Directive 2006/32/EC of the European Parliament and of the Council (repealed)Show full title

3. How to deal with uncertainty

All the methods listed in point 2 may entail some degree of uncertainty. Uncertainty may derive from(1):

(a)

instrumentation errors: these typically occur because of errors in specifications given by the product manufacturer;

(b)

modelling errors: these typically refer to errors in the model used to estimate parameters for the data collected;

(c)

sampling errors: these typically refer to errors resulting from the fact that a sample of units was observed rather than the entire set of units under study.

Uncertainty may also derive from planned and unplanned assumptions; these are typically associated with estimates, stipulations and/or the use of engineering data. The occurrence of errors is also related to the chosen system of data collection that is outlined in points 2.1 and 2.2. A further specification of uncertainty is advised.

Member States may choose to use the method of quantified uncertainty when reporting on the targets set out in this Directive. Quantified uncertainty shall then be expressed in a statistically meaningful way, declaring both accuracy and confidence level. For example, ‘the quantifiable error is found with 90 % confidence to be ± 20 %’.

If the method of quantified uncertainty is used, Member States are also to take into account that the acceptable level of uncertainty required in energy savings calculations is a function of the level of savings and the cost-effectiveness of decreasing uncertainty.

(1)

A model for establishing a level of quantifiable uncertainty based on these three errors is given in Appendix B in the International Performance Measurement & Verification Protocol (IPMVP).