In this paper, we present an implementation of a model predictive controller (MPC) for wind farm power tracking problem. The controller is evaluated in the high-fidelity PAral-lelized Large-eddy simulation Model (PALM). By taking measurements from PALM, we show that the closed-loop MPC can provide power reference tracking while reducing force variations on a farm level by solving a constrained optimization problem at each time step. A six turbine wind farm case study is presented in which the controller operates with yawed turbines that increases the potential power that can be harvested with the wind farm, and we show that it is possible to track a reference power signal that temporarily exceeds the power harvested when operating under the so-called greedy control settings.

@en

Distributed fault diagnosis has been proposed as an effective technique for monitoring large scale, nonlinear and uncertain systems. It is based on the decomposition of the large scale system into a number of interconnected subsystems, each one monitored by a dedicated Local Fault Detector (LFD). Neighboring LFDs, in order to successfully account for subsystems interconnection, are thus required to communicate with each other some of the measurements from their subsystems. Anyway, such communication may expose private information of a given subsystem, such as its local input. To avoid this problem, we propose here to use differential privacy to pre-process data before transmission.

@en

In the context of increasingly strict requirements for building energy efficiency, Aquifer Thermal Energy Storage (ATES) systems have emerged as an effective means to reduce energy demand for space heating and cooling in larger buildings. In the Netherlands, over 2000 systems are currently active, which has already raised issues with spatial planning in some areas; current planning schemes may lack the flexibility to properly address variations in ATES operation, which are driven by uncertainties across a broad range of time scales – from daily changes in building energy demand, to decadal trends for climate or groundwater conditions. This work is therefore part of a broader research effort on ATES Smart Grids (ATES-SG), which has focused on more adaptive methods for ATES management and control. In particular, improved control schemes which allow for coordination between neighboring ATES systems may offer more robust performance under uncertainty (Rostampour & Keviczky, 2016). The case studies for the ATES-SG project have so far focused on idealized cases, and on a historical simulation of ATES development in the city center of Utrecht. This poster will present an additional case study for the city center of Amsterdam, which poses several geohydrological challenges for ATES: for instance, variable density flow due to salinity gradients in the local aquifer, and varying depths for ATES systems due to the thickness of the aquifer. To study the effect of these conditions, this case uses an existing 15-layer geohydrological model of the Amsterdam region, cropped to an area of 4500m x 2500m around the Amsterdam Zuidas district. This rapidly developing business district is one of the densest areas of ATES use in Amsterdam, with 32 well doublets and 53 monowells currently registered. The geohydrological model is integrated with GIS data to accurately represent ATES spatial planning; simulated well flows are provided by a model predictive control component. This model is then simulated for two cases: a baseline decoupled configuration without coordination, and a case in which a subset of adjacent ATES systems is managed centrally to avoid overlaps between stored thermal volumes. Given that the thickness of the local aquifer offers significant potential for further ATES adoption in the area, such a coordinated approach could help maximize the benefits of future ATES development. @en

Traditional deterministic robust fault detection threshold designs, such as the norm-based or limit-checking method, are plagued by high conservativeness, which leads to poor fault detection performance. On one side they are ill-suited at tightly bounding the healthy residuals of uncertain nonlinear systems, as such residuals can take values in arbitrarily shaped, possibly non-convex regions. On the other hand, they must be robust even to worst-case, rare values of the modeling and measurement uncertainties. In order to maximize performance of detection, we propose two innovative ideas. First, we introduce threshold sets, parametrized in a way to bound arbitrarily well the residuals produced in healthy condition by an observer-based residual generator. Secondly, we formulate a chance-constrained cascade optimization problem to determine such a set, leading to optimal detection performance of a given class of faults, while guaranteeing robustness in a probabilistic sense. We then provide a computationally tractable framework by using randomization techniques, and a simulation analysis where a well-known three-tank benchmark system is considered.

