Search

Accurately predicting estuaries water quality is essential to support immediate intervention to water quality problem management. Deep learning is used to improve forecasting of water quality parameters in many aquatic systems. However, it is frequently constrained by low data frequency and quality. High-frequency, continuous monitoring using integrated in-situ water quality and environmental sensors can be an input to deep learning models resulting in highly accurate water quality and environmental predictions. This paper proposes a novel approach to improve forecasting of water quality and environmental variables. The results of the real-world data from the Swan Canning Estuary sites show how well the suggested model works. With different sizes of training and testing sets, the model can still predict the increased number of hours in high scores. Eliminating highly correlated variables impact the model's performance, emphasising the usefulness of strongly correlated variables in scarce data scenarios. • This article proposes a framework to enhance water quality variable forecasting. • The framework reframes the time series problem as a supervised learning problem. • The results on the set of data from the estuary show how well the framework works. • The approach is envisaged to support management in detecting water quality issues. [ABSTRACT FROM AUTHOR]

The peak period of an energy-generating wave is one of the most important parameters that describe the spectral shape of the oceanic wave, as this indicates the duration for which the waves prevail with respect to their maximum extractable energy. In this paper, a half-hourly peak wave energy period (T P) forecast model is constructed using a suite of statistically significant lagged inputs based on the partial auto-correlation function with an extreme learning machine model developed and its predictive utility is benchmarked against deep learning models, i.e., convolutional neural network (CNN/CovNet) and recurrent neural network (RNN) models and other traditional M5tree, Conditional Maximization based Multiple Linear Regression (MLR-ECM) and MLR models. The objective model (ELM) vs. the comparison models (CNN, RNN, M5tree, MLR-ECM, and MLR) were trained and validated independently on the test dataset obtained from coastal zones of eastern Australia that have a high potential for implementation of wave energy generation systems. The outcomes ascertain that the ELM model can generate significantly accurate predictions of the half-hourly peak wave energy period, providing a good level of accuracy relative to deep learning models in selected coastal study zones. The study establishes the practical usefulness of the ELM model as being a noteworthy methodology for the applications in renewable and sustainable energy resource management systems. • A real-time peak wave energy period forecasting model is constructed using ELM model. • The advance ELM is benchmarked against a deep learning and other traditional models. • The advance ELM generates better predictions than deep learning models in this region. • ELM model provides significant energy management implications for ocean waves. [ABSTRACT FROM AUTHOR]

Despite the considerable success of deep learning methods in modelling physical processes, they suffer from a variety of issues such as overfitting and lack of interpretability. In hydrology, conceptual rainfall-runoff models are simple yet fast and effective tools to represent the underlying physical processes through lumped storage components. Although conceptual rainfall-runoff models play a vital role in supporting decision-making in water resources management and urban planning, they have limited flexibility to take data into account for the development of robust region-wide models. The combination of deep learning and conceptual models has the potential to address some of the aforementioned limitations. This paper presents a sub-model hybridization of the GR4J rainfall-runoff model with deep learning architectures such as convolutional neural networks (CNN) and long short-term memory (LSTM) networks. The results show that the hybrid models outperform both the base conceptual model as well as the canonical deep neural network architectures in terms of the Nash–Sutcliffe Efficiency (NSE) score across 223 catchments in Australia. We show that our hybrid model provides a significant improvement in predictive performance, particularly in arid catchments, and generalizing better across catchments. • Hybrid modelling features parsimony of data-driven and conceptual hydrologic models. • We present a deep learning-based hybridization for the GR4J rainfall-runoff model. • We present a hierarchical training approach consisting of gradient-based and evolutionary optimization of the respective models. • The results show that hybrid models outperform the canonical (base) hydrologic and deep learning model. • The addition of surface temperature and antecedent streamflow further improves the predictive performance. [ABSTRACT FROM AUTHOR]

