Estonia’s Renokratt prototype shows how knowledge-based expert systems can bridge fragmented building data, helping non-experts plan deep renovations and supporting scalable, EPBD-aligned decision-making through typologies, rules, and structured measures databases.
Joosep Viik, Project Expert, Tallinn University of Technology | LinkedIn profile, ETIS profile
Ergo Pikas, Tallinn University of Technology | LinkedIn profile, ETIS profile
(Note: Opinions in the articles are of the authors only and do not necessarily reflect the opinion of the European Union)
The revised Energy Performance of Buildings Directive (EPBD) sets a clear direction: Europe must accelerate deep renovation, improve building performance transparency, account for whole-life impacts and enable building owners to move towards zero-emission buildings. Yet across Member States delivering on these goals depends on one enabling factor: high-quality, interoperable and continuously updated building data. This data need spans multiple layers. Member States must be able to connect building registries and digital building logbooks with energy performance certificate (EPC) workflows, technical system data (heating, ventilation, controls), smart readiness information, and where relevant life cycle and material datasets.
The renovation context makes the data challenge particularly visible. Renovation decisions are typically made under uncertainty, by non-professional building owners, and often over many years. To make renovation passports and EPCs actionable, the underlying datasets must support not only technical calculations but also clear sequencing of measures, realistic cost expectations, and understandable performance outcomes.
The data needed to plan and scale renovation is often incomplete, inconsistent, or inaccessible for real-world decision-making.
The problem is not the lack of technical renovation solutions. The real bottleneck is that expert knowledge, reliable building data, and decision support tools are not yet integrated into a scalable workflow that non-expert renovation initiators can use independently.
A promising pathway to bridge this gap is emerging through knowledge-based expert systems, which combine building data, typologies, renovation measures databases, and rule-based logic into an interactive renovation planning environment. This article introduces the concept of a knowledge-based configuration expert system (KBCES) for renovation planning, based on which the Estonian prototype Renokratt was developed, and explains why such tools are becoming essential for full EPBD implementation. The same challenge, making expert knowledge understandable and actionable for homeowners, is addressed through a renovation guide that consolidates results from the LIFE IP BuildEST and LIFE heritageHOME projects.
Deep renovation is a complex multi-stakeholder process. In many cases, the renovation initiator is not a construction professional, but rather a private homeowner, an apartment association representative, or a building manager. Research shows that these renovation initiators often struggle with basic early-stage questions and may not have the expertise to define renovation requirements or communicate them clearly to designers and contractors.
In Estonia, this challenge is especially visible in the renovation of Soviet-era apartment buildings, where scaling renovation rates is necessary to meet long-term climate targets. The renovation process is hindered by barriers such as limited awareness, the undervaluation of design, and difficulties in defining renovation goals and design tasks.
At the same time, training every non-expert into a renovation specialist is unrealistic. What is needed instead is a system that embeds expert knowledge into an accessible digital workflow, supporting users ‘independently and at their convenience’.
Traditional renovation planning relies heavily on expert-led audits and manual development and evaluation of different renovation scenarios to make informed decisions. While valuable, this approach faces major limitations when scaled to the national level due to limited expert capacity, inconsistent quality of outputs, high time cost, and weak reuse of previous renovation knowledge.
Digitalisation can improve renovation efficiency by enabling simulation, comparison, and optimisation of renovation scenarios but only if reliable data and structured knowledge are available as input. This is where the EPBD data gap becomes tangible: public building registries often contain baseline information, but not the detailed, validated, and renovation-relevant dataset required for confident decision-making.
A KBCES is a hybrid concept that combines [1]:
In renovation planning, this means the system can:
This turns renovation planning into a structured, replicable workflow—rather than a one-off expert exercise.
The Renokratt prototype was developed for typical Estonian Soviet-era apartment buildings using a Design Science Research methodology [2]. Its objective was to support non-experts in establishing and configuring renovation scenarios, allowing early-stage assessments to be carried out to understand their impacts and produce clear renovation requirements for design tenders.
The prototype guides users through six stages:
This workflow directly responds to the EPBD implementation challenge: making renovation planning accessible and consistent at scale.
A key finding from the prototype evaluation was clear: the system's usefulness depends heavily on reliable data. In Estonia, baseline building data can be retrieved from open sources such as Address Data System (ADS), Topographic Database (ETD), Building Registry (EBR). However, open registry data can be incomplete or, in fields where data entry responsibilities are left to non-experts, even incorrect.
This creates a direct policy implication for the implementation of the EPBD. Better data structures and improved building data quality are prerequisites for scaling digital renovation. In practice, this means the ‘data gap’ is not only a missing dataset problem, it is also a governance problem: who maintains data, who validates it, and how it becomes renovation-ready.
One powerful method for overcoming missing data is building typology enrichment. Renokratt complements open registry data with typology-based statistical parameters, for example, by assigning typical façade window proportions or thermal properties based on archetype classification. This allows the system to function even when detailed measured data are not available, while still producing reasonable early-stage estimates.
But typology alone is not enough. The system also requires a structured renovation measures database that is technically feasible, sufficiently comprehensive, and continuously updated. To be used for the described expert system, it also must be easy to understand for non-experts and machine-readable for automation. To provide meaningful guidance, a renovation expert system must understand not just measures, but also relationships and constraints between them. This is crucial for the implementation of the EPBD because it is not about ‘single measures’ it is about coherent renovation pathways that avoid technical lock-ins and unintended consequences.
