Electrical power systems
Price-based control of ancillary services
This study focuses on the real-time power imbalance in the power net, which arises as a consequence of errors in the prediction of both production and demand. As this power imbalance will increase both in size and in frequency, present arrangements to cope with this imbalance are no longer valid. This study will be formulated to study advanced ICT and price-based control frameworks for ancillary services (reserve capacity), which give consumers and producers clear, real-time financial incentives to adapt their consumption/production according to the actual needs of the power system. Decisions by consumers, producers, power exchanges and TSOs can be taken locally, based on local or national preferences and regulation, and a global quantity price. Still, the embedded incentives of the proposed framework should guarantee that all these local decisions together contribute to the global objectives of the EE power net: a reliable electric energy supply at the lowest costs. An important aspect of reliability of advanced ICT systems is to verify their robustness against cyber attacks and external manipulation. A hacker can potentially alter control signals and sensor data, and thereby threaten the stability of these systems. A security and fragility investigation is thus an essential part of this study. The study should demonstrate that instead of investing in additional expensive and environment-unfriendly reserve production or storage facilities with a low utilization rate, the reliability and economy can be enforced by intelligent ICT and control, based on both secure and robust systems.
Distributed control of power networks
This group of studies will serve to demonstrate the utility of distributed control systems theory to formulate, analyze and systematically design optimal control laws for power systems of various configurations and levels of complexity. In some studies time-varying price signals will be employed to reflect the current state of the physical system, as well as the currently required actions to optimally ensure its future integrity, reliability and efficiency. In others, local control agents will cooperate to attain the common goal(s), and to coordinate their actions to prevent any negative impacts of control actions of one agent in the region managed by another agent. Also, one or more higher-level supervisors may be present to resolve conflicts in those cases where the agents cannot reach any mutual agreement. Apart from simulation studies in cooperation with companies that have significant presence and experience in this domain (one of the partners is collaborating with EDF, for example), a specific experimental investigation is envisioned as well. We will look at distributed control methods for the incorporation of micro-combined heat power (micro-CHP, systems installed in local house holds) systems in the electricity grid. We have access to such a test facility with 6 of these micro-CHP's at Gasunie Research and Technology. The application of the new algorithms should enable the micro CHP (and possible other local sustainable power supplies) to be embedded in our electricity system, and to result in more efficient power generation and lower energy costs for households.
SmartGrids– New ICT tools for complex networks, Generation-Transmission-Distribution levels
New technologies in control, computers and communications have allowed the possibility of a future power grid, or Smartgrids, quite different from the present one. The objective of this AD is to integrate these new technologies to the grid and to present new tools for handling specific control problems related to the distributed control of large-scale electric power generation, transmission and distribution networks. To attain this goal, it is needed interactions from Power systems, Control systems, Computer science and Industry in order to bridge the gap between classical control engineering techniques and the more high-level intelligent control techniques from modern control and computer science. At Generation level, it is necessary to study the integration of large share of renewables, and in particular the integration of large off-shore wind farms that will need better communication and control with higher level controllers. At Transmission level, Wide Area Measurement Systems – Phase Measurement Units will propose new possibilities to control engineers. In fact, these devices merges distributed electrical measurements with GPS signals, and send them to other points of the network. This procedure allows receiving remote information with precise time stamps. Which signals, where to send and how to use them are the key questions for new distributed networked (intercommunicating) controllers. Finally, at Distribution level, new smart metering devices in each house, connected by information network to distributed controllers will allow (based in price signals) dynamic demanding as a new control input to the network. In the same way, current European regulation enforces house-owners to install small renewable energy generation devices (mainly photo-voltaic solar panels). This distributed generation brings the problems of bi-directional power-flow, monitoring, pricing and power quality. Distributed generation and dynamic demanding are very important topics since they affect the stability of the distribution network and its topology. Even if each system may be very simple, the huge number of individuals may bring the system to chaotic behaviour. All these aspects will be even enhanced with the planned large introduction of electrical and hybrid cars.
Integrated chemical and petrochemical plants
Coordination of batch reactor operation under constraints on the utilization of infrastructures
The computation of optimal policies for the operation of chemical batch reactors to optimize product yield or, most often, to minimize the batch time has become state of the art not only in research but also is increasingly applied in industry. Such policies are implemented by online controllers and state estimators are used to monitor the progress of the batch and to trigger corrective actions if needed. In larger production facilities, several large batch reactors are operated in parallel. These reactors require common infrastructure services, most importantly cooling capacity, but also emergency relief systems, exhaust systems and downstream storage and processing capacity. The joint request for these services currently is either ignored in the optimization of the operation of the individual batches and reactors or handled by a higher-level assignment of batch starting times under the assumption of fixed, known processing times for the different stages of the batches. The idea of this demonstrator is to include the competition for scarce or limited resources into the optimization of the individual batch runs by an auction-like decentralized negotiation process, as successfully applied in scheduling problems. Such decentralized coordination schemes will be investigated theoretically and for simplified examples in the theoretical research in TA 3. Their potential will be demonstrated for a case study on a batch reactor system with limited cooling capacity, which will be provided by BASF.
Optimal management of hydrogen networks
Hydrogen has become one of the main products in petrol refineries. As a result, hydrogen management plays a key role in the production of the different commercial oil fractions. All such plants are interconnected by several kilometers of diverse pipes forming a distribution network with different purities, capacities and operating at several pressures. Hydrogen networks have received attention in the literature from the point of view of their (re)design, but very few from that of real-time operation and, as far as we know, no commercial software is available in the sector for this purpose. We intend to test an integrated optimization based framework to optimize the distribution of the available hydrogen from producers to consumer facilities, so as to take advantage of low purity hydrogen supplies by combining streams of different purity levels and flows and, at the same time, ensuring operational restrictions. The major aim is to provide an effective and integrated decision support system for on-line, open-loop optimization and data reconciliation.
Transition management in chlorine plants
Chlorine is one of the base chemicals that are needed in a large number of chemical synthesis, e.g. for the production of vinylchloride and then PVC. Chlorine currently is produced by electrolysis of salt in a membrane process. A chlorine plant consists of many steps of mixing, purification and pH regulation before the salt brine enters the electrolysis cells. The cells are switched on and off individually. Due to the strong nonlinearity of the titration curves, the local pH regulation problems are difficult, especially when load changes occur during start-up and when cells are shut-down in an unplanned fashion. Large variations in pH lead to reduced efficiency and can damage the electrolysis cells. It will be investigated whether such load transitions can be realized more smoothly by a hierarchical control layer such that the production rate is higher and in particular the sensitive cells are not damaged, compared to the present manual upper layer control with decentralized standard controllers on the lower level.