Bridging EPBD Compliance with Real Energy Efficiency Performance

Keywords: Energy Performance of Buildings Directive (EPBD); Measurement and Verification (M&V); Energy savings validation; Building energy performance; HVAC upgrades; Data-driven compliance; Energy efficiency policy

Europe’s ambition for greener buildings is entering a new era: one where real performance matters as much as design expectations. But how do we prove energy savings actually materialize? This article explores how measurement and verification transforms compliance under the EPBD, bridging the gap between theory and reality in building efficiency.

The link between building-level M&V and the EPBD requirements

The 2024 revision of the Energy Performance of Buildings Directive (EPBD) significantly strengthens requirements around transparency, accountability, and measurable progress in building energy performance across the EU. To this end, the EPBD requires the existence of a monitoring, reporting and performance verification framework for evaluating the progress towards the national goals for energy efficiency in buildings.

While the directive emphasizes Energy Performance Certificates (EPCs) as the main means for verification, it leaves room for more granular, evidence-based approaches. Article 9 states that “compliance of individual non-residential buildings … shall be checked on the basis of energy performance certificates or, where appropriate, other available means”. In this context, measurement and verification (M&V) of energy savings at the individual building level can be a highly valuable tool for fulfilling the verification and reporting obligations introduced by the EPBD.

Figure 1. M&V in the framework of EPBD requirements.

When improving the energy efficiency of a commercial building, M&V can help assess whether energy performance improvements were actually achieved, rather than merely predicted. Traditional compliance methods often rely on theoretical performance, which can diverge significantly from real-world outcomes due to occupant behavior, operational changes, or installation quality. By contrast, building-level M&V uses monitored data to quantify actual savings. This offers reliable information to the building owners and aligns closely with the EPBD’s emphasis on operational performance and real energy use.

Furthermore, M&V enhances the credibility and comparability of reported data. The revised directive calls for improved data quality and harmonization across Member States. Verified, building-level performance data can feed into national databases and EU-wide monitoring systems with greater reliability than modeled estimates alone.

Finally, the EPBD revision promotes the use of digital tools to store and update building performance data over time. Integrating M&V processes into these logbooks allows for continuous verification rather than one-off assessments, enabling authorities and building owners to demonstrate sustained compliance.

Current challenges with M&V

Despite of all its benefits, M&V at the individual building faces a series of challenges. In particular:

1.    High-granularity data may not be available before the retrofit. The core idea behind the International Performance Measurement and Verification Protocol (IPMVP) is that energy savings can be estimated by comparing the actual energy consumption after the retrofit to a counterfactual energy consumption that reflects the case where the retrofit had not taken place. The counterfactual energy consumption is calculated from a predictive model that has been trained on pre-retrofit data (baseline). However, it is often the case that pre-retrofit data is not available and, as a result, a baseline of the energy consumption cannot be estimated at all.

2.    Energy savings estimation is fragile with respect to changes in the way a building and its HVAC systems are used. Energy consumption is influenced by weather, occupancy patterns, and operational changes. Even with normalization techniques, it is difficult to isolate the effect of a renovation measure from these external variables. This introduces uncertainty that regulators may find hard to accept compared to standardized calculations for deemed savings.

3.    Cost and complexity. Robust M&V requires metering infrastructure, data collection systems, and expert analysis. For many buildings, this can be disproportionately expensive compared to standardized approaches like EPCs. This makes it harder to scale across entire national building stocks, which is a core goal of the EPBD.

Figure 2. Limitations of M&V as a means for complying with EPBD requirements.

The aforementioned limitations support the argument that M&V should be used as a complementary approach to verification only for upgrades in large, commercial buildings, where metering infrastructure may be installed after the upgrade anyway, and the cost of the data analysis can be justified.

At the same time, these limitations can be re-framed as requirements for M&V methods and tools that can help scale up the use of M&V for program-level and policy-level evaluation of energy performance. These requirements are outlined below:

1.    Deal with lack of pre-retrofit data. There is a need for methods that allow users to substitute data with assumptions about the pre-retrofit state of a building and the expected impact of the upgrades. High granularity data, if made available after the retrofit, should be utilized to verify the extent to which the expectations were met. This requires a shift from modelling the energy consumption dynamics before the upgrade to modelling the dynamics after it.

2.    Separate the impact of the upgrade from changes in the building’s operation. Typical M&V approaches model a building’s energy consumption by blending together its sensitivity to outdoor temperature and recurrent patterns that can be captured by calendar features (hour of day, day of week and week of year). This approach makes it easy to create predictive models, but makes it very difficult to adapt to changes in the building’s operating schedule, the number of occupants at peak occupancy or the setpoint settings of its HVAC systems.

3.    Simplify the data analytics workflow. M&V of energy savings requires expertise in data analysis and energy consumption modelling. Any method or tool for scaling up M&V should implement a full workflow that guides users to carry out the required calculations in a consistent and replicable manner.