Renokratt formalises renovation logic through configuration rules, such as:
The system therefore becomes more than a calculator: it becomes an expert reasoning environment, capable of filtering choices and guiding users to technically valid configurations.
The KBCES [3-4] concept is suggested as a response to the challenges in initiating renovation projects. It is a synthesis of expert systems, knowledge-based systems and configuration systems. Its architecture is based on the key components of a knowledge-based system: A) a user interface, B) a knowledge base and C) an inference engine (Figure 1).
Figure 1. System architecture of the KBCES for deep renovation planning [3-4].
A. User interface (Figure 2) must guide non-professional users through the system, while providing sufficient information to make informed renovation decisions.
Figure 2. User interface for the Renokratt. Source: Renokratt.
B. Knowledge base is the core element that enables the system to provide expert level guidance in a scalable and repeatable way. It stores the structured renovation knowledge required to transform raw building inputs into technically valid renovation pathways. In practice, this includes a combination of (i) building archetype and typology information, (ii) renovation measures and solution packages, and (iii) formalised rules and constraints that capture dependencies between measures and building systems. The knowledge base therefore goes beyond a static catalogue of options: it represents renovation logic in a machine-readable form, enabling the system to identify feasible configurations, exclude incompatible choices, and propose coherent renovation strategies. For example, it can link envelope upgrades and airtightness improvements to ventilation requirements or connect heating system upgrades to expected energy performance outcomes. To support early-stage planning, the knowledge base can also include default values and statistical parameters derived from typical building characteristics (e.g., geometry assumptions, window-to-wall ratios, thermal performance of typical walls), allowing the system to generate credible scenarios even when registry data is incomplete. By structuring renovation expertise into reusable components, the knowledge base makes it possible to deliver consistent advice to non-expert users, while providing a foundation for continuous updates as regulations, technologies, and cost assumptions evolve.
C. Inference engine (Figure 3) includes an interpretation module (connects inputs to knowledge), an explanation module (communicates pros/cons), and a solving module (runs assessments and generates results). The latter includes functional blocks such as visualisation, compliance checking, energy performance assessment, greenhouse gas calculation, cost estimation, and results generation.
Figure 3. Configuration model rule R3 Unified Modeling Language (UML) schema example. [3-4].
The EPBD sets ambitious goals but reaching them depends on what happens between policy and construction sites. That space is filled with uncertainty, fragmented information, and limited expert capacity. To fully implement the EPBD, Member States need more than compliance frameworks. They need digital capacity to deliver reliable building performance data and traceability that can produce scalable pathways for millions of buildings, and decision support that also works for non-experts.
One of the most important evaluation insights was that for apartment associations, the primary decision-making driver is financial impact, especially expected changes in monthly expenses for residents. This is a critical reminder for the implementation of the EPBD: even if policy targets focus on energy and emissions, renovation decisions often depend on affordability and clarity. Therefore, bridging the data gap is not only about energy performance modelling it is about making costs transparent, and scenarios comparable and impacts understandable.
Knowledge-based renovation expert systems offer a practical bridge by connecting public open data + typology enrichment + renovation measures databases, rule-based reasoning + automated indicators into one end-user workflow.
If we want to bridge the data gap for the implementation of the EPBD for building renovation, the priorities are clear:
To accelerate deep renovation and succeed in walking the path set by the EPBD, Member States must close the gap between building data and actionable renovation decisions. This gap is most visible in the initiation phase, where non-expert owners and apartment associations need clear, comparable pathways under uncertainty.
Renokratt shows how a KBCES can bridge this gap by combining public registry data, typology enrichment, a renovation measures library, and rule-based logic into a workflow that produces comparable scenarios that translate directly into design-tender-ready requirements. Scaling this approach requires improved data quality and continuously maintained, machine-readable renovation knowledge.
[1] Liao, S.-H.: Expert system methodologies and applications—a decade review from 1995 to 2004. Expert Syst. Appl. 28(1), 93–103 (2005). https://doi.org/10.1016/j.eswa.2004.08.003; R. Akerkar and P. Sajja, Knowledge-Based Systems. Jones & Bartlett Learning, 2010; Felfernig, A., Hotz, L., Bagley, C., Tiihonen, J.: Knowledge-based configuration: from research to business cases. Newnes (2014)
[2] Peffers, K., Tuunanen, T., Rothenberger, M., Chatterjee, S.: A design science research methodology for information systems research. J. Manag. Inf. Syst. 24, 45–77 (2007)
[3] Viik, J., Knowledge-based configuration expert system for renovating Estonian apartment buildings, Tallinn University of Technology, Master thesis (2024)
[4] Viik, J., Pikas, E., Kalamees, T. (2026). Knowledge-Based Configuration Expert System for Deep Renovation Planning of Buildings. In: Jurelionis, A., Fokaides, P.A., Mazzarella, L., Hartmann, T. (eds) Building Digital Twins. BDTIC 2025. Lecture Notes in Civil Engineering, vol 775. Springer, Cham. https://doi.org/10.1007/978-3-032-09040-9_13
