1,721,302 research outputs found
Stochastic Scheduling of Wind-Integrated Power Systems
Full text linkThe cost of balancing supply and demand will increase as power systems are decarbonised,
because the requirement for operating reserve will increase with the wind
penetration, while the flexible fossil-fuel generators, which have been the traditional
providers of reserve, will be displaced. While these costs can be mitigated through increased
interconnection, energy storage, and demand-side market participation, a fundamental
review of system operational policy is also needed to ensure that the available
reserves are scheduled optimally. Stochastic Unit Commitment can find the commitment
and dispatch decisions that minimise the expected system costs, including the potential
costs of unserved energy, given the short-term uncertainties of wind and other variables.
It therefore has the potential to provide the most efficient possible paradigm for the operation
of wind-integrated systems. Because the system’s ability to respond to wind fluctuations
is constrained by intertemporal limitations of the other components, time domain
simulations are needed to assess the performance of different operational strategies or
generator fleet characteristics. However, Stochastic Unit Commitment has demanding
computational requirements that can render it impractical for long-term simulations of a
large power system.
This thesis develops a new tool for simulating the operation of large, wind-integrated
power systems using stochastic scheduling, with the emphasis on computational efficiency.
Embedded within it are new models for characterising time series of aggregated
wind output and wind forecast errors; these models are integrated with a Stochastic Unit
Commitment algorithm within a Monte Carlo framework. We explore simplifications
that can mitigate the computational burden without unduly compromising the quality
of the analysis. Simulations with the tool show that fully stochastic scheduling can reduce
operating costs by around 4% relative to traditional deterministic approaches, in a
system with a 50% wind penetration
Alternative design strategies of distribution systems
Full text linkIn contrast with traditional approaches based either on the analysis of a small
specific area or on idealistic networks, the proposed methodology determines optimal
network design policies by evaluating alternative planning strategies on statistically similar
networks. The position of consumers influences the amount of equipment used to serve
them. Therefore, simple geometric models or randomly placed points used in previous
researches are not adequate. Using an algorithm based on fractal theory, realistic consumer
sets are generated in terms of their position, type and demand to allow statistical evaluation
of the cost of different design policies.
In order to systematically deal with the problem of determining justifiable network
investments, the concept of economically adapted distribution network was investigated and
applied in the context of a loss-inclusive design promoting efficient investment policies
from an overall social perspective.
The network’s components are optimized, after yearly load flow calculations, based
on the minimum life-cycle cost methodology, balancing annuitised capital investments and
maintenance costs against the cost of system operation. Evaluating the cost of each
particular design over statistically similar networks allows statistically significant
conclusions to be drawn.
The main results include the optimal number of substations for typical urban and
rural LV, HV and EHV distribution systems, network costs (investment, purchasing and
maintenance) and losses as well as the sensitivity of optimal network design to future energy prices and cost of equipment.
The impact of the increasing amount of microgeneration on networks has not been
fully addressed to date. There have not been clustering problems in existing networks as a
result of customers choosing to install microgenerators, either as a new device or as a
replacement of a previous heating system. The operation of microgeneration connected to
the distribution network can cause statutory voltage limits, recommended voltage unbalance
levels and switchgear fault ratings to be exceeded. However, there are a range of
distribution network designs and operating practices and thus the impact will vary
accordingly.
The operation of distribution networks is approached considering the existence of
single or three-phase loads and microgeneration. This would however cause the network to
be unbalanced and hence, traditional methods that consider a three-phase balanced system
would provide misleading results.
Every residential daily load’s behaviour shows rapid shifts from “load valleys” to
high peaks due to the random and frequent “switch on/off” of appliances. Modelling each
load individually will reveal problematic operating conditions which were not considered
when using a smooth load profile. Thus, each and every domestic load was represented by a
different load profile and the impact on losses was evaluated.
