Using Uncertainty Modeling to Optimize MV in ESPC

Using Uncertainty Modeling to Optimize M&V in ESPC Projects Discussion Document 10/31/02

Overview n Context: Concern about level of M&V n n n Difficult to reach consensus on a specific project General concern about the degree of stipulation and utility bill reduction Can we add value to the current M&V decisionmaking process (largely qualitative) by incorporating uncertainty modeling (more quantitative)? n Can we do this without complex (expensive) analysis?

Role of M&V in ESPC n n Maintain the integrity of the ESPC program Identify risk parameters and assign responsibilities n n n Generally, ESCOs responsible for equipment “performance” and agency responsible for “operational” factors. Define required measurement parameters and protocol Future role: Help correlate interest rates with energy savings uncertainty

M&V Decision-making in ESPCs – Current Status n Available Tools n n n FEMP M&V Guide – based on IPMVP Risk/Responsibility matrix M&V decision support flowchart ASHRAE Guideline 14 Limitations – Lack of “context-specific” guidance n n What is the risk tolerance of ESCO and agency? What is the “optimal” (not ideal) level of analysis, given data, time, and budget constraints for a given project?

M&V Decision-making in ESPCs – Proposed Improvement n Apply quantitative uncertainty analysis to optimize M&V, considering: n n Savings uncertainty Cost of reducing uncertainty via M&V Risk tolerances of the agency and ESCO This allows agency and ESCO to make a more informed decision n Tradeoff M&V value (uncertainty reduction) and M&V cost

This is not new! n FEMP M&V Guidelines already recommend use of uncertainty analysis n n n FEMP Method A Detailed Guidelines n n n Section 2. 5: Selection of M&V methods & rigor Appendix D: Sampling guidelines Section 3. 1: Considering uncertainty Section 3. 2: How stipulations apportion risks and responsibilities We also don’t need to invent new science! n Apply well-established risk management techniques for FEMP context

ESPC Cash-flow Model n Payments to ESCO <= Total cost savings Total Cash flow during ESPC is (supposed to be) less than before ESPC Savings Agency Cash flow ($) n ESCO Payment Energy + O&M Before ESPC Energy + O&M During ESPC Total Energy + O&M Cost Savings

ESPC Cash-flow Model In reality, there are uncertainties around each cash flow component weather, energy prices, etc. Agency Cash flow ($) n + Savings = Energy + O&M Equip performance, model assumptions, etc. Before ESPC + During ESPC ESCO Payment Typically fixed = ESCO Payment Energy + O&M

Cash-flow Uncertainty vs. Risk Tolerance Expected Cash-flow Uncertainty Agency Risk Tolerance Risk Management Need $/yr Agency Cash-flow

M&V and Risk Management n Optimize M&V based on risk management needs n M&V is not an end in itself! Expected Cash-flow Uncertainty w/ no risk mgmt $0 for M&V Agency Risk Tolerance Uncertainty w/ some risk mgmt $X for M&V Uncertainty w/ more risk mgmt $2 X for M&V $/yr Agency Cash-flow

M&V and Risk Management – Illustrative Example Project Capital Uncertainty M&V – Risk Management Lighting Retrofit Medium Low Agency mostly stipulates savings and absorbs risk Controls/ Commission Low High Since capital is low, agency mostly stipulates, in lieu of increased M&V costs Chiller retrofit High Medium Agency requires increased M&V due to savings uncertainty and high capital

Proposed LBNL Work (FY 2003) n Pilot Projects n n n Assess lessons learned - policy implications for FEMP n n Select 2 -3 ESPC projects as pilots before M&V decisions are made Analyze uncertainty (work with ESCO, PF) Assess agency risk tolerance (discussions with agency) Analyze costs/benefits of different M&V options Record final decision-making process Risk assessment/management requirements Incorporation of risk management items in contracts Need, if any, for guidelines and tools Leverage FEMP M&V Committee for peer-review, guidance and dissemination