A smart building HVAC control competition crowdsourced and compared algorithms on fair and equal ground using the standardized BOPTEST framework. The competition attracted 138 participants, but only 9% submitted valid solutions for the final stage, highlighting the complexity of advanced HVAC control design. The winning solutions showed significant potential to reduce energy use and cost by shifting demand, without compromising occupant comfort. Across scenarios, thermal energy cost reductions of 36–76% relative to a baseline, were achieved. In peak heat periods, the cost reduction leveraged limited energy use reduction (0–15%), but more significant energy price reduction (34–62%). This shows smart controls' ability to avoid as much as possible consumption during the morning peak hours, when spot prices are tendentially the highest. Hosting the competition has highlighted challenges in creating competitions that both are fair and promotes solutions that are transferable to real life implementation.

This positional paper summarizes a special issue, named ‘Building Sustainability and Performance Through Simulation’. Modelling and simulation approaches are increasingly utilized to investigate methods for enhancing building sustainability and performance. Although existing research has made significant progress in evaluating building performance and providing suggestions to improve building sustainability at various geographic scales, numerous research questions and gaps remain that need to be addressed. To answer these questions and address these gaps, IBPSA-USA organized the SimBuild 2024 conference. The scientific committee, which includes the authors of this positional paper, selected top candidates from the authors who contributed to the conference papers to extend their work into journal articles for this special issue. Thirteen papers appear in this special issue, addressing critical research questions and urgent research gaps that support our research community in continuing to explore building sustainability and performance using simulation approaches.

This paper proposes a support vector machine (SVR) thermal comfort prediction model based on genetic algorithm (GA), gray wolf optimization algorithm (GWO) and sparrow search algorithm (SSA), and compares and analyzes the effects of different optimization algorithms on SVR performance. A data set was constructed by measuring six key parameters of a shopping mall, namely air temperature, relative humidity, air velocity, mean radiant temperature, human metabolic rate and clothing thermal resistance, to evaluate the performance of different optimization methods in terms of prediction ermal comfort prediction and provides theoretical support for optimizing building performance simulation and improving indoor thermal comfort indicators.

This study presents a refined method for constructing Typical Meteorological Year (TMY) weather files in high-latitude regions, using Norway as a case study. We generate municipality-specific TMYs based on the CERRA, a high-resolution (5.5 km) dataset well-suited for Norway’s complex topography and varied climate. Standard TMY methods apply uniform weighting across all months, with heavy emphasis on solar radiation. However, this approach is suboptimal for high latitudes, where winter solar gains are minimal, and temperature is the primary driver of building energy demand. We introduce a modified approach, TMY-NO, which retains conventional multi-variable weighting during summer but shifts to temperature-dominant weighting in winter. This seasonal adjustment improves the thermal representativeness of the TMYs in heating-dominated contexts. Validation against various datasets confirm improved winter temperature accuracy without substantial loss in solar fidelity. A complete set of 358 municipality-level TMYs was produced and adopted as an annex to the Norwegian Standard NS3031.

In performance simulations of residential buildings, manual light-switching behaviour is typically represented through deterministic rules or fixed schedules. While stochastic methods have been developed to replicate human control behaviour for non-residential environments, such methods are currently lacking for residential settings, despite the increased complexity and diversity of occupant actions. This study seeks to address this by introducing RELICS: REsidential Lighting Interaction and Control Simulation. The first step of the model involves characterizing behavioural variability at the household and room-level. Subsequently, manual control is stochastically generated, utilizing assigned routines and predefined probability functions. The efficacy of RELICS is assessed through a case study analysis. The simulation outcomes were compared with monitored data, revealing that the model accurately captures essential switching behaviour patterns. While conventional deterministic approaches may suffice for annual energy estimates, the probabilistic methods employed by RELICS offer a more realistic representation of spatiotemporal dynamics in artificial lighting control.

