7.3.1 International Performance Measurement and Verification Protocol

The International Performance Measurement and Verification Protocol (IPMVP) is the single-most recognized M&V protocol in the industry, and is used in more than 40 countries. This protocol sets out a framework to determine energy and water savings achieved compared to baselines defined for new buildings or retrofit projects. It was originally instigated by the U.S. Department of Energy (DOE) and developed by a coalition of international organizations to respond to the calls for a standardized protocol to verify the savings achieved from energy conservation measures. The IPMVP is currently maintained by the Efficiency Valuation Organization (EVO), a nonprofit organization that promotes energy and water efficiency. The EVO also promotes the implementation of M&V by qualified experts that have achieved the status of Certified Measurement and Verification Professional (CMVP) awarded by the Association of Energy Engineers (AEE) (EVO, 2016).

Volume I of the IPMVP defines the principal concepts of M&V and available options for determining energy and water savings. It also provides examples of various applications of M&V (EVO, 2012). Volume II provides the concepts and practices for improved indoor environmental quality (IEQ) in the context of energy efficiency projects (DOE, 2002). There are concerns about unintended consequences of improving energy efficiency and de-carbonization of building stock (Shrubsole et al., 2014). For example, improving the airtightness of a building may increase the risk of overheating in summer. It is important that the operational energy performance targets are not achieved by compromising the occupants’ thermal comfort or the indoor air quality. Volume II of the IPMVP follows a holistic approach to energy efficiency and the IEQ to make sure the pursuit of energy efficiency does not compromise the IEQ performance. Finally, Volume III of the IPMVP provides the concepts and practices for determining energy saving in new constructions (Part I), and in the applications that involve renewable energy technologies (Part II), (EVO, 2006; IPMVP Inc., 2003).

7.3.2 Federal Energy Management Program

The U.S. Federal Energy Management Program (FEMP) promotes energy efficiency and the use of renewable energy resources in federal buildings, which constitute a large component of energy consumption in the United States. The M&V guidelines issued by the U.S. DOE are used for this program (DOE, 2015). These guidelines provide a structured framework to determine and verify energy savings. This M&V framework includes the activities shown in Fig. 7.1. A major focus of the FEMP is performance-based contracts and therefore it takes a long-term view about performance of ECMs and requires regular M&V activities after the initial verification.

The M&V guidelines for FEMP also include guidance and examples for specific ECMs related to building envelope, heating, ventilation and air conditioning (HVAC) systems, and renewable technologies. The FEMP M&V guidelines contain detailed producers for applying the IPMVP concepts to federal buildings. The principal M&V methods used in the FEMP are therefore underpinned by the IPMVP.

7.3.3 ASHRAE (American Society of Heating, Refrigerating, and Air-Conditioning Engineers) Guideline 14

ASHRAE Guideline 14 defines the requirements for instrumentations, data management, measurement methods, and building performance simulations used to determine energy and demand savings (ASHRAE, 2002). It also describes methods to analyze the uncertainties associated with measurements and modeling. The overall uncertainty of the estimated savings must be below the thresholds defined in this guideline. ASHRAE Guideline 14 covers the technical foundations of M&V and does not deal with the wider aspects of M&V such as M&V framework and plans, risks, costs, and third-party verification. There is a significant overlap between the methods described in ASHRAE Guideline 14 and the M&V options defined in the IPMVP. This will be subsequently explained in this chapter.

7.3.4 ISO (International Standards Organization) 50015

ISO 50015 (2014) follows the introduction of ISO 50001 (2011) for energy management systems. ISO 50015 specifically covers measurement and verification of energy performance of organizations. This international standard defines the terms of reference and sets out general principles for M&V in several organizational contexts. It can be used for the M&V of energy performance or energy performance improvement for all or part of an organization. Energy performance improvement in the context of this standard can be achieved through technological, managerial or operational, behavioral, economical, or other changes. The scope of ISO 50015 is thus broader than the abovementioned M&V protocols. It does not provide specific methods for M&V at the building level. Organizations that follow the IPMVP guidelines to verify the savings achieved as a result of implementing ECMs in their buildings can satisfy most of the requirements of this standard. There are private companies accredited to provide certification to this standard.

7.3.5 Superior Energy Performance Protocol

The Superior Energy Performance (SEP) protocol was developed by several members of the U.S. Council for Energy-Efficient Manufacturing (CEEM) in 2012 and is administered by the U.S. DOE (DOE, 2012). It is meant to help organizations that conform to the ISO 50001 standard and have an energy management system in place to verify the savings achieved as a result of implementing this standard. It is currently used in the United States and Canada, and verification bodies accredited by the American National Standards Institute (ANSI) can perform audits on facilities to confirm compliance with the SEP requirements. Although the SEP protocol was primarily developed for industrial facilities, it can also be used for commercial buildings.