@en

Aquifer Thermal Energy Storage (ATES) technology can lead to major reductions in energy demand for heating and cooling in buildings. ATES systems rely on shallow aquifers to seasonally store thermal energy and have become popular in the Netherlands, where a combination of easily accessible aquifers and strict energy regulations makes the technology especially relevant. However, this rapid adoption has made their management in dense urban areas more challenging. For instance, thermal interferences between neighboring systems can degrade storage efficiency. Policies for the permitting and spatial layout of ATES thus tend to be conservative to ensure the performance of individual systems, but this limits the space available for new systems – leading to a trade-off between individual system performance, and the overall energy savings obtained from ATES in a given area. Furthermore, recent studies show that operational uncertainties contribute to poor outcomes under current planning practices; systems in the Netherlands typically use less than half of their permitted water volume. This further reduces energy savings compared to expectations and also leads to an over-allocation of subsurface space. In this context, this work investigates the potential of a more flexible approach for ATES planning and operation, under which neighboring systems coordinate their operation. This is illustrated with a three-building idealized case, using a model predictive control approach for two control schemes: a decoupled formulation, and a centralized scheme that aims to avoid interferences between neighboring systems (assuming perfect information exchange). These control schemes are compared across a range of scenarios for spatial layout, building energy demand, and climate, using a coupled agent-based/geohydrological simulation. The simulation indicates that centralized operation could significantly improve the spatial layout efficiency of ATES systems, by allowing systems to be placed more densely without penalizing their individual performance. This effectively relaxes the trade-off between individual system performance and collective energy savings as observed in the decoupled case. The continued adoption of ATES technology provides a window of opportunity to revisit existing practices for the layout and operation of urban ATES systems, as information exchange – supported by appropriate spatial planning – could offer significant potential towards improved performance under operational uncertainties. @en

This paper presents a stochastic model predictive control approach for a thermal grid with uncertainties in the consumer demand profiles. This approach leads to a finite-horizon chance-constrained mixed-integer linear optimization problem at each sampling time, which is in general non-convex and hard to solve. Earlier approaches for such problems are either suboptimal ad-hoc methodologies, or computationally intractable formulations. We provide a unified framework to deal with production planning problems for uncertain systems, while providing a-priori probabilistic certificates for the robustness properties of the resulting solutions. Our methodology is based on solving a random convex optimization problem to compute the uncertainty bounds using the so-called scenario approach and then, solving a robust mixed-integer optimization problem with the computed randomized uncertainty bounds at each sampling time. Using a tractable approximation of uncertainty bounds, the proposed problem formulation retains the complexity of the problem without chance constraints. In the presented thermal grid application this implies that a robust mixed-integer program is solved to provide a day-ahead prediction for the thermal energy production plan in the grid. The performance of the proposed methodology is illustrated using Monte Carlo simulations and employing two different problem formulations: optimization over input sequences (open-loop MPC) and optimization over affine feedback policies (closed-loop MPC).

@en

In this paper, we present an energy management framework for building climate comfort systems that are interconnected in a grid via aquifer thermal energy storage (ATES) systems in the presence of two types of uncertainty namely private and common uncertainty sources. The ATES system is considered as a large-scale storage system that can be a heat source or sink, or a storage for thermal energy While the private uncertainty source refers to uncertain thermal energy demand of individual buildings, the common uncertainty source describes the uncertain common resource pool (ATES) between neighbors. To this end, we develop a large-scale uncertain coupled dynamical model to predict the thermal energy imbalance in a network of interconnected building climate comfort systems together with mutual interactions between the local ATES systems. A finite-horizon mixed-integer quadratic optimization problem with multiple chance constraints is formulated at each sampling time, which is in general a non-convex problem and hard to solve. We then provide a computationally tractable framework based on an extension to the so-called robust randomized approach which offers a less conservative solution for a problem with multiple chance constraints. A simulation study is provided to compare two different configurations, namely: completely decoupled, and centralized solutions.

@en

This paper presents a complete model of a building heating and cooling equipment and a ground source heat pump (GSHP) coupled with an aquifer thermal energy storage (ATES) system. This model contains detailed mathematical representations of building thermal dynamics, ATES system dynamics, heat exchanger (HE), and GSHP together with heating and cooling circulated water pipelines. In particular, we derive a static (steadystate) model for a countercurrent HE and an electrical GSHP as a function of water flow rates and input/output water temperatures in both sides of HE and GSHP to determine the amount of heat transfer from ATES to HE then HE to GSHP, and vice versa. We then develop a model predictive control framework to keep track of the desired building zone temperature, while meeting operational constraints. In such a framework, a finite-horizon mixed-integer nonlinear optimization problem is solved at each sampling time. A simulation study illustrates the functioning of our proposed predictive model.@en

