Publicación (Tesis doctorales)

Contribution to the design, implementation and standardization of semantic middleware architectures for the Smart Grid

Molina, Jesus Rodriguez
The Smart Grid, conceived as the power grid enhanced with Information and Communication Technologies aimed to optimizing electricity consumption, enabling a bidirectional participation in the energy provided and improving the power grid with features like Demand Response, Demand Side Management or Optimal Power Flow, is becoming one of the most compelling Cyber-Physical Systems that is being currently developed. Its capacity to increase usable energy in a sustainable manner, along with collect more information about its actual usefulness, makes possible the improvement of living standards for many people all around the world in a more transparent and open way. What is more, since it is able to integrate the comparatively small power supplied by the Renewable Energy Sources provided by the end users that participate in the Smart Grid (which become “prosumers”, as they both produce and consume electricity), it democratizes access to energy and enables a higher degree of competitiveness between traditional actors in the energy markets and newcomers, thus granting the overall improvement of the services that can be provided by the utility companies. However, there are still many open issues that must be solved before completely using the Smart Grid to our advantage. Among these open issues, interoperability of its installed devices is a major one. The equipment that is used in deployments of this kind (Advanced Metering Infrastructure, Phasor Measurement Units, Remote Terminal Units, etc.) is manufactured by different companies with different backgrounds and interests. Therefore, the implementation of their products has usually different ways to transmit information or even proprietary solutions with a low degree of compatibility with other pieces of hardware. In a way, the status quo is similar to the situation of computer networks before the first standards were released: a plethora of manufacturers offer their own solutions to provide services and connectivity, but they struggle to work cooperatively with developments of other equipment vendors that may have different perspectives on the technologies that can be used for data transmission. In addition to that, the services that should be available for either a Smart Grid or one of its smaller scale counterparts (microgrids, nanogrids) are not made clear, neither in terms of what services they should be or where they should be located. While there are some high level functionalities that are usually regarded as almost mandatory (the aforementioned Demand Side Management, device registration, Demand Response, Optimal Power Flow), other more data-centric facilities are often portrayed in a way vaguer manner. Security specifications, the existence of semantic capabilities, how to access the capabilities in a specific deployment or even how the hardware devices become integrated is not described with enough detail. That issue jeopardizes the main purpose of installing and developing components in this area of knowledge, because it makes difficult a further integration of both legacy systems that may have been used by large utility companies for a long time and new developments done by smaller companies that want to play a role in the Smart Grid. Fortunately, many of these challenges can be solved by implementing a software layer located between the applications that can be included for the benefit of the end users and the hardware and network infrastructure installed for package and binary data interchange. This software layer, commonly referred as middleware, has as its main purpose abstracting the heterogeneity and complexity of the underlying distributed hardware components, so that it will offer to the high, more application-based layer a collection of facilities of homogeneous, centralized appearance, usually shaped as an Application Programming Interface that can be accessed by the application developers. Middleware is a very useful software tool for Cyber-Physical Systems and distributed solutions because it grants the integration of almost any kind of device, either by adding software components in the device itself or in another part of the system, which must be open enough to have the components installed or have the required computational capabilities to have those components installed. The main original contribution to knowledge of this doctoral thesis is offering a proposal for a model of a semantic middleware architecture for the Smart Grid, based on software components for distributed solutions. This model is aimed to be used in any kind of deployment related to the Smart Grid, as well as providing a common set of components and interfaces to be observed in future implementations. This architecture has been called Common Middleware Architecture (CMA), as it aims to provide the necessary software components for middleware development under any imaginable use case within this application domain. It has been designed based on the experience accumulated from several research projects where the implementation of a middleware layer was one of the main achievements. CMA has been designed with the main needs of a middleware solution in mind, such as hardware abstraction, context awareness, device registration, interfaces for the upper level, securitization and device integration. While the main domain of CMA is the Smart Grid, and demonstrators based on the Smart Grid have been used to validate it, CMA can also be adapted to other environments. All in all, the main objective of this thesis is creating a reliable framework for the development of middleware solutions for the Smart Grid, which can be used in other application domains where there are requirements of hardware abstraction and service availability resembling the ones that can be found in this area of knowledge. Another major objective of this thesis is making contributions to the standardization of middleware development for the Smart Grid, so that there will be a specific set of services to be developed in order to comply with the most important functionalities of middleware (hardware abstraction, homogeneous set of services for applications, encasing services based on semantic capabilities and security). These two objectives have been achieved with the contributions done in the study of the State of the Art, the inference of open issues and challenges, the establishment of a list of functional and non-functional requirements and the validation of the proposal put forward in this manuscript. The solutions developed can be regarded as the background of the architecture described here, and therefore its performance should be good enough for the functionalities carried out for this kind of software layer, which should be present in any distributed or Cyber-Physical System that uses a collection of deployed pieces of equipment with different capabilities. Besides, since this is a middleware solution solves problems for issues present in other distributed and/or Cyber-Physical Systems (the Internet of Things, underwater robotics) it can be ported to other domains with ease, as services as high level interface access or device registration are used in those situations as well.
Tipo de publicación:
Tesis doctorales
Tipo de publicación:
PhD Thesis