Fig 6 - uploaded by Hamid Shaker
Content may be subject to copyright.
Net demand forecast of direct and indirect methodologies using WNN for the first day of November in Ireland case study.
Context in source publication
Context 1
... settings and training sets are selected the same as Alberta's case study. Fig. 6 represents the net demand forecasting results of the WNN for both direct and indirect methodologies in the first day on November 2012 as a sample. Net demand forecast accuracy in this test case is lower than that for Alberta since wind power penetration is remarkably high in Ireland power system. Intermittent behavior of wind power ...
Citations
... Net load is defined as the difference between consumption and renewable energy generation [3]. NLF can be achieved directly (i.e., a single forecast of the net load) or indirectly (i.e., by calculating the difference between the load and the RES generation forecasts) [4]. Direct NLF has become increasingly important lately due to its computational advantage over indirect NLF and the availability of net load data [5]. ...
Modern microgrids require accurate net load forecasting (NLF) for optimal operation and management at high shares of renewable energy sources. Machine learning (ML) principles can be used to develop precise and reliable NLF models. This paper evaluates the performance of different ML models, that are optimally trained using supervised learning regimes, for direct short-term net load forecasting (STNLF) in renewable microgrids. Different categories of ML models, such as neural network, ensemble, linear regression, nearest neighbor, and support vector machine were used. The comparative assessment was conducted utilizing historical net load, meteorological, and time-related categorical data acquired from the renewable integrated microgrid of the University of Cyprus in Nicosia, Cyprus. The results showed that all STNLF ML models achieved normalized root mean square error (nRMSE) values below 10%. Amongst the investigated models, the Bayesian neural network (BNN) presented the highest forecasting accuracy, exhibiting a daily average error of 3.58%. In addition, the BNN model yielded robust forecasts regardless of the season and weather conditions. Finally, the results demonstrated that optimally constructed ML models can be applied to provide STNLF in renewable integrated microgrids, which can be used by microgrid operators to efficiently control and manage their assets.
... Specifically, the direct involves the use of only one forecast (i.e., net load). On the other hand, the indirect is the difference of the load and the renewable energy sources (RES) generation forecasting [3]. Although the indirect NLF strategy is well addressed in the literature, the direct NLF strategy has emerged in recent years due to: (a) increased accessibility of net load data compared to load consumption and RES generation data [4], and (b) computational advantage (as only a single model is required to be trained) compared to the indirect NLF strategy [5]. ...
Net load forecasting (NLF) is a key component for the efficient operation and management of microgrids at high shares of renewables. Depending on the forecasting strategy followed, NLF is classified as direct or indirect. In this paper, a performance comparison was conducted between indirect and direct short-term NLF (STNLF) strategies in renewable microgrids. A STNLF model was constructed by utilizing Bayesian neural network (BNN) principles applied to datasets obtained from the University of Cyprus microgrid and buildings. For the indirect STNLF, historical load and photovoltaic (PV) generation data, along with weather and categorical time-related data were used as inputs to develop the optimized BNN models for load and PV generation forecasting. The direct STNLF model achieved lower error (3.98% at the microgrid level) compared to the indirect one.
... The direct NLF strategy consists of a single forecast of the net load. The indirect NLF strategy involves forecasting the load and renewable production separately, and then calculating their difference [3]. ...
Accurate net load forecasting is a cost-effective technique, crucial for the planning, stability, reliability, and integration of variable solar photovoltaic (PV) systems in modern power systems. This work presents a direct short-term net load forecasting (STNLF) methodology for solar-integrated microgrids by leveraging machine learning (ML) principles. The proposed data-driven method comprises of an initial input feature engineering and filtering step, construction of forecasting model using Bayesian neural networks, and an optimization stage. The performance of the proposed model was validated on historical net load data obtained from a university campus solar-powered microgrid. The results demonstrated the effectiveness of the model for providing accurate and robust STNLF. Specifically, the optimally constructed model yielded a normalized root mean square error of 3.98% when benchmarked using a 1-year historical microgrid data. The
k
-fold cross-validation method was then used and proved the stability of the forecasting model. Finally, the obtained ML-based forecasts demonstrated improvements of 17.77% when compared against forecasts of a baseline naïve persistence model. To this end, this work provides insights on how to construct high-performance STNLF models for solar-integrated microgrids. Such insights on the development of accurate STNLF architectures can have positive implications in actual microgrid decision-making by utilities/operators.
