This report describes the forecasting model which was developed by team "4C" for the global energy forecasting competition 2017 (GEFCom2017), with some modifications added afterwards to improve its accuracy. The model is based on neural networks. Temperature scenarios obtained from historical data are used as inputs to the neural networks in order to create load scenarios, and these load scenarios are then transformed into quantiles. By using a feature selection approach that is based on a stepwise regression technique, a neural network based model is developed for each zone. Furthermore, a dynamic choice of the temperature scenarios is suggested. The feature selection and dynamic choice of the temperature scenarios can improve the quantile scores considerably, resulting in very accurate forecasts among the top teams.

This paper addresses the generation maintenance scheduling (GMS) dilemma in a deregulated power system. At first, under a centralized cost minimization framework, a GMS is formulated and set as the benchmark (cost-minimization GMS). Then, the cost-minimization is changed into a profit-maximization problem of generation companies (GENCOs) and the GMS is developed as a bi-level optimization. Karush-Kuhn-Tucker conditions are applied to transform the bi-level into a single-level mixed-integer linear problem and subsequently, Nash equilibrium is obtained as the final solution for the GMS under a deregulated market (profit-maximization GMS). Moreover, to incorporate reliability and economic regulatory constraints, two rescheduling signals (incentive and penalty) are considered as coordination processes among GENCOs and independent system operators. These signals are based on energy-not-supplied and operation cost, and ensure that the result of profit-maximization GMS is in the given reliability and social cost limits, respectively. These limits are obtained from the cost-minimization GMS. Lastly, the model is evaluated on a test system. The results demonstrate applicability and challenges in GMS problems. 

Wind turbine anomaly or failure detection using machine learning techniques through supervisory control and data acquisition (SCADA) system is drawing wide attention from academic and industry While parameter selection is important for modelling a wind turbine's condition, only a few papers have been published focusing on this issue and in those papers interconnections among sub-components in a wind turbine are used to address this problem. However, merely the interconnections for decision making sometimes is too general to provide a parameter list considering the differences of each SCADA dataset. In this paper, a method is proposed to provide more detailed suggestions on parameter selection based on mutual information. First, the copula is proven to be capable of simplifying the estimation of mutual information. Then an empirical copula-based mutual information estimation method (ECMI) is introduced for application. After that, a real SCADA dataset is adopted to test the method, and the results show the effectiveness of the ECMI in providing parameter selection suggestions when physical knowledge is not accurate enough.

Energy forecasting provides essential contribution tointegrate renewable energy sources into power systems. Today,renewable energy from wind power is one of the fastest growingmeans of power generation. As wind power forecast accuracygains growing significance, the number of models used forforecasting is increasing as well. In this paper, we propose anautoregressive (AR) model that can be used as a benchmarkmodel to validate and rank different forecasting models andtheir accuracy. The presented paper and research was developedwithin the scope of the European energy market (EEM) 2017wind power forecasting competition.

In this article, a parametric model for health condition monitoring of wind turbines is developed. The study is based on the assumption that a wind turbine’s health condition can be modeled through three features: rotor speed, gearbox temperature and generator winding temperature. At first, three neural network models are created to simulate normal behavior of each feature. Deviation signals are then defined and calculated as accumulated time-series of differences between neural network predictions and actual measurements. These cumulative signals carry health condition–related information. Next, through nonlinear regression technique, the signals are used to produce individual models for considered features, which mathematically have the form of proportional hazard models. Finally, they are combined to construct an overall parametric health condition model which partially represents health condition of the wind turbine. In addition, a dynamic threshold for the model is developed to facilitate and add more insight in performance monitoring aspect. The health condition monitoring of wind turbine model has capability of evaluating real-time and overall health condition of a wind turbine which can also be used with regard to maintenance in electricity generation in electric power systems. The model also has flexibility to overcome current challenges such as scalability and adaptability. The model is verified in illustrating changes in real-time and overall health condition with respect to considered anomalies by testing through actual and artificial data.