This paper presents a solution method for a day-ahead stochastic reserve scheduling (RS) problem using an AC optimal power flow (OPF) formulation. Such a problem is known to be non-convex and in general hard to solve. Existing approaches follow either linearized (DC) power flow or iterative approximation of nonlinearities, which may lead to either infeasibility or computational intractability. In this paper we present two new ideas to address this problem. We first develop an algorithm to determine the level of reserve requirements using vertex enumeration (VE) on the deviation of wind power scenarios from its forecasted value. We provide a theoretical result on the level of reliability of a solution obtained using VE. Such a solution is then incorporated in RS-OPF problem to determine up-and down-spinning reserves by distributing among generators, and relying on the structure of constraint functions with respect to the uncertain parameters. As a second contribution, we use the sparsity pattern of the power system to reduce computational time complexity. We then provide a novel recovery algorithm to find a feasible solution for the RS-OPF problem from the partial solution which is guaranteed to be rank-one. The IEEE 30 bus system is used to verify theoretical developments together with a comparison with DC counterpart using Monte Carlo simulations.

@en

The application of seasonal Aquifer Thermal Energy Storage (ATES) contributes to meet goals for energy savings and greenhouse gas (GHG) emission reductions. Heat pumps have a crucial position in ATES systems because they dictate the operation scheme of the ATES wells and therefore play an important role in utilizing the storage potential of the subsurface. In the Netherlands, suitable climatic and geohydrological conditions in combination with progressive building energy efficiency regulation have caused the adoption of ATES to take off, resulting in a situation where demand for ATES exceeds the available subsurface space in many urban areas. The most important aspects in this problem are A) the permanent and often unused claim resulting from static permits for ATES operation, and B) excessive safety zones around wells to prevent interaction between wells. Both aspects result in an artificial reduction of subsurface space for potential new ATES systems. Recent research has shown that ATES systems could be placed much closer to each other, and that a controlled/limited degree of interaction between them can actually benefit the overall energy savings of an entire area. Two different simulation experiments were carried out to evaluate the effect of an adaptive permit capacity policy, as well as revised layout guidelines for ATES wells. Our solution provides a framework in which smaller distances between wells and adaptability of the permit volume plays a key role, to allow for optimal utilization of subsurface space for ATES and maximize GHG emission reduction. This paper shows how the total GHG emission reduction of an area can be increased by intensifying the use of the aquifer by allowing (some) interaction between ATES wells, which opens up unused but claimed subsurface space, and increase the number of heat pumps and ATES systems installed.@en

Aquifer Thermal Energy Storage (ATES) is an innovative building technology that can be used to store thermal energy in natural subsurface formations [1, 4, 10]. In combination with a heat pump, ATES can reduce the energy demand of larger buildings by more than half, which has made the technology increasingly popular in northern Europe (see Figure 1). Furthermore, the climate and subsurface conditions required for ATES use can be found in areas across Europe, Asia and North America. By the middle of the century, roughly half of the world’s urban population is therefore expected to live in areas technically suitable for ATES [2].@en

This paper presents a stochastic model predictive control problem for a class of discrete event systems, namely stochastic max-plus linear systems, which are of wide practical interest as they appear in many application domains for timing
and synchronization studies. The objective of the control problem is to minimize a cost function under constraints on states, inputs and outputs of such a system in a receding horizon fashion. In contrast to the pessimistic view of the robust approach on uncertainty, the stochastic approach interprets the constraints probabilistically, allowing for a sufficiently small violation probability level. In order to address the resulting nonconvex chance-constrained optimization problem, we present two ideas in this paper. First, we employ a scenario-based approach to approximate the problem solution, which optimizes the control inputs over a receding horizon, subject to the constraint satisfaction under a finite number of scenarios of the uncertain parameters. Second, we show that this approximate optimization problem is convex with respect to the decision variables and we provide a-priori probabilistic guarantees for the desired level of
constraint fulfillment. The proposed scheme improves the results in the literature in two distinct directions: we do not require any assumption on the underlying probability distribution of the system parameters; and the scheme is applicable to high dimensional problems, which makes it suitable for real industrial applications. The proposed framework is demonstrated on a twodimensional production system and it is also applied to a subset
of the Dutch railway network in order to show its scalability and study its limitations.@en