7.4.1 Retrofit Isolation: Key Parameter Measurement

This option deals with simple ECMs that can be isolated from the rest of the building and assessed individually. The key performance parameters that define the energy use of the system are measured while a number of other parameters may be estimated. A typical example of a retrofit measure that can be isolated from other systems is replacement of lighting installations with more efficient lights. This can increase the heating demand and reduce cooling loads. These interactive effects that are related to the impact of the ECM on other systems within the building can also be estimated. Estimations can be made by using historical data, manufacturer’s specification, or engineering judgment agreed upon by all parties involved in the contract. The saving error related to estimations must also be evaluated.

7.4.2 Retrofit Isolation: All-Parameter Measurement

In this option, energy performance of the system affected by the ECM is either directly measured or calculated based on measurements of proxies used for energy performance. All determinants of energy use will be measured. Measurements can be carried out on a short-term or continuous basis, depending on the expected variations in savings and the length of the monitoring period allowed in the contract.

7.4.3 Whole Facility

Where the performance of a building is fundamentally affected by the retrofit, as in deep renovation, or where a number of interrelated ECMs are implemented, it is reasonable to take into account the whole building to determine savings. In this option, energy use of the whole facility is measured before implementing the ECM(s) and after the implementation during the monitoring period allowed in the contract. It is important to ensure comparison between baseline and postretrofit performance is carried out under consistent operating conditions. Consequently, simple adjustments or regressions analysis may be required to adjust the baseline performance. A simple example is the changes in weather conditions before and after retrofit. Heating and cooling degree-days may be used to adjust measured energy. However, if the building context and operating conditions are significantly different before and after retrofit it may be difficult to follow this option. Calibrated energy performance simulation offers more flexibility in such circumstances.

7.4.4 Calibrated Simulation

Similar to the whole facility option, this option takes into account the whole building. However, it utilizes building energy performance simulation (BEPS) to determine savings.

The computer model developed for simulation must be calibrated with the actual performance of the building. Calibration is achieved by adjusting the computer model of a building to reflect the as-built status (e.g., as-built fabric U-values, pressure-test results, and commissioning results of HVAC systems) and actual operating conditions (e.g., occupancy pattern, operational schedules of HVAC systems, temperature set points, and actual weather conditions). The outputs of the adjusted computer model are then compared against the measured performance to check if the model can reasonably reflect actual operation of the building. Once calibration is achieved, the effect of implementing the ECM(s) can be calculated by the computer model.

The calibration process is based on hourly or monthly energy data and is determined by the coefficient of variation of the root mean square error (CVRMSE) and normalized mean bias error (NMBE). Table 7.1 provides the calibration criteria used for hourly and monthly calibration of a computer model.

The calibration indices are defined as follows:

$\text{CVRMSE}=100\text{×}\sqrt{\left[{\displaystyle {\sum }_{i=1}^n{\left(y_i-{\stackrel{⌢}{y}}_i\right)}^2}/\left(\text{n}-1\right)\right]}/\overline{y}$ (7.3)

(7.3)

$\text{NMBE}=\frac{{\displaystyle {\sum }_{\text{i}=1}^{\text{n}}(y_i-{\stackrel{⌢}{y}}_i)}}{(\text{n}-1)\text{×}\overline{y}}\text{×}100$ (7.4)

(7.4)

The CVRMSE criterion ensures that hourly or monthly energy errors do not cancel out and are all taken into account in the calibration process. The NMBE criterion, on the other hand, looks for systematic bias in the model. Both criteria must therefore be satisfied for calibration. These calibration indices represent how well a mathematical model describes the variability in measured data.

The calibrated simulation option of the IPMVP is underpinned by ASHRAE Guideline 14. ASHRAE Guideline 14 provides the following equation to link the modeling error to the uncertainty associated with the projected saving:

$\text{U}=t\text{×}\frac{1.26\text{×}\text{CVRMSE}}{\text{F}}\text{×}\sqrt{\frac{\text{N}+\text{2}}{\text{N}\text{×}\text{M}}}$ (7.5)

(7.5)

The t-statistic used in Eq. (7.5) would be generally determined, for a given confidence level, by the number of data points in the baseline period and the number of parameters in the baseline model. ASHRAE Guideline 14 (2002, p. 14) provides a table that could be used to infer the t-statistic. For 68% confidence, t = 1 could be used for all numbers of data points and baseline periods. The uncertainty calculated using Eq. (7.5) must not be greater than 50% of annual projected savings at 68% confidence level (ASHRAE, 2002).

7.4.5 Examples for M&V Options

Table 7.2 includes examples for various M&V options reviewed in this chapter. IPMVP Options A and B are simpler and more cost effective. However, if the significance of some determinants of energy use is not clear, Option A must be avoided. Furthermore, if the interactive effects of ECMs are significant or unmeasurable, Options A and B must be avoided. Generally, as the number, complexity, and interaction of ECMs increase, whole-building options (IPMVP Options C and D) become more relevant.

