Summary:

Summary:

Summary:

Summary:

Summary:

Urban scale building simulation refers to the extension of building energy modelling to the urban scale. This can be seen as an aggregated city model, built upon aggregated building models. Another interpretation is to build city-wide models from top information at a country level. Simulation at a district scale and city scale is still in its infancy, therefore a review of different urban simulation workflows is illustrated with a focus on bottom-up approaches and identification of the major obstacles to building model accuracy and reliability.

References:

Summary:

In this module we will review tools and methods to match building energy simulation results to measured real data. Firstly we will analyse the main causes for discrepancy between building energy models and measured results. Secondly we will differentiate between different terms such as: validation, calibration, verification, replicability, uncertainty and sensitivity, uncertainty propagations. Workflows for different calibration techniques and tools (brute force, Bayesian calibration, etc.) are reviewed.

References:

Summary:

Summary:

Summary:

In this first module we focus on the Artificial Intelligence (AI) methods used to analyse large quantities of data in order to perform two main tasks: (1) to observe anomalies, trends and errors in variables and, (2) use this AI methods to predict and forecast different variables such as building energy consumption thus preventing abnormal settings or faults in energy production. Machine Learning techniques are a relevant predictions tools for demand-response applications in the context of the smart grid e.g.: to avoid matching of peak loads over small periods.

Summary:

In this second module the practical application of Machine Learning (ML) techniques to adjust system behaviour in the occurrence of anomalies and faults is reviewed. This means a ML system is able to detect and adjust HVAC settings or trigger alarm events when a condition is not met. Typical examples are: (1) the use of reinforcement learning applied to monitoring of occupancy patterns and the interaction with thermostats or lighting, (2) use of weather forecast or weather anomalies to predict energy demand and adjust operational settings accordingly.