... For a power system of just wind power and the load demand. [14] forecasted the net load using Wavelet Neural Network (WNN) trained by Levenberg-Marquardt learning algorithm. The authors took data from two different power systems of different penetration levels. ...
... Hourly historical load, wind, and solar power used as inputs to forecast the net load for the next 24 hours. Furthermore, net load has been forecasted directly and indirectly in [12] for a power system of load demand and wind power. Wavelet Neural Network WNN trained by Levenberg-Marquardt learning algorithm was used to forecast the net load over 6-hours-ahead and in time steps of 30 minutes. ...
Abstract Electricity generated from renewable resources such as wind, solar, geothermal, biomass, ocean waves and tidal is considered sustainable and emissions free. Among these kinds of clean energy resources, wind and solar energy are contributing the most to supplying the load demand all over the world. The stochastic nature of wind and solar power resources injects additional variability and uncertainty to the power system and makes it difficult for the system operator to maintain a continuous balance between the generated and the consumed power especially during high wind and solar power penetration levels. To secure a reliable and economic hybrid power system operation, it is important to provide the system operator with accurate net load forecasts. This research aims to reach optimal net load forecast by forecasting the net load directly using Adaptive Neuro Fuzzy Inference System (ANFIS) depending on the historical net load data.
... Due to the growth of volatile generation and storage systems in the distribution system and the resulting changes of power flows at the hand-over points to the TSO's system, the importance of the substation-related load, generation and storage forecasts is increasing. Researchers have already developed several short-term forecasting models for RE, load and net demand [7][8][9][10][11][12][13][14]. The net demand is the resulting deviation between RE power generation and load. ...
... The net demand is the resulting deviation between RE power generation and load. Distinctions between direct and indirect net demand forecast approaches were made in [7][8][9][10][11]. Direct methods forecast the net demand itself, whereas indirect methods forecast the RE generation and load to subsequently subtract them. ...
... In this way, an indirect forecast achieves greater results and is especially recommended for applications with an increasing share of volatile RE [7,9]. Indirect net demand forecasts, such as those developed in [7][8][9][10], only consider the generation and load behavior and focus on the distribution level. Thus, these models do not consider the storage behavior and are not suitable for TSO requirements. ...
The growth in volatile renewable energy (RE) generation is accompanied by an increasing network load and an increasing demand for storage units. Household storage systems and micro power plants, in particular, represent an uncertainty factor for distribution networks, as well as transmission networks. Due to missing data exchanges, transmission system operators cannot take into account the impact of household storage systems in their network load and generation forecasts. Thus, neglecting the increasing number of household storage systems leads to increasing forecast inaccuracies. To consider the impact of the storage systems on forecasting, this paper presents a new approach to calculate a substation-specific storage forecast, which includes both substation-specific RE generation and load forecasts. For the storage forecast, storage systems and micro power plants are assigned to substations. Based on their aggregated behavior, the impact on the forecasted RE generation and load is determined. The load and generation are forecasted by combining several optimization approaches to minimize the forecasting errors. The concept is validated using data from the German transmission system operator, 50 Hertz Transmission GmbH. This investigation demonstrates the significance of using a battery storage forecast with an integrated load and generation forecast.
... As mentioned before, only few papers have focused on these aspects. And finally, inspired by Shaker et al. [41], we aim to examine the difference in accuracy between a direct and indirect forecasting strategy, where a direct strategy constitutes forecasting net demand, whereas the indirect strategy implies forecasting electricity consumption and PV power generation separately and subtracting these afterwards. The contributions of this study can therefore be summed up as follows: ...
This paper presents a study into the utilization of Gaussian Processes (GPs) for probabilistic forecasting of residential electricity consumption, photovoltaic (PV) power generation and net demand of a single household. The covariance function that encodes prior belief on the general shape of the time series plays a vital role in the performance of GPs and a common choice is the squared exponential (SE), although it has been argued that the SE is likely suboptimal for physical processes. Therefore, we thoroughly test various (combinations of) covariance functions. Furthermore, in order bypass the substantial learning and inference time accompanied with GPs, we investigate the potential of dynamically updating the hyperparameters using a moving training window and assess the consequences on predictive accuracy. We show that the dynamic GP produces sharper prediction intervals (PIs) than the static GP with significant lower computational burden, but at the cost of the ability to capture sharp peaks. In addition, we examine the difference in accuracy between a direct and indirect forecasting strategy in case of net demand forecasting and show that the latter is prone to producing wider PIs with higher coverage probability.
