Natural ventilation in buildings : detailed prediction of energy performance
MetadataShow full item record
Natural ventilation has during the last decade experienced a renaissance in Europe. The penalty of electrical energy use of traditional mechanical ventilation and air conditioning is the main reason. The present thesis, “Natural Ventilation in Buildings -Detailed Prediction of Energy Performance”, is the result of a PhD study funded by Hydro Aluminium/Wicona and The Research Council of Norway. The study has been carried out in close collaboration with fellow researchers Tommy Kleiven and Tor Arvid Vik. The focus for the present dissertation is detailed prediction of energy performance in buildings with natural ventilation. This is more difficult than in a similar but mechanically ventilated building because the energy flow in a naturally ventilated building is generally: • More dynamic due to variation in airflow paths and airflow rates. • More sensitive to internal parameters such as inter-zonal pressure differentials and control of airflow openings. • More sensitive to the outdoor environment. The dissertation adresses topics related to meteorological data, presssur characteristics in buildings with natural ventilation, and detailed energy performance predictions. The meteorological input data needed for detailed prediction of energy performance in buildings with natural ventilation are discussed. It was found that useful records of meteorological data exist for several hundred locations only in Norway. Based on such meteorological data records, methods for the creation of complete and reliable weather data files for use in building simulation were devised, which involved: • Development of methods for evaluating and, if necessary, calibrating the recorded radiation data. The calibration is often needed as the quality of field pyranometers, as well as the calibration and maintenance-routines, are limited. • Evaluation, implementation and validation of diffuse fraction model for predicting beam and diffuse radiation from measured global radiation. • Evaluation of suitable methods for applying wind data records for close locations to obtain realistic wind data for the location in question. • Methods for prediction of the ground temperatures based on recorded temperature at a certain depth in order to obtain realistic ground temperature profiles. These methods can be particularly useful for environments with significant variation in surface conditions on the ground due to e.g. variations in snow conditions or vegetation. All methods related to meteorological data records were implemented in computer codes. This allows for processing of such records to create typical statistical information such as wind-roses and statistical tables, and reliable ready-to-use input data for building simulation. This treatment of data, which describes the local climate at the building site in detail, can enable the use of the local climate as a guiding factor in the design of a building with natural ventilation. A detailed field experiment was carried out on a case study building utilising natural ventilation. The pressure differentials over the various components and airflow openings were accurately measured. The total pressure drop for the whole ventilation system was found to be 29.1Q+2.3Q2, where Q is the airflow rate [m3/s], indicating that viscous friction rather than turbulent losses dominates the pressure drop. Results also indicate that it is fully possible to use regular mechanical ventilation components as filters and heat exchangers in a system based purely on natural driving forces. The present work demonstrates how a detailed simulation model, able to realistically predict the energy performance of a building with natural ventilation, can be developed. The following topics are discussed: • How to model a detailed airflow network capable of realistically predicting the inter-zone airflow, and how to handle frequency controlled fans. • How to simplify the three-dimensional ground heat transfer problem down to one dimension for the ground-coupled part of the building fabric. • How to account for the occupants’ behaviour related to energy performance e.g. the opening and closing of external doors and the operation of curtains. • How to utilise poor and lacking meteorological data at the location in question to create reliable input data for simulation through use of good data at a close location. • Modelling of heat transfer coefficients in embedded ducts. • How to realistically model a run-around heat recovery system without doing detailed plant modelling. • Typical aspects related to dynamic building simulation such as imposed gains, dynamic modelling of heat transfer coefficients and modelling of the building construction elements. The simulation results were validated extensively against monitored and measured data in a real building, namely the Mediå School in Grong, Norway. The ventilation airflow rates were accurately predicted through simulations, both when the building was ventilated solely through utilisation of natural driving forces and when frequency controlled fans were used. The predicted energy consumption matched the measured consumption of the case study building closely throughout the whole year, both on an hourly, daily and monthly basis. The high level of detail enabled reliable quantification of all the main parameters related to energy performance in the Mediå School. One of the findings was that the introduction of thermal bridges in the building was responsible for a 40% increase in the hydronic energy consumption compared to the same building without thermal bridgesi. Further, the analysis of the embedded duct system revealed the importance of fans in order to enhance convective heat transfer at duct walls. During a hot period in June it was found that more than 60% of the cooling effect of the embedded intake duct was caused by conduction into the ground, the remaining being caused by the diurnal energy storage and lag in the duct walls. Results from the field experiments showed that the pressure transducers installed were not capable of detecting the low pressure differentials associated with pure natural ventilation, and thus not being suitable to deduce airflow rates. Further, significant calibration was needed for the transducers to give reliable values for pressure differentials associated with fan operation. Despite the calibration it was concluded that the transducers were unreliable even for the higher pressure differentials.One solution can be to use predictive control in buildings with natural ventilation. Thus, the installation of advanced and/or unreliable monitoring equipment can be avoided through use of simulation. The work demonstrates how state-of-the-art simulation tools can be used to optimise an existing building with natural ventilation. The other apparent use of such a tool is for a final validation of a suggested building and related energy-system design.