Parametric residential heating energy, thermal comfort and CO2 concentration data for an energy-efficient semi-detached house in Ireland

Description

The dataset is a synthetic, simulation-based collection of numerical data representing the energy performance, thermal comfort, and indoor air quality (specifically CO₂ concentration) of a low-energy, naturally ventilated semi-detached residential dwelling in Ireland, an oceanic temperate climate zone. Generated through 45,000 parametric simulations using EnergyPlus and the jEPlus tool, the dataset captures a wide range of building design and operational scenarios for an A-rated semi-detached house, the most prevalent residential archetype in Ireland. It is stored in a comma-separated values (.csv) format and hosted on Mendeley Data, with the data originating from University College Dublin, Ireland. The dataset comprises 19 input parameters and 24 output parameters, detailed as follows: Input Parameters Sampled via LHS from Irish EPCs and TUS data: Window U-value (0.53–3.21 W/m²·K): Glazing heat loss. Floor U-value (0.06–0.85 W/m²·K): Floor insulation. Roof U-value (0.06–0.88 W/m²·K): Roof insulation. Wall U-value (0.12–0.85 W/m²·K): Wall heat loss. Orientation (0–360°): Solar exposure. HSPT (18–22°C): Heating setpoint. Lighting Density (1–6 W/m²): Lighting load. Occupancy Density (0.0125–0.3289 persons/m²): Occupant load. Equipment Density (16–24 W/m²): Equipment load. HVAC Efficiency (0–1): Heating efficiency. Infiltration (0.35–1.5 ACH): Air leakage. Metabolic Rate (60–360 W/person): Occupant heat. Clothing Value (0–1 clo): Thermal insulation. SHGC (0.1–1): Solar gain factor. WWR (40–90%): Window area ratio. Window Open Time (10–180 min): Ventilation duration. Window Close Time (10–180 min): Closed duration. Window Area (0.1–1 factor): Openable area. Ventilation (3.35–6.5 ACH): Air exchange rate. Output Parameters Calculated per simulation: c0: Heating (kWh/year): Annual heating energy. c1: Lighting (kWh/year): Annual lighting energy. c2–c8: Thermal Discomfort (hours/year): Hours outside comfort for kitchen, dining, sitting, bedrooms 1–3, facility. c9–c15: Setpoint Not Met (hours/year): Unmet heating hours for kitchen, dining, sitting, bedrooms 1–3, facility. c16–c20: CO₂ > 1000 ppm (hours/year): High CO₂ hours for kitchen, dining, bedrooms 1–3. c21: EUI (kWh/m²·year): Energy intensity. c22: Total Site Energy (kWh): Total energy use. c23: CO₂ Average (hours/year): Mean CO₂ > 1000 ppm hours. c24–c33: Fanger Index: PMV (-3 to +3) for 10 zones (kitchen, dining, corridor, sitting, top toilet, bedrooms 1–3, top corridor, bath). We see this dataset as a vital resource for understanding how heating energy, thermal comfort, and CO₂ levels interplay in low-energy, naturally ventilated homes in Ireland. It’s useful for us as architects to optimise building designs, for engineers to refine HVAC and ventilation systems, for policymakers to shape energy-efficient regulations, and for researchers like us to study occupancy impacts on indoor environmental quality.

Steps to reproduce

We begin by defining the archetype for our study. We select an A-rated semi-detached house, a prevalent type in Ireland, and model it using geometrical details like floor area and window-to-wall ratio from Irish Energy Performance Certificates (EPCs). We incorporate realistic occupancy schedules from Time Use Survey (TUS) data and pair this with weather data for Ireland’s oceanic temperate climate, such as a Dublin TMY file, to set the environmental context. Next, we establish the input parameters driving our simulations. We define 19 variables based on EPCs and TUS, including window U-value (0.53–3.21 W/m²·K), floor U-value (0.06–0.85 W/m²·K), roof U-value (0.06–0.88 W/m²·K), wall U-value (0.12–0.85 W/m²·K), orientation (0–360°), heating setpoint (18–22°C), lighting density (1–6 W/m²), occupancy density (0.0125–0.3289 persons/m²), equipment density (16–24 W/m²), HVAC efficiency (0–1), infiltration (0.35–1.5 ACH), metabolic rate (60–360 W/person), clothing value (0–1 clo), solar heat gain coefficient (0.1–1), window-to-wall ratio (40–90%), window opening time (10–180 min), closing time (10–180 min), window area factor (0.1–1), and ventilation rate (3.35–6.5 ACH). Then, we build our simulation model in EnergyPlus (v9.x+). We create a multi-zone representation with rooms like kitchen, dining, sitting, and bedrooms 1–3, using the AirflowNetwork module to simulate natural ventilation. We integrate window operation schedules based on occupancy and temperature, and model infiltration through surface cracks, ensuring accurate airflow dynamics. To generate the parametric runs, we use jEPlus (v2.x+) and apply Latin Hypercube Sampling (LHS). We construct an LHS matrix with 45,000 rows and 19 columns, covering our input ranges, and link it to our EnergyPlus IDF file. We run these 45,000 simulations on a robust system, outputting 24 parameters per run: heating (kWh), lighting (kWh), thermal discomfort and setpoint unmet hours per zone, CO₂ > 1000 ppm hours per zone, EUI (kWh/m²·year), total energy (kWh), average CO₂ hours, and Fanger PMV across 10 zones. We validate our base case to ensure accuracy. We install low-cost sensors in real Irish homes to collect energy, temperature, and CO₂ data at 1-minute intervals, calibrating weekly with high-grade sensors. We compare a simulation with median inputs to this data, expecting errors of 6–8% (e.g., heating 7.5% low, CO₂ 6.43% low, thermal discomfort 6.7–8% low per WHO/CIBSE TM59). We adjust the model if discrepancies exceed this range. Afterward, we clean the data. We compute z-scores for each output (z = (x - μ) / σ), keeping values between -3 and +3 to remove outliers, and export the results into a .csv file with 45,000 rows and 43 columns.

Institutions

Categories

Funding

Related Links

Licence