Aquifer Thermal Energy Storage (ATES) can yield significant reductions in the energy use and greenhouse gas (GHG) emissions of larger buildings, and the use of these systems has been rapidly growing in Europe – especially in the Netherlands, where over 3000 systems are currently active in urban areas. However, the successful management of this technology poses a range of policy challenges, due to its reliance on subsurface resources and to the possibility of thermal interactions across adjacent systems. In particular, recent research suggests that ATES planning policies should acknowledge a potential trade-off between the total energy or GHG savings which can be obtained by ATES systems within a given area, and the economic returns realized by individual system operators. To better understand this compromise, this paper follows a simplified version of the multi-objective robust decision making framework (Kasprzyk et al., 2013), using an idealized agent-based model of ATES adoption and operation coupled with a geohydrological subsurface model. This simulation approach was used to investigate suitable options for the spatial planning of ATES systems, by exploring the behaviour of the coupled system under a set of sociotechnical and geohydrological uncertainties. A multiobjective evolutionary optimization algorithm was then applied to search for ATES well layout parameters which perform well in relation to the assessment criteria. The optimization identified a set of planning parameters which describe a Pareto-efficient trade-off between the individual and collective performance of ATES systems under uncertainty. @en

This paper proposes a building energy management framework, described by mixed logical dynamical systems due to operating constraints and logic rules, together with an aquifer thermal energy storage (ATES) model. We develop a deterministic model predictive control strategy to meet building thermal energy demand. At each sampling a mixed integer quadratic optimization problem is formulated. We then provide a simulation study using an agent-based model and a geohydrological simulation environment (MODFLOW) to illustrate the performance of the framework.

@en

Aquifer Thermal Energy Storage (ATES) systems can significantly reduce the energy use and greenhouse gas emissions of buildings in temperate climates. However, the rapid adoption of these systems has evidenced a number of emergent issues with the operation and management of urban ATES systems, which require careful spatial planning to avoid thermal interferences or conflicts with other subsurface functions. These issues have become particularly relevant in the Netherlands, which are currently the leading market for ATES (Bloemendal et al., 2015). In some urban areas of the country, the adoption of ATES technology is thus becoming limited by the available subsurface space. This scarcity is partly caused by current approaches to ATES planning; as such, static permits tend to overestimate pumping rates and yield excessive safety margins, which in turn hamper the energy savings which could be realized by new systems. These aspects are strongly influenced by time-dependent dynamics for the adoption of ATES systems by building owners and operators, and by the variation of ATES well flows under uncertain conditions for building energy demand. In order to take these dynamics into account, previous research (Jaxa-Rozen et al., 2015) introduced a hybrid simulation architecture combining an agent-based model of ATES adoption, a Matlab control design, and a MODFLOW/SEAWAT aquifer model. This architecture was first used to study an idealized case of urban ATES development. This case evidenced a trade-off between the thermal efficiency of individual systems and the collective energy savings realized by ATES systems within a given area, which had already been suggested by other research (e.g. Sommer et al., 2015). These results also indicated that current layout guidelines may be overly conservative, and limit the adoption of new systems. The present study extends this approach to a case study of ATES planning in the city centre of Utrecht, in the Netherlands. This case is particularly relevant due to a combination of dense ATES development and complex subsurface conditions. An agent-based model of ATES adoption was thus parameterized to represent historical development patterns in the area over the 1998-2015 period, as well as plausible future adoption dynamics under a range of socio-technical uncertainties. An existing geohydrological model (Deltares, 2009) was used to represent local subsurface conditions. Preliminary results from this case study indicate that the idealized dynamics obtained in the previous case can also be observed under more realistic conditions; the geographic constraints introduced by building plot layouts and other spatial features tend to further constrain the adoption of new systems, emphasizing the risk of a scarcity of space under current layout guidelines. Furthermore, order effects appear to play a more significant role for system efficiency than in the idealized case. Earlier adopters thus tend to benefit from higher thermal efficiency due to the transient development of thermal bubbles, which could make older systems more robust to thermal interactions. In order to better understand the relationships between these processes and the operation of ATES wells under uncertainty, the case study will be extended by incorporating a Model Predictive Control approach for simulated ATES operation. @en