... To optimize the control and guarantee the quality-of-service in the electricity grid, it is important to predict the power flows accurately, as this favors a sensible management of the available flexibility. It has been shown that forecasting accuracy can be improved when the production and consumption in the grid are disaggregated and predicted separately (Monforte et al., 2017;Shaker et al., 2014). The disaggregation of solar generation from the total grid load can be achieved by using on ground irradiance measurements (Sossan et al., 2017). ...
In this paper, we present a method to determine the global horizontal irradiance (GHI) from the power measurements of one or more PV systems, located in the same neighborhood. The method is completely unsupervised and is based on a physical model of a PV plant. The precise assessment of solar irradiance is pivotal for the forecast of the electric power generated by photovoltaic (PV) plants. However, on-ground measurements are expensive and are generally not performed for small and medium-sized PV plants. Satellite-based services represent a valid alternative to on site measurements, but their space-time resolution is limited. Results from two case studies located in Switzerland are presented. The performance of the proposed method at assessing GHI is compared with that of free and commercial satellite services. Our results show that the presented method is generally better than satellite-based services, especially at high temporal resolutions.
... Using signal processing tools, it is shown that direct ND forecast will lead to less volatile time series and so better forecast accuracy than indirect ND prediction. Note that ND has recently been defined in [30] as difference between load demand and wind generation. However, in this research work a more generalized form of ND as difference between load demand and renewable generations is presented and its direct forecast is proposed. ...
... WNN is a more professional forecasting engine, which uses wavelets as the activation functions of the NN. The results of these four comparative methods are quoted from [30]. For the sake of a fair comparison, the same test periods of [30], including the first week of February, May, August, and November, are also considered for the proposed approach. ...
... The results of these four comparative methods are quoted from [30]. For the sake of a fair comparison, the same test periods of [30], including the first week of February, May, August, and November, are also considered for the proposed approach. Additionally, the same error criteria of [30] are used in this numerical experiment, which include normalized mean absolute error (nMAE) and normalized root mean squared error (nRMSE) defined as follows: ...
Integration of renewable generations, such as wind and photovoltaic, into electrical power systems is rapidly growing throughout the world. Stochastic and variable nature of these resources makes some operational challenges to power systems. The most effective way to tackle these challenges is short-term prediction of their available powers. Despite various developed methods to forecast generation of renewable resources, still they have large errors, which may lead to under/over-commitment of conventional generators in power systems. Prediction of net demand (ND), defined as electrical load minus renewable generations, can provide useful information for accurate scheduling of conventional generators. In this article, characteristics of the time series of electric load, renewable generations and ND are analyzed, and a new hybrid prediction strategy is presented for direct prediction of ND. The training mechanism of the proposed forecasting engine is composed of a new stochastic search method and Levenberg–Marquardt learning algorithm based on an iterative procedure and greedy search. The suggested prediction strategy is tested on different real-world power systems and its obtained results are compared with the results of several other forecast methods and published literature figures. These comparisons confirm the validity of the developed forecasting strategy. © 2016 Wiley Periodicals, Inc. Complexity, 2016
... 4) Wavelet Neural Networks (WNN): The WNN [15] is a type of neural network whose activation functions in the hidden layer are Wavelet functions. WNNs have been used in different forecasting applications [15], [23], [33], [34] and have shown promising results. Thus, a WNN with multidimensional Morlet Wavelet as the activation function is explored in this paper for the estimation stage. ...
Roof-top solar photovoltaic systems are normally invisible to system operators, meaning that their generated power is not monitored. If a significant number of systems are installed, invisible solar power could significantly alter the net load in power systems. In this paper, a data-driven methodology is proposed to estimate the power generation of invisible solar power sites by using the measured values from a small number of representative sites. The proposed methodology is composed of a data dimension reduction engine and a mapping function. A number of established methods for reducing the dimension of large-scale data is investigated, and a hybrid method based on k-means clustering and principal component analysis is proposed. The output of this block provides a small subset of sites whose measured data are used in the mapping function. We have implemented several mapping functions to estimate the total generation power of all sites based on the measured output of the selected subset of sites. Numerical results based on data from California's power system are presented.
