Simulating the Diffusion of Smart Meters Through an Artificial Society

Abstract

The need to create a more flexible and modern electricity system, which additionally can attribute to energy savings, has been started in the Netherlands in the 1990’s beginning with the introduction of liberalization and unbundling of those corporations responsible for electricity production and transport. Creating a more flexible and modern electricity system requires for instance the integration of small scale decentralized electricity production by individual consumers, with large scale centralized electricity production by corporations. Considerable technical and social modifications are required to make this system work. An electricity system that makes considerable use of ICT and information is required which is called a smart grid. One essential element in this smart grid is the smart meter. Small scale residential consumers can equip their homes with a smart meter. This smart meter can collect information concerning energy flows, can be used to automate certain tasks and can be employed to provide feedback to the consumer concerning energy usage. This feedback can be used by the consumer to engage in energy saving strategies. Consumers will need to accept smart meters, by which we mean that consumers will voluntarily make use of a smart meter. When many consumers accept these smart meters, the diffusion of smart meters is high. The Dutch government would like to have a situation where the diffusion of smart meters is at least 80%. This level is required by European law and is set in such a way as that it enables effective functioning of the smart grid. The acceptance of smart meters by an individual consumer is not a static phenomenon. Consumers have certain attitudes towards smart meters based on the social class to which they belong, but these attitudes can be influenced. Consumers can exchange information with each other and can receive information from for instance the media. These information exchanges can lead to adjustment of their attitudes towards smart meters. We take advantage of these facts and have made use of agent-based modeling and simulation to simulate the effect of information exchanges between consumers. The “Diffusion of Smart Meters Model and Simulation” can be used to study how over time the average acceptance of smart meters by a group of consumers changes due to the exchange of opinions between the consumers within this group. Additionally we can examine the effect of exposing consumers to positive information through the media. The design of the agent-based model and simulation presented in this thesis is guided by a conceptual framework and metamodel which has been created specifically to assist researchers in the design of agent-based models and simulations of socio-technical systems. This conceptual framework and metamodel deals with the sanctioning of these models sand simulations. We contend that the notion of traditional validation of a model and simulation is not applicable in the case of agent-based models and simulations of socio-technical systems which often exhibit complex and non-linear behaviors. This notion of traditional validation, which is intended at proving that a model and simulation corresponds well to the real world, is only applicable in the case where we have relatively simple problem situations which we intend to study with well understood models and simulations. When we have deal with complex problem situations and relatively complex models and simulations, we cannot prove that a model and simulation corresponds well to the real world. We now sanction our model and simulation, during which we test if we THINK if the model and simulation corresponds well to our problem situation. This sanctioning is a more subjective assessment of the usefulness of the model and simulation. The results from this simulation study indicate that a higher level of smart meter acceptance can be realized by making use of media influence and by letting consumers of a higher social class meet with each other. A high level of acceptance of smart meters by many individual consumers again will lead to a high diffusion of smart meters throughout society. The conceptual framework and meta-model have proven to be useful to be able to sanction the model and simulation compared to a traditional validation of the model and simulation. Sanctioning of this model and simulation has lead to the development of a simulation heuristic which has been used to provide a solution to the smart metering problem. the conceptual framework and metamodel are still in their initial design stage and will require additional improvements over time.