This paper presents a control-oriented model for combined building climate comfort and aquifer thermal energy storage (ATES) system. In particular, we first provide a description of building operational systems together with control framework variables. We then focus on the derivation of an analytical model for ATES system dynamics. The dynamics of stored thermal energy over time in each well of an ATES system is the most important concept for a building climate control framework. This concept is proportional to the volume and temperature of water in each well of an ATES system at each sampling time. In this paper we develop a novel mathematical model for both dynamical behavior of volume and temperature of water in each well of an ATES system and provide detailed steps for estimating the model parameters. To illustrate the applicability of our proposed model, a comparison based on an extensive simulation study using an aquifer groundwater simulation environment (MODFLOW) is provided. @en

Background: ATES is application is growing Application of seasonal Aquifer Thermal Energy Storage (ATES) contributes to energy saving and Greenhouse Gas (GHG)-reduction goals (CBS, 2015; EU, 2010, 2014). Recently it was shown that ATES is applicable in several parts of the world (Bloemendal et al., 2015). While in most parts of the world adoption is just beginning, in the Netherlands progressive building energy efficiency regulation already caused the adoption of ATES to take off (Heekeren and Bakema, 2015; Sommer et al., 2015). As a result of the large number of ATES systems in the Netherlands, the subsurface plays a crucial role in the energy saving objectives of The Netherlands (Kamp, 2015; SER, 2013). Problem: suboptimal use of the subsurface for energy storage ATES systems accumulate in urban areas, as can be expected with a large growth of ATES systems; at many locations in Dutch cities demand for ATES transcends the available space in the subsurface (Li, 2014; Sommer et al., 2015). Within in the Dutch legal framework and state of technology optimal use of the subsurface is not secured; i.e. minimizing the total GHG emissions in a certain area. (Bloemendal et al., 2014; Li, 2014). The most important aspects in this problem are A) the permanent and often unused claim resulting from static permits and B) excessive safety zones around wells to prevent interaction. Both aspects result in an artificial reduction of subsurface space for potential new ATES systems. Recent research has shown that ground energy storage systems could be placed much closer to each other (Bakr et al., 2013; Sommer et al., 2015), and a controlled/limited degree of interaction between them can actually benefit the overall energy savings of an entire area. Solution: the approach and first results of our research project on ATES Smart Grids The heating and cooling demand of buildings is a dynamic and hard to predict process, due to effects such as weather, climate change, changing function and usage of buildings over time. This naturally also applies to the required storage capacity in the subsurface. Because of these uncertainties the only way to optimally use the subsurface is to shift the organization of the subsurface space use from the planning phase to the operational phase. Our solution therefore provides a framework in which adaptability plays a key role. Optimal use can only be achieved when users have insight in the status and their effect on the resource they are exploiting (Ostrom, 1990). Therefore exchange of information is necessary for individual users to adapt their operation based on the current state of the subsurface and their energy demand via a controller and compensation measures. To arrive at a proof-of-concept based on our approach, we use expertise and models from different fields such as administrative policy and decision making, systems and control, and hydrogeology. A central element is the so-called agent-based model (ABM), which is a technique widely used in administrative policy and decision making in order to simulate the behavior of actors. Each agent has a controller that determines how the geothermal energy system must satisfy the energy demand of the building. Thus properties and characteristics of the building and the system are included in this controller. We apply a so-called Model Predictive Control (MPC) approach, which means that the controller takes into account the expected energy demand in the future, and also how its control strategy influences the performance of its own resources. The computed control action is implemented on a groundwater model of the area including all geothermal energy systems, which then serves as the basis for planning the control actions over the next period. This concept is developed in 3 steps up to the scale of a typical Dutch town. Currently we have a proofof- concept based on a fictitious academic model (1), which we are currently scaling up to the city-center area of Utrecht (2) after which we then incorporate the entire city of Amsterdam (3). The first results of the academic model are promising (Jaxa-Rozen et al., 2015), it shows that stable situations result even for ATES system that are placed a lot closer together than what current regulations would dictate. This has a positive effect on the total @en