Optimal measurement and verification plan on lighting

Abstract

Measurement and Verification (M&V) has become an indispensable process in various incentive energy efficiency and demand side management (EEDSM) programmes to accurately and reliably measure and verify the project performance in terms of energy or cost savings. Due to the uncertain nature of the un-measurable savings, there is an inherent trade-off between the M&V accuracy and M&V cost. Practically, there are three types of quantifiable uncertainties coupled with the M&V process including measurement, modelling and sampling uncertainties. For large-scale lighting retrofit projects that require long-term continuous measurements, the desired sampling effort for savings determination contributes to a significant increase to the M&V cost. On the contrary, the measurement and modelling uncertainties are considered less significant in the lighting M&V process. In order to handle the sampling uncertainties and achieve the required M&V accuracy cost-effectively, three metering cost minimisation (MCM) models are proposed, namely spatial MCM model, longitudinal MCM model, and the combined spatial and longitudinal MCM model to assist the design of optimal M&V metering plans, by which the minimal metering cost is achieved with the satisfaction of the required M&V metering and sampling accuracy. In the proposed MCM models, the objective functions are the M&V metering cost that covers the procurement, installation and maintenance of the M&V metering system whereas the M&V accuracy requirements in terms of confidence and precision levels are formulated as the constraints. Generally, for lighting projects that have multiple homogeneous lighting groups with different sampling uncertainties, the spatial MCM model is most applicable when the lighting population are properly maintained to avoid lamp population decay. If no project population maintenance activities are carried out, then the lamp population will decay as time goes by. In such a case, the longitudinal MCM model is most suitable to optimise the sample sizes within adjacent reporting years for each lighting group. The combined spatial and longitudinal MCM models exhibits the best performance in terms of metering cost minimisation whilst satisfying the required M&V accuracy, especially for the lighting projects that have multiple lighting groups with different sampling uncertainties and different population decay dynamics. Optimal solutions to the proposed MCM models offer useful information in designing the optimal M&V metering plan, such as the required lighting samples to be measured in each lighting groups, the achieved sampling accuracy in terms of confidence and precision levels as well as the annual and total M&V metering cost for the studied lighting project. The advantages of the proposed MCM models are demonstrated by several lighting retrofit case studies. For the case studies, metering solutions obtained with or without optimisations are calculated and compared. The comparisons highlight the advantageous performance of the proposed MCM models. These MCM models are widely applicable to M&V projects with different technologies, population sizes, and sampling accuracy requirements. Since the lighting population decays as time goes by, the lighting project performance is not sustainable and vanishes rapidly without proper maintenance activities. The scope of the maintenance activities refers to the replacements of the failed lamps due to the occurrence of non-repairable lamp burnouts. Full replacements of all the failed lamps during every maintenance activity contribute to a tighter project budget due to the expense for the lamp failure identifications as well as the procurement and installation of new lamps. Since neither “no maintenance” nor “full maintenance” is preferable to the lighting project developers, an optimal maintenance planning (OMP) approach is also proposed to decide the optimal number of failed lamps to be replaced, such that the EE lighting project achieves sustainable energy savings whereas the project developers obtain their maximum benefits in the sense of a maximum cost-benefit ratio. The OMP problem is aptly formulated under a control system framework. According to existing studies on the lamp population decay modelling, the lamp population decay dynamics are taken as the plant of the control system. The number of lamps to be replaced is designed as the inputs of the control system. As different lighting technologies have different population decay dynamics, different procurement prices and different rebate tariffs, the control inputs can be optimally decided to satisfy the project budget constraints and project boundary constraints. The optimal maintenance planning problem is then translated into an optimal control problem and solved by a model predictive control (MPC) approach. Since the lighting population has a close relationship with the sample size determination, the optimal maintenance planning approach is also integrated with the proposed MCM models, which further improves the performance and flexibility to the applications of the proposed MCM models for the M&V metering plan designing.