Denis Fan

Real-time building occupancy sensing using neural-network based sensor network

Current occupancy sensing technologies may limit the effectiveness of buildings controls, due to a number of issues ranging from unreliable data, sensor drift, privacy concerns, and insufficient commissioning. More effective control of Heating, Ventilation and Air-conditioning (HVAC) systems may be possible using a smart and adaptive sensing network for occupancy detection, capable of turning off services out of hours, and not over-ventilating, thus enabling energy savings, and not under-ventilating during occupied periods, giving comfort and health benefits. A low-cost and non-intrusive sensor network was deployed in an open-plan office, combining information such as sound level, case temperature, carbon-dioxide (Co2) and motion, to estimate occupancy numbers, while an infrared camera was implemented to establish ground truth occupancy levels. Symmetrical uncertainty analysis was used for feature selection, and a genetic based search to evaluate an optimal sensor combination. Selected multi-sensory features were fused using a neural network. From initial results, estimation accuracy reaching up to 75% for occupied periods was achieved. The proposed system offers promising opportunities for improved comfort control and energy efficiency in buildings.

A design model for building occupancy detection using sensor fusion

Building occupancy sensing is useful for control of building services such as lighting and ventilation, enabling energy savings, whilst maintaining a comfortable environment. However, a precise and reliable measurement of occupancy still remains difficult. Existing technologies are plagued with a number of issues ranging from unreliable data, maintaining privacy, sensor drift, change of use, and short-term financial pressures, including low quality parts and insufficient commissioning. A major performance barrier is currently the fitness to purpose, or otherwise of sensing technologies used. Sensor fusion techniques offer a way to make up for this, aiming to more reliably determine occupancy using a range of different indoor climatic variables. Over the last decade, artificial intelligence (AI) techniques have found some application for building controls, and can also be applied to occupancy estimation. We describe a novel methodology for building occupancy detection using a sensor fusion model based on the Adaptive Neuro-Fuzzy Inference System (ANFIS) algorithm. The system monitors indoor climatic variables, indoor events and energy data obtained from a non-domestic building to infer occupancy patterns.

Enabling urban-scale energy modelling: a new spatial approach

Urban-scale energy modelling provides an ideal tool for studying non-domestic energy consumption and emissions reduction at the community level. In principle, an approach based on the characteristics of individual commercial premises and buildings is attractive, but it poses a number of challenges, the most immediate of which is deciding precisely what to model. For a range of reasons connected with their self-contained nature, individual non-domestic buildings would ideally be selected. However, the main information sources available – digital mapping and business taxation data – are not based on ‘buildings’ and do not use the concept, thus making an automated approach problematic. At the same time, manual identification of the distinct buildings in a city is not a practical proposition because of the numbers involved. The digital mapping and business taxation data are brought together in the Local Land and Property Gazetteer (LLPG). An analysis of the relationships between the relevant elements, namely building polygons and premises attracting business taxation, allowed a unit to be defined that matches the definition of a ‘building’ in most circumstances and can be applied without the need for human intervention. This novel approach provides a firmer basis for modelling non-domestic building energy at the urban scale.

A dynamic thermal network approach to the modelling of Foundation Heat Exchangers

Foundation Heat Exchangers (FHX) are a novel form of ground heat exchanger for residential applications. The recently developed dynamic thermal network approach has been applied to formulate a model of the FHX that includes the basement, pipes and adjacent ground. This response factor approach allows complex three-dimensional geometries, such as this, to be represented and simulated efficiently. The formulation of the method and its application to the FHX is described along with a numerical procedure to calculate the required weighting factor series. An improved method of calculating these data and reducing it to a compact form is presented. Some modification of the original method has also been necessary to implement the boundary conditions associated with the heat exchanger pipes and ground surface. Data from an installation at an experimental house have been used to validate the model.

Advanced Occupancy Sensing for Energy Efficiency in Office Buildings

Control systems for heating, ventilation and air conditioning in non-domestic buildings often operate to fixed schedules, assuming maximum occupancy during business hours. Since lower occupancies usually mean less demand for heating, ventilation and air conditioning, energy savings could be made. Air quality sensing, often combined with temperature sensing, has performed sufficiently in the past for this if maintained properly, although sensor and control failures may increase energy use by as much as 50%. As energy costs increase, coupled with increased complexity in building services and reduced commissioning time, all placing ever higher demands on sensing, building controls must meet increasingly stringent environmental requirements, whilst also improving reliability. Sensor fusion offers performance and resilience to meet these demands, while cost and privacy are key factors which are also met. This article describes a neural network approach to sensor fusion for occupancy estimation. Feature selection was carried out using symmetrical uncertainty analysis, while fusion of sensor features used a back-propagation neural network, with occupant count accuracy exceeding 74%.