Heat pump use for heating and cooling in buildings is increasing significantly due to renewable energy integration. While cascade optimization has been extensively studied for cooling applications achieving up to 17% energy savings, limited research addresses this approach for heat pump heating systems despite similar performance characteristics. This work presents an analogous optimization strategy using heuristic mathematical optimization approach is used to calculate optimal threshold values for heat pump cascade. A simulation of three heat pumps with the optimized cascade reduces electricity consumption by 13.80% compared to a baseline configuration. Economic assessment is conducted at the design phase to evaluate the capital expenditure and operational expenditure of different heat pump capacity configurations meeting the same nominal demand. While single high-capacity heat pumps minimize investment costs, a multi-unit configuration offers superior long-term savings over five years. The proposed optimization further accelerates cost recovery. Abbreviations: AI: Artificial intelligence; BEM: Building energy model; BIM: Building information model; CAPEX: Capital expenditure; CAPFT: Capacity modifier curve with temperature; CO2: Carbon dioxide; COP: Coefficients of performance; DMS: Demand side management; EER: Energy efficiency ratio; EIR: Energy input ratio; EIRFPLR: Energy input ratio with part load ratio; EIRFT: Energy input ratio with temperature; HVAC: Heating ventilation and air conditioning; OPEX: Operational expenditure; PLR: Part load ratio; RES: Renewable energy sources; ROI: Return of investment; VRF: Variable refrigerant flow;

Research on earthen materials has significantly increased in recent decades, but a few studies evaluate their hygrothermal performance at wall or building scale. The heterogeneous nature of earthen materials leads to significant uncertainties regarding their hygrothermal properties, often neglected in modelling. This study presents a transient hygrothermal model, developed in Python, and applied to a rammed earth wall of a house subjected to a continental climate in France. A sensitivity analysis was conducted using the RBD-FAST method, focusing on the water content within the wall and indoor heat and vapour fluxes. The uncertain inputs were the hygrothermal properties of the material and the environmental parameters. The results show that the variability of the water content is significant, mainly due to uncertainties in the sorption isotherm, which could be detrimental to the structure. Notable discrepancies appeared in indoor fluxes, which could affect energy consumption and hygric comfort predictions in whole building simulations.

Despite the growing popularity of model predictive controllers (MPCs) in building automation, there are few investigations of MPC in residential buildings. However, existing studies show promise for MPCs in buildings with high glazing, and radiant conditioning. This paper presents the development and results of a long-term MPC implementation at a full-scale research house during the heating season. The MPC manages the operation of hydronic floors to maintain thermal comfort while minimizing energy use. This work includes a novel MPC approach, forecasting approach, and parameter estimation technique. An existing data-driven model and sequential parameter estimation approach were modified and used in this work. Disturbance forecasting (including solar gains, equipment, and heat transfer to the ground, and water storage tanks) employs a clustering algorithm, then rule extraction with a decision tree. Compared to a rule-based controller (RBC), the MPC reduced energy use by 10%, and decreased the magnitude and duration of overheating.

Building performance simulations (BPS) are essential for managing rising energy demands but are computationally intensive, limiting their use in early-stage design. Surrogate models (SMs) offer fast approximations using machine learning but lack generalizability for changing building morphologies. This study introduces three morphology-based feature extraction methods – Zernike image moments, polar raycasting, and principal component analysis – to train a generalizable SM for predicting annual energy use intensity (EUI) of 2.5D residential buildings in Singapore. The best SM achieved a 9.2% RMSPE, with features yielding low error and avoiding overfitting. This framework enables efficient design space exploration, aiding architects and urban designers in optimizing masterplans for energy efficiency. In the final section, we demonstrate a design space exploration (DSE) framework with our generalizable SM using a novel parametric model, showcasing its speed, adaptability and convenience over conventional optimization workflows. Highlights Novel approach explores building morphology's impact to annual EUI . Using Zernike moments, ray casting and PCA to extract features from any 2.5D buildings. 4000 tuned SMs trained from hyperparameter optimisation. Most robust SM achieved 9.2% RMSPE when tested with 800 samples. SM's generalizability is proven by applying it to a design model with previously unseen parameters.

Ventilated photovoltaic curtain walls reduce buildings’ reliance on the electricity grid, transforming them into producers and consumers. The airflow and heat transfer characteristics within curtain walls are necessary for better photovoltaic and thermal efficiency. This paper establishes a natural convection model of the photovoltaic curtain walls, solved using the finite element method, focusing on the impact of geometric parameters on flow and heat transfer to determine the optimal channel geometry. Results show that the thickness significantly affects the photovoltaic curtain wall’s performance, with 200 mm thickness being optimal. Compared to direct contact with the building surface, using a 200-mm-thick channel reduces the photovoltaic module’s average temperature by 5.27%, increasing the photovoltaic efficiency and overall efficiency by 0.21 and 12.56%. Increasing the width and height enhances heat accumulation and thermal efficiency. Large heights improve the chimney effect and air velocity but minimally affect the average temperatures of the air and panels.