Relating losses, voltages, currents and load unbalance ratio leads to conclusions
about the way how to optimise the network with DG. The aim was to investigate and
develop methodology for evaluation of the long-term loss-inclusive optimal network design
strategies and to determine the effect of the penetration of microgeneration, such as CHP
and PV, in realistic distribution networks and optimal network planning. The need for
reinforcement of network components will depend on the level of generation and on the
extent to which reverse power flows occurs. In most parts of the network, microgeneration
exports will not be sufficient to result in any need for network investment. However, if the
network was to be planned accounting with DG, capital investment scenarios are presented
and compared to existing networks trying to accommodate clusters of microgeneration
A geographically disaggregated approach to integrate low-carbon technologies across local electricity networks
Full text linkMeeting climate targets requires widespread deployment of low-carbon technologies such as distributed photovoltaics, heat pumps and electric vehicles. Without mitigating actions, changing power flows associated with these technologies would adversely impact some local networks. The extent of these impacts, and the optimal means of avoiding them, remains unclear. Here we use local-level data and network simulation to estimate variation in future network upgrade costs in over 40,000 geographical regions comprising all of Great Britain. We find that costs vary substantially between localities, and are typically highest in urban areas, and areas with highest deployment of heat pumps and electric vehicles. We estimate reductions in required upgrades associated with local flexibility, which vary substantially between localities. We show that using geographically disaggregated data to inform flexibility deployment across the country could reduce network upgrade costs by hundreds of millions of pounds relative to an approach that treats localities as homogeneous
Assessing local costs and impacts of distributed solar PV using high resolution data from across Great Britain
Full text linkHighly spatially resolved data from across Great Britain (GB) are combined with a distribution network modelling tool to assess impacts of distributed photovoltaic (PV) deployment up to 2050 on local networks, the costs of avoiding these impacts, and how these depend upon context. Present-day deployment of distributed PV, meter density, and network infrastructure across GB are found to be highly dependent on rurality, and data on these are used to build up three representative contexts: cities, towns, and villages. For each context, distribution networks are simulated, and impacts on these networks associated with PV deployment and growth in peak load up to 2050 calculated. Present-day higher levels of PV deployment in rural areas are maintained in future scenarios, necessitating upgrades in ambitious PV scenarios in towns and villages from around 2040, but not before 2050 in cities. Impacts of load growth are more severe than those of PV deployment, potentially necessitating upgrades in cities, towns, and villages from 2030. These are most extensive in cities and towns, where long feeders connect more customers, making networks particularly susceptible to impacts. Storage and demand side response are effective in reducing upgrade costs, particularly in cities and towns
Low-complexity decentralized algorithm for aggregate load control of thermostatic loads
No full textThermostatically controlled loads such as refrigerators are exceptionally suitable as a flexible demand resource. This article derives a decentralized load control algorithm for refrigerators. It is adapted from an existing continuous time control approach, with the aim to achieve low computational complexity and an ability to handle discrete time steps of variable length - desirable features for embedding in appliances and high-throughput simulations. Simulation results of large populations of heterogeneous appliances illustrate the accurate aggregate control of power consumption and high computational efficiency. Tracking accuracy is quantified as a function of population size and time step size, and correlations in the tracking error are investigated. The controller is shown to be robust to errors in model specification and to sudden perturbations in the form of random refrigerator door openings.Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.Intelligent Electrical Power Grid
Optimal Transmission Investment Strategies for Sustainable Power Systems
Full text linkMaintaining security and reliability in the electricity supply is fundamental to the functioning
of a modern society and drives the need for adequate transmission capacity for both market
participants and customers. Planning the investment in transmission has always been a
complicated undertaking due to the high development costs and long lead times. Furthermore,
to anticipate the future needs of customers is a task as difficult as that of cost-effective
planning and construction of new facilities. Trying to find treatments for some of these issues
represents a major motivation for this thesis.
This thesis investigates the problem of how much reinforcement a transmission system
requires when a significant proportion of wind generation is integrated into an existing
transmission system. A multi-period transmission planning model is developed for
determining optimal transmission capacity by balancing amortised transmission investment
costs and annual generation costs subject to network security constraints, The model employs
the security-constrained DC optimal power flow formulation and applies a solver
(DashXpress) to obtain the results of the remaining linear large-scale optimisation problem.