This article proposes a maintenance management and risk reduction approach. The approach introduces two reliability-based indexes called condition indicator and risk indicator. Condition indicator is a unit-less parameter that comes directly from monitored condition of a component and converts the categorical condition into a numerical value. Risk indicator in megawatt represents the risk imposed by the health of a component onto the system. To demonstrate application of the indicators, they are implemented through an hourly network constraint unit commitment problem and applied in a test system where the analysis of impact of condition of the generators to the operation is the new contribution. The results demonstrate how addition of such indicators will impact the operation of the grid and maintenance scheduling. The results show the benefit for the system operator as the overall failure risk in the system is taken into account, and the benefit for the asset owner as the direct impact of the maintenance to be carried out can be investigated. Two of the main outcomes of the maintenance management and risk reduction approach are as follows: asset owners can analyze their maintenance strategies and evaluate their impacts in the maintenance scheduling, and system operators can operate the grid with higher security and lower risk of failure.

This paper proposes an approach for maintenancemanagement of wind turbines based on their life. The proposedapproach uses performance analysis and anomaly detection(PAAD) which can detect anomalies and point out the originof the detected anomalies. This PAAD algorithm utilizes neuralnetwork (NN) technique in order to detect anomalies in theperformance of the wind turbine (system layer), and then appliesprincipal component analysis (PCA) technique to uncover theroot of the detected anomalies (component layer). To validatethe accuracy of the proposed algorithm, SCADA data obtainedfrom online condition monitoring of a wind turbine are utilized.The results demonstrate that the proposed PAAD algorithm hasthe capability of exposing the cause of the anomalies. Reducingtime and cost of maintenance and increasing availability and inreturn profits in form of savings are some of the benefits of theproposed PAAD algorithm.

This paper addresses passive and active preventive maintenance scheduling in an islanded microgrid with storage and renewable energy sources. At first, under a centralized framework, a single-level cost-minimization formulation for passive maintenance scheduling is developed and used as a benchmark in the operation. An independent microgrid operator is responsible for the operation in this framework. Then, through a bi-level formulation, the active maintenance scheduling and operation is carried out with profit-maximization objective. These two developed frameworks provide the houses with opportunity to earn profit and the regulator and the operator to analyze the performance of the system. The bi-level formulation is transformed into a single-level problem through Karush–Kuhn–Tucker conditions. Furthermore, the proposed model provides the capability of incorporating condition monitoring data into the operation. The model is validated through a test system and the outcomes demonstrate the advantages, applicability and challenges of utilizing the proposed model.

This paper proposes a model for strategic maintenance scheduling of offshore wind farms (SMSOWF) in a deregulated power system. The objective of the model is to plan the maintenance schedules in a way to maximize the profit of the offshore wind farm. In addition, some network constraints, such as transmission lines capacity, and wind farm constraints, such as labor working shift, wave height limit and wake effect, as well as unexpected outages, are included in deterministic and stochastic studies. Moreover, the proposed model provides the ability to incorporate information from condition monitoring systems. SMSOWF is formulated through a bi-level formulation and then transformed into a single-level through Karush-Kuhn-Tucker conditions. The model is validated through a test system, and the results demonstrate applicability, advantages and challenges of harnessing the full potential of the model.

This paper proposes an approach for stress condition monitoring and maintenance assessment in wind turbines(WT) through large amounts of collected data from supervisory control and data acquisition (SCADA) system. Theobjectives of the proposed approach are: to provide a stress condition model for health monitoring, to assess the WT’smaintenance strategies, and to provide recommendations on current maintenance schemes for future operations ofthe wind farm. At first, several statistical techniques, namely Principal component analysis, Pearson, Spearman andKendall correlations, mutual information, regressional ReliefF and decision trees are used and compared to assessthe data for dimensionality reduction and parameter selection. Next, a normal behavior model is constructed by anartificial neural network which performs condition monitoring analysis. Then, a model based on mathematical form ofProportional hazards model is developed where it represents stress condition of the WT. Finally, those two modelsare jointly employed in order to analyze the overall performance of the WT over the study period. Several cases areanalyzed with a five-year SCADA data and maintenance information is utilized to develop and validate the proposedapproach.