Skip to content

Chapter 5 — Measurement Uncertainty & Data Acquisition Systems

5.1 All measurements are uncertain

No reading is exact. Understanding and managing measurement uncertainty is what separates trustworthy monitoring from deceptive precision.

5.2 Key concepts

  • Accuracy — closeness to the true value.
  • Precision / repeatability — closeness of repeated readings to each other.
  • Hysteresis — output depends on the direction of change (loading vs. unloading).
  • Noise — random fluctuation masking the signal.
  • Error budget — the sum of individual uncertainties (sensor, cable, readout, environment) that bounds the total uncertainty.

A small, precise but inaccurate instrument can be more misleading than a larger, honest one.

5.3 Transducers: how instruments "speak"

Most modern instruments are transducers converting a physical quantity into an electrical or pneumatic signal. Common types:

  • Vibrating-wire — frequency output, robust, widely used.
  • Electrical (strain-gage, resistance, capacitance) — versatile, sensitive to lead effects.
  • Pneumatic — gas-pressure balance, no power needed.
  • Hydraulic — liquid-column balance, long-established.

5.4 Data acquisition systems (ADAS)

A data acquisition system scans sensors, applies conversion, time-stamps, and stores readings. For automated monitoring it includes telemetry and a database. Selection considers channel count, scan rate, power, environmental hardening, and data formats.

5.5 Managing uncertainty in practice

  • Choose instruments whose accuracy matches the decision, not maximal spec.
  • Establish and document the error budget.
  • Calibrate and verify (Chapter 11).
  • Report readings with their uncertainty, not as false-precision decimals.

Precision is not accuracy

A readout showing 12.3456 kPa is not more true than 12.3 kPa. Report to the resolution the instrument and installation actually support.

5.6 Key takeaways

  • Every reading carries uncertainty; manage the error budget.
  • Accuracy ≠ precision; both matter, neither alone suffices.
  • Match instrument accuracy to the decision, not to the brochure.
  • Automated systems need robust ADAS, telemetry, and data handling.