Architecture, Building and Planning - Theses
Permanent URI for this collection
Search Results
1 - 1 of 1
-
ItemDevelopment of an optimal control strategy for hybrid ventilation of office buildings( 2016)Hybrid ventilation can save operational building energy, if it is designed and controlled appropriately. Peak electricity demand during summer time can also be reduced. It is important that the control strategy should be developed together with the ventilation system at the design stage of the buildings. However, it is not known how to develop the control strategy before buildings are built. Current practice is that the required data needs to be obtained from the actual building. This research proposes a method to develop control strategies for hybrid ventilation at the design stage. The research method was devised by combining the advantages and two modelling techniques: phenomenological modelling, and data-driven modelling. First, the hybrid ventilation system is modelled using phenomenological simulation software tools. The phenomenological model developed was used to generate data. Then the simulated data was used to develop the thermal network model and the simplified airflow model which could be used for identifying the optimal control strategies. Through literature review, computer simulation, and the three experimental case studies, it was found that TRNSYS Type 56 & COMIS were reliable transient simulation software tools which could simulate an acceptable representative of the actual building behaviour. The most suitable simplified model for hybrid ventilation system was the thermal network model. The investigation was carried out on Shed D, a small single room building located at the Burnley Campus, The University of Melbourne. The thermal network model could provide reasonable accuracy, and used less computing time than the phenomenological model. Most importantly the thermal network model offered physical insight into building’s thermal characteristics. In other words, the building’s physical parameters could be incorporated into the model coefficients. With the known properties of building components, the air change rate for natural ventilation in the building can be estimated in term of the ventilation heat loss, one of the lumped parameters of the thermal network model. In addition to single zone stand alone buildings, multi-zone buildings can also be modelled by the thermal network technique. The research showed that the thermal network model worked for actual multi-zone buildings through the case study, the ATC building, Swinburne University of Technology. The clear correlation between the results of TRNSYS Type 56 & COMIS and the thermal network model was also demonstrated. The simplified air flow model was created from TRNFLOW results which were verified with the measured air flow rate. Air exchange measurement was implemented on the Doug McDonell building, University of Melbourne using CO2 as the tracer gas. The coupling between the simplified airflow model and the thermal network model was developed. This technique could handle the variation of the air flow rate over time. The air flow rate is in the form of Infiltration/ventilation heat loss conductivity which is a parameter of the thermal network. The coupling between the two models could improve the accuracy of the predictions. When there was high fluctuating wind speed, the improvement in fit between the measured data and the simulated data was approximately 20%. Based on the concept originally proposed by Spindle (2004) [Spindle, HC 2004, ‘System Identification and Optimal Control of Mixed Mode Cooling’, PhD thesis, Massachusetts Institute of Technology, Massachusetts], the optimal control program was developed using MATLAB, 2012b. The crucial part of the models for switching between modes was discovered, which is the disturbance of the model. This needed to be identified because it contains the previous time step output information, enabling the previous output to be used for calculating the next time step output. The simulation results showed that the optimal control strategy offered 20% saving in energy consumption compared to the Static mode (windows are closed all times). The computational time was approximately one minute per day of simulation time. With this computing time, it is possible to practically apply this technique for a real building management control system. The research presents a seven-step procedure for a hybrid ventilation system: TRNFLOW modelling; Data selection; Air flow modelling; Thermal network modelling; Integration of airflow model and thermal network model; Mechanical ventilation modelling, and Optimal control strategy. The seven-step procedure enables the optimal control strategy of a hybrid ventilation system can be identified at the design stage.