The cost of measurement uncertainty

No measurement is free from uncertainty. Uncertainty is defined as the deviation from the true value within a stated probability. The paper looks at flow measurement uncertainty costs in relation to oil and gas applications. Examples are given of the sources of uncertainty and their impact on the overall measurement. What is not commonly appreciated is the extent to which largely untreated influences can have on the overall measurement. A good example here is the recommended upstream straight length ahead of the flowmeter in order to stabilize the flow for metering purposes. Typical distances for turbine meters range from 10 up to 20D and for some other types this can be more. However this 10-20D is given independent of pipe size. Fluid dynamically, it is not correct to state that 10D is required for a DN50 pipe and that 10D is also required in a DN500 pipe. Here the existing standards do require revision. Experiments have shown that the straight upstream length should increase as pipe diameter increases. The paper treats uncertainty in cost terms and concludes by comparing the cost of the measurement uncertainty to the cost of the installation itself. Bids always try to minimize the purchase cost but this is only a small fraction of the overall cost of measurement. Even the stability of dedicated provers can impact the overall measurement uncertainty and the Author presents data quantifying this largely neglected cost. Recent audits in several refineries have confirmed the correctness of the proposed methodology. It is thought that most if not all refineries and chemical plants can reduce costs of metering by simply working to better and tighter guidelines.