Gros Cul BBW Pov
Janet Denzel
My main current research concerns safe Reinforcement Learning and Data-Driven MPC. I currently have 6 PhD students working on connected topics. This work is strongly connected to the industry, with large industrial partners such as Equinor, DNV, Konsberg, but also smaller ones focusing on energy.
I also co-supervise several PhD projects related to the Artificial Pancreas problem, and multi-rotor wind turbine control. Finally, I supervise and co-supervise several industrial PhD projects based at Volvo (Sweden), focusing on powertrain control, 2nd life of batteries, and traffic management.
The green transition of the European energy system will require storage capacities in various forms. Strongly decentralized energy storage, placed directly at the consumers end, have a high value because they allow for reducing the energy and power flows in the transmission and distribution infrastructures. This form of energy storage is currently commercially developed by several companies in Scandinavia. One bottleneck of domestic energy storage is the high investment cost of batteries. It is therefore critical to exploit domestic energy storages in an optimal way, so as to maximize their value for the consumers.
The optimal usage of domestic energy storage is a complex task. Indeed, the domestic energy demand can be forecasted, but not accurately predicted. Besides, the prices of electric energy (spot prices) are only known in a day-ahead fashion. If domestic energy production is available (typically PV), that production is also difficult to predict accurately. In addition, energy storage is best used if it operates not only based on the spot prices, but also on various flexibility markets. However, this creates fairly complex requirements on the energy storage management system.
In this project, we will work with the company Pixii to investigate the optimal use of battery storage using data-driven and AI-driven approaches. The selection of the tools we will use will be done together with the MSc candidate.
Managing the energy consumption of buildings more efficiently will have a high benefit for the cost and safety of energy in the future power system. In particular, the energy demand of buildings presents opportunities for delivering flexible power and energy consumption to the power system, which can have a high value if exploited optimally.
However, important difficulties exist in achieving this optimal exploitation. Buildings are all different, they are very costly and difficult to model using first-principles, and they can have very stochastic responses. These issues call for using data-driven, AI-based methods for energy management. However, it can be difficult to have a sufficiently large and rich data set to use black-box AI for energy management in buildings. This calls for dedicated AI methods where prior and physical knowledge can be embedded.
In this project, we will work with SINTEF on experimental buildings and test several dedicated AI methods for creating Digital Twins for decision making and AI-driven control techniques for energy management in buildings. The selection of the tools we will use will be done together with the MSc candidate. The tasks will involve:
Intelligent energy consumption for buildings has become necessary to utilize energy resources efficiently and economically. The recent energy crisis and high price variations have brought this topic to more social focus lately. However, optimally managing the energy consumption in houses is a complex topic. Developing control models for homes must be automated to achieve optimal residential consumption. In that context, reliable data-driven control techniques are necessary to solve this problem.
Here at ITK, we have developed our own set of methods to perform data-driven MPC for energy management in houses. These solutions can be tested and refined via experiments on a real smart home that ITK operates. This project will extend an earlier MSc project on the topic and bring the ITK methods to a higher level of maturity. The tasks will involve:
This project aims at being either a MSc thesis, or a specialization project followed with a M.Sc. thesis. The student will working the research group of Prof. Sebastien Gros (sebasti...@ntnu.no). The work will be part of the research project MAIDOM.
The IoT is quicly entering our homes. Data from the IoT regarding domestic energy consumption are of high value for the society. Several companies are building software tools for integrating the data flow of IoT devices for energy, with various objectives in mind. ITK has also developed an in-house software to achieve that integration. An important component missing in the existing solutions is to provide advance data analytics and inform the users on their energy consumption and on how to improve it. Indeed, the current industrial solutions are basic, and provide incorrect feedback to the users. In this project, we will aim at improving the existing software with onboarding solutions, and develop advance data analytic methods for domestic consumers. The end ambition is to create a social media of energy with exciting vizualisations, where users can compare their performance, and compete in a friendly way to improve it.
Small-scale to large-scale indoor farming has been widely adopted across the globe in recent years as a sustainable food production method. It is an area of tremendous innovation and economic activity during the post covid period. Indoor farming, especially vertical farming has a wide range of benefits, for example a healthy source of food, low environmental impacts, local sourcing, and all-year-around production. Still, in the very early phase of development, vertical farming offers a great opportunity for us to contribute to making the system efficient to feed the world with a minimal environmental footprint. From this perspective, such highly controlled environment farming allows us cyberneticians to use various tools at our disposal to make them optimal.
This project is planned as a joint specialization project and master thesis. The students will be working in the research group of Prof. Sebastien Gros (sebasti...@ntnu.no). During the project phase, the students are expected to do the following tasks:
A good example on building an automated hydroponics system can be found in =nyqykZK2Ev4&t=11s . The students will be provided with a basic design and possibly with most of the necessary electronics and mechanical structure to start building the units immediately upon starting the project.
During the master thesis, the students could utilize these hydroponics units for exploring the scope of optimization and control in indoor farming. There is a wide range of problems to explore in this area. We might focus on problems like stress analysis on plants upon varying different climatic conditions. But this is open for students to be more creative and explore different aspects of the problem during the thesis.
Ocean waves represent a substantial source of renewable energy and will be part of the solution when transitioning to emission-free energy systems world-wide. The Swedish company CorPower Ocean ( ), one of the leading companies in wave energy conversion, is currently in the demonstration phase with a system that partly uses technology originating from NTNU. It utilizes the combination of innovative mechanical solutions and optimal control methods to maximize power output while ensuring survival.
Optimal model-based operation of wave energy converters (WEC) requires sufficiently accurate representation of the physical system. Numerical models are inevitably affected by model uncertainties due to approximations and simplifications. The performance of model-based estimation, prediction, and control algorithms is directly linked with the overall power output of the WEC, which makes the improvement of the model accuracy an important task.
The goal of this master thesis is the development and evaluation of learning algorithms for uncertain model parameters. For this task, both the estimation of first-principle model parameters and phenomenological representations may be investigated. A simulation environment is provided for evaluation of the developed learning algorithms.
This project is planned as a joint specialization project and master thesis. The students will be working in the research group of Prof. Sebastien Gros (sebasti...@ntnu.no) in collaboration with CorPower Ocean. The following points should be addressed during the project/thesis:
Various gross receipts taxes are imposed upon private bankers; pipeline, conduit, steamboat, canal, slack water navigation and transportation companies; telephone, telegraph and mobile telecommunications companies; electric light, water power and hydroelectric companies; express companies; palace car and sleeping car companies; and freight and oil transportation companies. The categories are as follows:
Transportation: The tax rate is 50 mills and the tax is based on gross receipts from passengers, baggage and freight transported within Pennsylvania, and on intrastate shipment of freight and oil. This does not include transportation by motor vehicles or railroads. The gross receipts tax on transportation services is reported to the PA Department of Revenue on RCT-113a. Firms are required to file reports and remit tax payments annually by March 15 for taxable gross receipts in the prior year.
Private Bankers: The tax is imposed on private bankers doing business in Pennsylvania at a rate of 1 percent on gross receipts from commissions from loans; banking services; discounts on loans; charges or fees on depositor accounts; rents; rentals of safe deposit boxes; interest from bonds, mortgages, premiums and dividends; profits from the purchases and sales of securities; and many other related services. Form RCT-131 must be filed annually with payment by Feb. 15 following the close of the prior calendar year.
92d504bec8