chapter has shown that a single calibration line can in principle be used for many individual analyses. It thus seems natural to combine these two ideas, and to use control charts to monitor the performance of a calibration experiment, while at the same time obtaining estimates of The procedure recommended by ISO involves the use of q =2 OT 3) standards or reference materials, which need not be (and perhaps ought not to be) from among those used to set up the calibration graph. These standards... 

Set up logbooks for the equipment and keep track of performance by plotting appropriate control charts. As a minimum, charts monitoring the high (say 1332.5 keV) and low (say 122.1 keV) energy resolution and the measured activity of a control sample should be maintained. 

In SPC, the most critical part of the process is the validation that you are measuring the right thing and are thereby motivating the correct response. In addition, if one measure can take the place of several measures, then that one measure should be identified, thereby simplifying the measurement process. Once a measurement has been selected, we are ready to set up the data collection process and to establish control charts that will monitor the performance of this data. 

S ystem identification is the term used to define a procedure to characterize the process response. In this case, system identification can be accomplished by setting the default level controller set point at 50 per cent (under Liquid Valve ), adjusting the steam flow to the heater in steps, up and down, and then observing the temperature response on a strip chart. This is termed step response testing and is the same as was done in the previous workshop. 

---