Przedziałowa postać wyniku pomiaru jako podstawa wyrażania niedokładności w systemach pomiarowo-sterujących

W artykule opisano koncepcję przedziałowego wyrażania wyniku pomiaru oraz jego niepewności w sposób specyficzny dla systemów pomiarowo-sterujących. Niepewność rozumiana jest tu jako parametr błędu wyniku pomiaru interpretowanego w kategoriach probabilistycznych. Wyrażanie wyniku w postaci przedziału przedstawiono na przykładach obliczanych symulacyjnie przy użyciu metody Monte Carlo.

---

An approach to the interval representation of measurement result and its uncertainty in measuring and control system is presented in the paper. Nowadays, the measurement result is characterized by the measurement uncertainty [1], which is defined as the radius of the interval built around the measured value in which the true value lies with given probability. A rapid growth of measurement systems application area leads to introduce more usable definition of inaccuracy which basis on the interval representation of a measurement result. This definition is more useful, particularly in real-time systems and when errors with asymmetrical distribution occur in systems [2, 4]. According to classical definition, to classify a system as real-time one delays in it have to be less than it is allowable. In such systems propagation of the signals from the input to the output is connected with arising of delays because all the system elements need time to perform their activities. However, to classify a measuring and control system as a real time, it should be taken into account all factors influencing on properties the system output signals, i.e. not only delays but also errors of measurement data. Therefore, the delay errors should be described as components of the total error being the basis of determination of the interval representing the system output measuring results [5]. Comparing the interval with critical acceptable values enables classifying the system as a real-time one. Theoretical consideration in the paper are illustrated by results of numerical experiments carried out by using Monte Carlo method.