Abstract
The aim of this case study was to determine how suitable time series are for assessing the stability over time of analytical systems based on sensor arrays. We show that control charts are unsuitable because of the autocorrelation in the data. The methodology we used was the following: first, we identified the model behind the data and then we plotted the prediction residuals in a control chart. We also evaluated the time constant for each series to determine whether the sampling frequency was correct. The data used were provided by a set of potentiometric sensors for barium, ammonium, magnesium, potassium and water hardness. The results enabled us to establish the correct sampling frequency for each of the sensors and to detect when a sensor is going out-of-control.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