This thesis begins by exploring the impact of wind generation on the determination of
appropriate levels of system capacity on the transmission network starting from the premise
that it is no longer cost effective to invest in sufficient network capacity to accommodate
simultaneous peaks from all generators. As such, a significant finding of this study is that
conventional and wind generation should share network capacity. Given the acknowledged
increase in uncertainty to security of supply due to difficulties in wind generation forecast this
thesis also explores the optimal sourcing of generation reserve, and investigates investment in
transmission capacity to exploit the cost benefits offered by standing reserve.
Finally, the thesis presents and evaluates an alternative associated with transmission operation
and investment level of risk and uncertainty by introducing more flexibility to the way the
transmission system is operated. Application of Quadrature Boosters and Demand Side as
model of corrective control, brings savings in operating costs without jeopardizing the level of
system security, enables better utilisation of existing facilities and reduces the demand for
new transmission investment
Dynamic time-of-use electricity pricing for residential demand response: design and analysis of the Low Carbon London smart-metering trial
Full text linkThis thesis describes the trial design and analysis of the Low Carbon London (LCL) residential dynamic Time-of-Use (dToU) trial. This trial investigated the potential for dToU tariffs to deliver residential demand response to the Supplier, where it may contribute to system balancing through Supply Following (SF) actions, and to the distribution network operator (DNO), where it may be used for network Constraint Management (CM). 5,533 households from the London area participated in the trial and their consumption was measured at 30 minute resolution. 1,119 of these received the dToU tariff, which subjected them to CM and SF price events that were designed according to the specific requirements of these respective use cases. A novel, data driven, engagement ranking index was developed that allowed stratification of subsequent results into sets of the most engaged consumers, who may be indicative of a future populace that is more experienced/engaged in home energy management. Demand response (DR) was calculated relative to baseline model that used the dToU group mean demand as an input, with aggregate response levels calculated over a range of time, socio-economic and household occupancy related variables. Taking a network perspective, the reliability of CM event response was examined and two simple linear models presented as candidate predictors of response level, which was found to be consistent with an 8% reduction in demand. The network capacity contribution of residential DR was theorised to consist of two components: “mean response” and “variance response”, and the real impact of these was investigated using the LCL gathered data. Potential risks to the network from low price induced demand spikes were explored empirically using the SF event data and the times of highest risk were identified. The extensive metadata set gathered from trial participants was processed into some 200 numerical variables. A correlation analysis was performed which was visualised using weighted correlation network graphs. A number of parameters were found to predict response level, but responsiveness
(the level of deliberate engagement) could only be reliably measured by engagement rank.Open Acces
Towards intelligent operation of future power system: bayesian deep learning based uncertainty modelling technique
Full text linkThe increasing penetration level of renewable energy resources (RES) in the power system brings fundamental changes of the system operating paradigms. In the future, the intermittent nature of RES and the corresponding smart grid technologies will lead to a much more volatile power system with higher level uncertainties. At the same time, as a result of the larger scale installation of advanced sensor devices in power system, power system engineers for the first time have the opportunity to gain insights from the influx of massive data sets in order to improve the system performance in various aspects. To this end, it is imperative to explore big data methodologies with the aim of exploring the uncertainty space within such complex data sets and thus supporting real-time decision-making in future power system. In this thesis, Bayesian Deep learning is investigated with the aim of exploring data-driven methodologies to deal with uncertainties which is in the following three aspects. (1) The first part of this thesis proposes a novel probabilistic day-ahead net load forecasting method to capture both epistemic uncertainty and aleatoric uncertainty using Bayesian deep long short-term memory network. The proposed methodological framework employs clustering in sub-profiles and considers residential rooftop PV outputs as input features to enhance the performance of aggregated net load forecasting. Numerical experiments have been carried out based on fine-grained smart meter data from the Australian grid with separately recorded measurements of rooftop PV generation and loads. The results demonstrate the superior performance of the proposed scheme compared with a series of state-of-the-art methods and indicate the importance and effectiveness of sub-profile clustering and high PV visibility. (2) The second part of this thesis studies a novel Conditional Bayesian Deep Auto-Encoder (CBDAC) based security assessment framework to compute a confidence metric of the prediction. This informs not only the operator to judge whether the prediction can be trusted, but it also allows for judging whether the model needs updating. A case study based on IEEE 68-bus system demonstrates that