Cost & Rigor in Savings Flashcards
Options A and B costings
Option A methods usually have lower costs and higher uncertainty than Option B methods. Since new measurement equipment is often involved in Options A or B, the cost of installing and maintaining this equipment may make Option C less costly for longer reporting periods, but this must be compared to the costs for tracking static factors, managing data, and making non-routine adjustments. Cost planning for Options A and B should consider all elements: analysis, meter installation and calibration, the ongoing costs to read and record data, and to perform verification activities. When multiple EEMs are installed at one site, it may be less costly to use Options C or D than to isolate and measure multiple EEMs with Options A or B.
Option C costings
Option C’s cost depends on the source of the energy data and the difficulty of tracking static factors impacting energy use within the measurement boundary to enable non-routine adjustments during the reporting period. Utility meters work well, or existing sub-meters if calibrated, and the data are properly recorded. This choice requires no extra metering cost. However, the cost of tracking changes in static factors depends on the facility’s size, the likelihood of static-factor changes, the difficulty of detecting changes on-site, the availability of frequent utility data (i.e., hourly or daily), and the surveillance procedures already in place.
Option D costings
An Option D simulation model is often time-consuming and costly because it requires the development of an often complex building energy simulation that considers all parts of the building envelope, systems, loads, schedules, or industrial process. Energy modeling and calibration efforts can take a substantial amount of specialized engineering time and will require updates to account for changes in static factors during the reporting period.
Managing Uncertainty
The determination of savings is uncertain because savings represents the absence of energy use and cannot be directly measured. In an M&V process, uncertainties in parameters used to determine savings prevent the exact determination of savings. We use the term error to compare a measurement or forecast when the true value is known, whereas the term uncertainty is used when the true value is not known.
Our goal is to minimize uncertainties in the determination of savings. Minimizing uncertainty can be done by minimizing the uncertainty of individual parameters in a savings determination process and conducting a thorough savings uncertainty analysis for the resulting savings.
Characteristics of a savings determination process that should be carefully evaluated to manage uncertainty in reported savings are:
» Instrumentation error – The measurement of any physical quantity includes errors because no measurement instrument is 100% accurate. Measurement equipment errors are due to calibration, the accuracy of the instrument, inexact measurement, and improper meter selection or operation. Verifying instrument specifications and accuracy relative to a calibrated instrument is typical.
» Modeling error – Mathematical forms from regression analysis or other techniques do not fully account for all variations in energy consumption or demand. Limited modeling error (uncertainty due to scattering in the data beyond what is characterized by appropriate independent variables) is expected and allowable within appropriate bounds. High levels of modeling error can be due to unusual variations in data, inappropriate functional form, the inclusion of irrelevant variables, or the exclusion of relevant variables. Evaluating modeling error using statistical parameters to ensure model validity is required and quantifying uncertainty is possible.
» Sampling error – Use of a sample of the full population of items or events to represent the entire population introduces error as a result of the variation in values within the population, or from biased sampling. Sampling4 may be performed in either a physical sense (i.e., only x number of the lighting fixtures are measured) or a temporal sense (instantaneous measurement only once per hour). Quantifying the actual precision of statistical sampling strategies to evaluate the viability of the sample is typical.
» Estimated values – Error introduced from using any non-measured parameters in savings computation method. Evaluating the potential impact on savings from estimated values using the range of the expected values is typical.
» Interactive effects – Energy impacts beyond the measurement boundary of the EEMs that are not fully included in the savings computation methodology. Any estimated interactive effects should be small compared to overall savings and conservatively estimated to limit impacts on reported savings.
» Data collection and analysis – The inherent errors arising from the (e.g., erroneous or missing data) must also be managed when developing and implementing the M&V Plan. Implementing rigorous quality assurance procedures can reduce these errors.
Error in the M&V Plan
Methods of quantifying, evaluating, and reducing some of these sources of errors and uncertainties should be used in the development of the M&V Plan to reduce the uncertainty associated with determining savings. Assessing optional M&V program characteristics can establish the level of rigor used and support stakeholder confidence in the reported savings.
Balancing rigor and costs example
Consider a project where stakeholders agree savings uncertainty will be determined, with an expected savings of $100,000 per year and $5,000 per year cost for an Option C M&V approach based on one independent variable (outside air temperature). At 90% confidence, an uncertainty of no better than ±
$25,000 (± 25% of savings) per year is possible. The expected uncertainty in reported savings can be reduced to ± $7,000 (± 7% of savings) by including an additional variable, but additional site visits would be required to collect the data. In this case, it may be seen as reasonable to increase M&V expenditures up to $10,000 per year (10% of savings), but not $20,000 per year (20% of savings).