The SRI2MARKETapproach to M&V

The EC LIFE project SRI2MARKET (Paving the Way for the Adoption of the SRI into National Regulation and Market) developed an M&V method for HVAC upgrade savings that addresses the limitations described above. Its key innovation is that it stays true to the core principles of the International Performance Measurement and Verification Protocol (IPMVP) while redefining how M&V is applied. Rather than only estimating a counterfactual baseline, it compares equivalent operating conditions before and after a retrofit. By identifying and classifying changing building states over time, the method can adapt continuously to activity shifts, operational changes, and other non-routine events.

At the core of the methodology is a modelling framework that detects relevant system states across time and establishes correspondences between them. Energy savings are then quantified as the difference in consumption between matching states, ensuring that comparisons remain valid even as building behavior evolves. This state-based approach enhances robustness, reduces bias, and improves the reliability of savings estimates in real-world conditions.

In terms of its implementation, the methodology utilizes a pair of models:

·         The first model is a physics-based model that describes the building and its HVAC systems. The model is flexible enough to capture thermal inertia, free-cooling operation and capacity saturation effects.

·         The second model is a predictive model that captures the deviations between the observed energy consumption and the output of the physics-based model, and maps them onto a hidden variable called indoor activity.

The modelling process is carried out for both pre- and post-retrofit periods, and energy savings are estimated by comparing energy consumption at similar outdoor temperature and indoor activity levels. The only assumption that the methodology makes is that there is a period after the retrofit where the distribution of the activity levels with respect to outdoor temperature remains stable. This is an assumption that all quantitative M&V methods implicitly make; the SRI2MARKET approach makes it in a transparent and unambiguous manner[1].

Figure 3. The core concept of the SRI2MARKET M&V approach.

The SRI2MARKET approach addresses the limitations of the previous section because:

1.    It combines physics-based models with predictive ones. The physics-based models are calibrated to the pre-retrofit data but, if pre-retrofit data is not available, their parameters can be filled in given the building’s EPC rating before the retrofit. The predictive models operate on post-retrofit data that is more likely to be available if the retrofit includes monitoring infrastructure and/or BMS installations.

2.    It supports activity shift events. Here, activity shifts are defined as changes in the way an existing system is used or changes in the context of the systems’ operation. To this end, the method allows users to explain energy consumption changes through two different causal paths:

o   Changes attributed directly to the implemented HVAC system improvements and upgrades.

o   Changes in system usage, such as different setpoints, number of heating/cooling hours, or number of occupants during peak occupancy.

3.    It offers a freely-accessible tool that implements a clear and replicable workflow. The tool is accessible at https://mevalapp.com. It requires user registration because it maintains model and estimation histories to support auditability and longitudinal tracking. Nevertheless, the tool will remain free for the foreseeable future as a platform for collaboration and testing of M&V methods that are robust to non-routine events.

A pragmatic way to introduce M&V

Even if M&V is not well suited to replace EPCs as a universal compliance tool under the EPBD, it can still play a strategically important role when applied selectively. Article 23 of the EPBD states that “Member States shall put in place inspection schemes or alternative measures such as digital tools and checklists to certify that the delivered construction and renovation works meet the designed energy performance”. Accordingly, a targeted population sampling approach, focusing on a limited number of carefully chosen buildings, can unlock much of the value of M&V without incurring the full cost and complexity of scaling it across the entire building stock.

Rather than attempting to implement M&V everywhere, regulators can prioritize large commercial or non-residential buildings undergoing major renovations. These buildings are more likely to justify the installation of advanced metering infrastructure as part of the upgrade, making high-quality data collection both feasible and cost-effective. By applying M&V to this targeted subset, authorities can generate a “ground truth” dataset of actual energy savings under real operating conditions. While this data is not directly representative of all buildings, it can be highly informative when combined with segmentation techniques. For example, results from monitored buildings can be extrapolated to similar building types, climates, and usage patterns.

Such an approach aligns well with the EPBD’s broader goals of improving data quality and supporting evidence-based policymaking, without undermining the scalability advantages of EPCs. EPCs can continue to serve as the primary, standardized compliance mechanism, while M&V-based sampling acts as a complementary layer of verification and insight. This hybrid model allows regulators to balance practicality with accuracy: EPCs provide coverage, while M&V provides depth.

Figure 4. Complementarity of EPCs and M&V to fulfill the EPBD requirements.

Conclusion

While the revised EPBD reinforces standardized tools like EPCs for large-scale compliance, building-level M&V offers a powerful complementary pathway grounded in real performance data. Approaches such as those developed in SRI2MARKET demonstrate that M&V can overcome traditional limitations through adaptive, data-driven methods. When applied strategically, particularly in larger, data-rich buildings, M&V can enhance transparency, strengthen trust in reported savings, and support more evidence-based policymaking without compromising scalability.