7.6.1 Energy Monitoring

Most new buildings have advanced metering strategies that allow for disaggregation of total energy use to specific end-uses such as heating, domestic hot water, cooling, auxiliary energy (energy used for fans, pumps, and control), lighting, and equipment. For example, the metering strategies designed and installed for new nondomestic buildings in the United Kingdom should be able to assign at least 90% of the estimated annual energy consumption of each fuel to end-use categories. Metering provisions should also facilitate benchmarking of energy performance; new buildings with total useful floor area greater than 1000 m² should have automatic reading and data-collection facilities (HM Government, 2006). The Leadership in Energy and Environmental Design (LEED) sustainability rating system that is used in North America and internationally also sets out stringent requirements for the metering strategies of new buildings and major renovations that facilitate disaggregation and benchmarking (USGBC, 2013). However, existing buildings that are subject to retrofit measures do not necessarily have advanced metering strategies. Total energy use of each fuel is often available through utility bills. However, it is sometimes difficult to get reliable data for end-use categories. This is a major challenge for energy benchmarking and identification of problem areas. Furthermore, verification of savings would be difficult in the absence of a detailed metering strategy. Information about energy end-uses is also necessary to calibrate the computer models, where the calibrated simulation method is pursued for M&V.

It is sometimes reasonable to install submeters to measure the energy use of key determinants of energy performance (e.g., the chiller’s energy supply in an air-conditioned office building). Clamp meters or portable electricity profilers, used at the distribution board for short period of time during typical operation of the building in accordance with relevant health and safety regulations, can provide useful insights about distribution of energy. More innovative approaches are being explored for energy data disaggregation. Fig. 7.2 shows the result of a disaggregation algorithm that is based on detailed monitoring of mains electrical supply (single-point measurement) to figure out and separate key energy loads based on a learning process. The learning process takes into account the effects of the loads on the mains supply and the timing and frequency of the effect. This approach can significantly reduce the metering requirements while providing useful information about energy use of the key loads within a building.

7.6.2 Monitoring of the Indoor Environmental Quality

Monitoring of the indoor environmental quality (IEQ) is important to ensure that improvement in energy performance has not been achieved at the expense of the IEQ. Trend logging of the IEQ parameters is also necessary to account for changes in operating conditions when the postretrofit performance is compared with the baseline to verify savings. Monitoring of indoor temperatures and relative humidity can establish thermal comfort conditions. Carbon dioxide concentrations are often used as proxy for the indoor air quality. If the occupancy details are known, CO₂ concentrations may also be used to infer the ventilation rates (CIBSE, 2005). Illuminance levels can be recorded and used to assess the performance of the lighting system. The operation of electrical lights can also provide information about occupancy patterns. The IEQ data can be used to calibrate the computer models where the calibrated simulation method is used for a project. The data can be collated via BMS or separate sensors installed for M&V. Fig. 7.3 shows an example of a wireless data-logging and alarm generation system based on ultrahigh frequencies (UHF) that was used by the authors in previous research (Burman et al., 2014). The transmitters are located at the monitoring points while the repeater and the receiver/data logger can be far from the monitoring points. A central receiver can record the data received from multiple transmitters. This data is remotely accessible via a GSM modem integrated into the receiver. The alarm generation system will let users know if there is any problem with the transmitters and data communication. Only the central receiver needs to be plugged into a power socket. It can be located in the facility manager’s office or the central plantroom. This system is not reliant on the local WiFi network and can provide a self-sufficient solution for wireless IEQ monitoring.

There is also a recognition among software developers for BEPS that it is necessary to facilitate transfer of data from monitoring platforms to BEPS platforms. For example, the IES software has recently developed a new module to link together various sources of operational data (from smart meters, BMS, and environmental sensors) and feed them back to the design models. This information can be used for model calibration and building diagnostics (IES, 2012). Another example is Apidae, a cloud-based platform to perform EnergyPlus simulations. It provides a calibrator module that facilitates transfer of operational data to BEPS. The calibrator module also uses the CVRMSE between the measured data and the simulation outputs as a fitness function, and its algorithms attempt to modify the parameter setting to minimize the CVRMSE (Apidae Labs, 2016).

7.6.3 Occupancy Monitoring

Investigations into the effects of occupant behavior on building energy use have found that behavioral parameters significantly influence energy use in both domestic and nondomestic sectors and vary according to the building type, size, and climate (Sunikka-Blank and Galvin, 2012; Azar and Menassa, 2012; Dasgupta et al., 2012). Longitudinal changes in behavioral parameters are also important in the context of retrofit projects and adjustments made to M&V calculations to determine savings. Occupant behavior can change the IEQ settings that can be captured by the sensory equipment. Determination of the number of occupants, occupancy pattern, and the small power used by occupants is also a major challenge. Depending on the context, proxies such as carbon dioxide concentrations and illuminance levels might be used to infer information about building occupancy. Records of WiFi connections in office-type buildings can also be used to infer occupancy (Martani et al., 2012). Fusion of data from several simple and cost-effective methods can lead to a better understanding of occupant behavior and energy use. For example, the following methods were tested in an office building to collate information about occupancy (Lam et al., 2014):

This information along with other energy and IEQ data enabled computer model calibration with NMBE of 1.27% and CVRMSE of 6.01%, significantly lower than the IPMVP limits (Table 7.1).

This level of monitoring may not be always practical for the whole building, especially in large nondomestic buildings. Nonetheless, these methods could be used in sample zones that are representatively chosen to provide feedback for model calibration and M&V.