CBDAC outperforms the state-of-the-art machine learning-based DSA methods and the models that need updating under different topologies can be effectively identified. Furthermore, the case study verifies that effective updating of the models is possible even with very limited data. (3) The last part of this thesis proposes a novel Bayesian Deep Reinforcement Learning-based resilient control approach for multi-energy micro-grid. In particular, the proposed approach replaces deterministic network in traditional Reinforcement Learning with Bayesian probabilistic network in order to obtain an approximation of the value function distribution, which effectively solves Q-value overestimation issue. The proposed model is able to provide both energy management during normal operating conditions and resilient control during extreme events in a multi-energy micro-grid system. Comparing with naive DDPG method and optimisation method, the effectiveness and importance of employing Bayesian Reinforcement Learning approach is investigated and illustrated across different operating scenarios. Case studies have shown that by using the Monte Carlo posterior mean of the Bayesian value function distribution instead of a deterministic estimation, the proposed BDDPG method achieves a near-optimum policy in a more stable process, which verifies the robustness and the practicability of the proposed approach.Open Acces
A stochastic optimization framework for anticipatory transmission investment
Full text linkAn unprecedented amount of renewable generation is to be connected to the UK grid in the coming decades, giving rise to new power flow patterns and warranting unprecedented amounts of transmission investment. However, significant uncertainty surrounds the state of the electricity system, primarily in terms of the size, location and type of new generators to be connected. These sources of uncertainty render the system planner unable to make fully informed decisions about future transmission investment.
This thesis presents a stochastic formulation for the transmission expansion planning problem under uncertainty in future generation developments. The problem has been modelled as a multi-stage stochastic optimization problem where the expected system cost is to be minimized. Uncertainty is captured in the form of a multi-stage scenario tree that portrays a range of possible future system states and transition probabilities. A set of investment options with different upgradeability levels and construction times have been included in the formulation to capture the diverse choices present in a realistic setting, where the planner can choose to invest in an anticipatory manner. A novel multi-cut Benders decomposition scheme is used to render the model tractable for large systems with multiple scenarios and operating points. The developed tool can identify the optimal long-term investment strategy based on the triptych of economic efficiency, adequate security provision and acceptable risk.
Simulation results on test systems validate that the stochastic approach can lead to further expected cost minimization when compared to methods that ignore the planner’s decision flexibility. Moreover, decisions are taken with subsequent adaptability in mind. The benefit of keeping future expansion options open is properly valued; investment paths that enable future delivery at lower costs are favoured while premature project commitment is avoided.Open Acces
Role and value of flexibility options in supporting cost-effective power system expansion planning under uncertainties
No full textDecarbonisation actions are driving progressive changes in the technological landscape and operational paradigms in power systems, leading to major challenges and uncertainties. Among else, substantial infrastructure investments are required to accommodate new assets and higher demand, but decision-makers face difficulties as the commitment to flawed investment plans could result in considerable economic, social, and environmental costs. This research aims to address the needs for enhanced coordination, advanced modelling frameworks, and flexible investment options in efforts to manage the risks associated with planning under uncertainties and to facilitate a cost-effective transition to a sustainable future. First, the computational performance of sophisticated modelling frameworks is addressed through the development of a machine learning-enhanced multicut Benders decomposition approach for the stochastic power system expansion planning problem. Its application in the case studies enables the solution of a previously intractable large-scale problem and leads to reductions in average master problem computation times of up to 72%. The focus is then placed on turning the challenge of accommodating future electrical demand of electric vehicles into an opportunity for strategic decision-making by employing smart charging concepts as investment options. For this purpose, the roles and values of Grid-to-Vehicle, Vehicle-to-Grid, and Vehicle-to-Building, in supporting cost-effective expansion planning, are evaluated through novel computationally efficient modelling in a multi-stage stochastic programming framework. Lastly, this research proposes a multi-stage integrated transmission and distribution expansion planning model that co-optimises investments in assets with diverse techno-economic characteristics to investigate the benefits of flexibility options in this context. The role of smart investment alternatives is evaluated with emphasis on investment flexibility as a means to hedge against uncertainties and their value is quantified using insights from Real Options Theory. The case studies validate flexibility options, including smart charging concepts, as viable and highly valuable alternatives to infrastructure reinforcements under all examined conditions.Open Acces
- …
