Meaning
A statistical validation protocol establishes the degree of equivalence between two different measurement systems used in quality control. When applied to polymer extrusion, bland altman analysis compares measurements from an inline sensor against manual laboratory micrometres to verify wall thickness consistency. This analysis determines whether a newer, automated instrument can replace an established manual method without introducing systematic errors.
The technique calculates the mean difference between the two methods and defines limits within which most differences are expected to fall.
Systematic Deviation
Processing variables in injection moulding often suffer from sensor drift that introduces hidden errors into the production log. The presence of a fixed or proportional bias between an inline melt temperature sensor and a nozzle thermocouple is revealed through the average difference line. When this value deviates from zero, the moulder must adjust the machine calibration to prevent thermal degradation of the virgin resin.
A stable bias allows for simple mathematical correction, whereas a variable bias requires deeper investigation of the sensor placement.
Error Distribution
Scatter plots of the differences against the averages reveal whether the disagreement between the measurements remains uniform across the entire operating range of the moulding machine. For instance, high melt temperatures might exhibit much greater variance than lower temperatures, which signals that the secondary measurement method loses reliability at the outer limits of the process window. Identifying these non-uniform patterns prevents the incorrect rejection of compliant batches during long runs.
It also highlights the specific processing zones where manual verification must remain the primary method, ensuring that regrind ratios do not inadvertently alter the viscosity measurements.
Moulding Application
Implementing this statistical tool during tool commissioning protects the moulder from making incorrect decisions about cavity dimensions based on faulty measurement data. The analysis defines the operational boundaries of automated inspection systems before they are integrated into high-speed production lines. Maintaining consistent measurement equivalence ensures that both the supplier and the buyer use compatible quality metrics.