Search

38th European Photovoltaic Solar Energy Conference and Exhibition; 967-972, Concurrently with the strong decline in installed PV system prices, the margin for operation and maintenance (O&M) operators has decreased, as O&M costs are commonly budgeted as a fixed percentage of installed system costs. Hence, there is a need to improve the efficiency of O&M operations. In the PV4.0 project, we are aiming to improve O&M efficiency across the value chain, by implementing Industry 4.0 concepts. One key part of these concepts is to implement an automatic fault detection system, that is used to quickly detect faults in PV plant operation, minimizing electricity generation losses. Furthermore, this fault detection system is to be used to automatically generate maintenance tickets, so corrective maintenance can be performed as soon as possible. In this paper, we present the development of two neural network (NN) based machine learning models. Model A is a four-layer NN, trained to classify individual datapoints as either normal or faulty. Detected faults are subsequently classified as either a hard (zero power output) or soft (power output higher than zero but lower than nominal) failures. Model B is a model that attempts to predict whether the number of failures in the next days exceed a certain threshold. Our results indicate that Model A has a fault detection rate of 79% in our testing dataset, while the performance of Model B is lower, at a fault detection rate of 49% for 1-day-ahead predictions, and only 30% for 2-day-ahead predictions. Further work will focus on improving the failure detection rate for both models. For Model B, we will additionally focus on improving the input dataset, so that better predictions might become possible. Model A will be implemented in a real-time O&M environment to test its performance in real operating conditions.

The physical and virtual dimensions of life are becoming more and more deeply interwoven. Society is merging with technology, giving rise to a global socio-technical ecosystem. In a society comprising people and machines as actors we often see people-to-people interactions mediated by machine and machine-to-machine interaction mediated by people. The speed and scale of this change and the differences in culture, language and interests make the problem of establishing effective means of communication and coordination increasingly challenging. Our vision, embodied in the SmartSociety project , is that a new generation of systems, hybrid (i.e., including humans and artificial peers, as well as social groups), distributed, open and large-scale, is needed to tackle these issues. In such systems, multitudes of heterogeneous peers will produce and handle massive amounts of data; peers will join/leave the system following unpredictable patterns with no central coordination and will interoperate at different spatial and temporal scales. Aware of the ethical issues, and by identifying the right incentive schemes and privacy levels, these systems will assist individuals and collectives in their everyday activities, coping with the diversity of the world and working in the presence of incomplete and incorrect information.

In this work we report on a method, based on the use of activation-inhibition mechanisms, for building and maintaining high-rate routing paths (data highways) in dense wireless sensor networks. We describe the algorithms devised and report on the outcomes of a simulation study, aimed at assessing the scalability and the performance of the proposed approach.

A major research challenge in distributed systems is the design of services that incorporate robustness to events such as network changes and node faults. In this paper we describe an approach ï½ which we refer to as EmbryoWare ï½ that is inspired by cellular development and differentiation processes. The approach uses ï½artificial stem cellsï½ in the form of totipotent nodes that differentiate into the different types needed to obtain the desired systemï½level behaviour. Each node has a genome that contains the full service specification, as well as rules for the differentiation process. We describe the system architecture and present simulation results that assess the overall performance and fault tolerance properties of the system in a decentralized network monitoring scenario.

Previous work has demonstrated the feasibility of distributed generation of data highways through a grid-based wireless sensor network. These highways carry packets across the network from source(s) to sink(s), and can be generated through a localised reaction diffusion process that does not require apriori knowledge of the global network topology. In this paper we consider the extension of this approach to a more realistic graph-based structure. We present the underlying algorithms, simulation results, and an analysis of the extent to which the algorithms can be fully distributed.