Radhakrishna S Aithal

Robust control and optimisation of energy consumption in daylight—artificial light integrated schemes

Energy efficiency strategies based on daylight—artificial light integrated schemes have proved to be efficient by many researchers worldwide. But much larger energy savings with the benefit of visual and thermal comfort can be achieved when systems integration strategies are competently designed. They require a high level of expertise and familiarity with new design techniques. This study describes the results of three computational models suitable for the optimum integration of visual comfort, thermal comfort, and energy consumption in schemes where daylight and artificial light are integrated. This mainly involves: (i) a system identification approach in lighting control strategy, (ii) a fuzzy logic based controller to reduce glare, increase uniformity and thermal comfort, and (iii) an adaptive predictive control scheme for the dimming of artificial light. In addition to the above models the scheme must take account of occupancy and user wishes. The anticipated synergetic effects of the computational models have been validated using climate data. A SIMULINK environment is established for the real time control and analysis of daylight—artificial light integrated schemes. Overall, the schemes maximise energy cost saving while optimizing the performance and the quality of the visual environment.

An adaptive neuro-fuzzy model for the prediction and control of light in integrated lighting schemes

Advanced lighting simulation tools as well as computationally intelligent systems present the possibility of using a model based on computation as a means of controlling lighting on the visual task. Lighting control has now become an essential element of good design and an integral part of energy management programmes. This paper presents a novel computational model suitable for the adaptive predictive control of artificial light in accordance with the variation of daylight. Simulated data and an adaptive neuro-fuzzy inference system are incorporated into the model. The software package Radiance is used to carry out the simulation. In this process, the role of a simulator is considered as the source of the system knowledge by which a supervised learner, implemented in adaptive neuro-fuzzy inference system is trained for faster predictions. The goal of this paper is to make use of the benefits of the hybridization between simulation and machine learning for the purpose of light control.

Authors’ response to GJ Levermore

We would like to thank Professor Levermore for commenting on our work. In this paper, we present a computational model based on adaptive neuro-fuzzy inference system (ANFIS); not a pure artificial neural network model. Fuzzy logic and neural networks are complementary technologies in the design of intelligent systems. Each method has its merits and demerits. Neural networks are essentially low-level computational structures and algorithms that offer good performance in dealing with sensory data, while fuzzy logic techniques often deal with issues such as reasoning on a higher level than neural networks. However, since fuzzy systems do not have much learning capability, it is difficult for a human operator to tune the fuzzy rules and membership functions from the training data set. Also, because the internal layers of neural networks are always opaque to the user, the mapping rules in the network are not visible and are difficult to understand; furthermore, the convergence of learning is usually very slow and not guaranteed. Hence the ANFIS technique, which is a promising approach for reaping the benefits of both fuzzy systems and neural networks, is adopted in our work. As for the information from recent literature, soft computing techniques (SCT) such as fuzzy systems, ANN and genetic algorithms are not new in the field of building, plant and lighting applications. International energy agency’s projects DELTA, EDIFICO, NEUROBAT4 and the research work of Seongju Chang (HISSTO) and Antoine Guillemin all have proved SCT to be a promising technology with respect to complex integrated and adaptive control problems encountered in the building energy sector. Our research aims at effective computational models for daylight /artificial light integrated schemes, a part of which is presented in this paper. Through ANFIS, we have achieved a more effective scheme than the above mentioned fuzzy/neural network systems, as it is more flexible for adaptive predictive control schemes. Training using measured data is time consuming and tedious, so we used simulated data for offline learning, and the model obtained is suitable for online adaptation. While considering lighting control schemes alone, this will not be beneficial, but to fulfill our target of self-adaptive daylight artificial light integrated schemes, this will be very suitable. Our work is progressing in this direction. A self-adaptive system such as ANFIS would not need very careful tuning, as is the case for classical adaptive control. Models based on soft computing techniques would progressively adapt themselves to the building and climate characteristics.