Geological and geophysical interpretation is characterised by large and localised datasets that are extremely expensive to acquire. There are clear advantages in applying deep learning techniques to such datasets, but this requires a large amount of suitable data for effective training. Creation of training data can be time-consuming, but novel countermeasures such as Generative Adversarial Networks (GANs) are an enticing alternative with the potential to alleviate this issue by providing synthetic data that is representative of the real data. This paper details an investigation of the potential of GANs for the augmentation of labelled seismic facies from a mass transport complex in training a convolutional neural network (CNN) for facies classification. The study adopts a specific GAN approach, known as the Conditional Style-Based Logo GAN (LoGANv2), because of its capability to generate conditioned output with improved training stability. By using LoGANv2 to synthesise examples that mimic the behaviour of the real data, based on a 3D seismic dataset from the North Carnarvon Basin in Australia, the accuracy of the traditional CNN for facies classification improved from 38% (the benchmark result where no augmentation was applied) to 52%, 61%, 94% and 74% for four trials with different degrees of augmentation. The results show that increasing training size, either through manual annotation (trial 2) or standard manipulations such as image rotation (trial 3 and trial 4), improves the performance of LoGANv2 and hence the CNN classification. The experiments indicate that although standard manipulation can increase the diversity of the training dataset, the balance between diversity and consistency within the datasets is also important, and this should be optimized for LoGANv2 to achieve better performance. In addition, we explored the adaptability of the proposed LoGANv2-CNN approach to unlabelled samples from another survey to demonstrate the robustness and flexibility of the model. • A LoGANv2-assisted CNN model was developed for augmentation and classification of seismic facies. • With LoGANv2 proving its capacity to augment different types of seismic facies with minimal manual labelling, the synthetic facies generated and the real ones are passed down to a CNN for further classification. • The streamlined process demonstrates that, with little manual annotation, the generation of synthetic seismic facies could lead to a better understanding of the underlying geological structures. [ABSTRACT FROM AUTHOR]

With the ever-increasing level of bidding freedom bestowed to participants in deregulated electricity markets, bidding strategies have become more diversified and complicated, inevitably giving rise to the growth of market uncertainties. Some researchers have developed tools to predict the bidding behaviors of generation companies (GENCOs) considering uncertainties. However, there still remains a gap in enhancing the performance of probabilistic Bidding Behavior Forecasting (BBF) and understanding the sources of bidding uncertainties in electricity markets. This paper proposes a holistic Bayesian Deep Learning (BDL) framework based on Accurate Variational Inference (AVI) to capture both aleatoric uncertainty and epistemic uncertainty, two important uncertainties in bidding behaviors. The introduced framework also procures higher BBF accuracy and reasonable computation cost compared with existing techniques. Derivatives of GENCOs' bidding uncertainty are ascertained by implementing a new metric to quantify the impact of influence factors. A numerical experiment is conducted using data from National Electricity Market (NEM) in Australia to demonstrate the performance of the proposed framework. • A BDL framework based on AVI is proposed for analyzing bidding uncertainty. • A modified BNN is constructed to perform BBF with higher efficiency. • The epistemic uncertainty and the aleatoric uncertainty in BBF are captured. • The derivatives of bidding uncertainty are ascertained. [ABSTRACT FROM AUTHOR]

Local and state governments depend on small area population forecasts to make important decisions concerning the development of local infrastructure and services. Despite their importance, current methods often produce highly inaccurate forecasts. Recent years have witnessed promising developments in time series forecasting using Machine Learning across a wide range of social and economic variables. However, limited work has been undertaken to investigate the potential application of Machine Learning methods in demography, particularly for small area population forecasting. In this paper we describe the development of two Long-Short Term Memory network architectures for small area populations. We employ the Keras Tuner to select layer unit numbers, vary the window width of input data, and apply a double training and validation regime which supports work with short time series and prioritises later sequence values for forecasts. These methods are transferable and can be applied to other data sets. Retrospective small area population forecasts for Australia were created for the periods 2006–16 and 2011–16. Model performance was evaluated against actual data and two benchmark methods (LIN/EXP and CSP-VSG). We also evaluated the impact of constraining small area population forecasts to an independent national forecast. Forecast accuracy was influenced by jump-off year, constraining, area size, and remoteness. The LIN/EXP model was the best performing method for the 2011-based forecasts whilst deep learning methods performed best for the 2006-based forecasts, including significant improvements in the accuracy of 10 year forecasts. However, benchmark methods were consistently more accurate for more remote areas and for those with populations <5000. • A novel deep learning method for small area forecasts is presented. • Forecast accuracy was influenced by jump-off year, constraining, area size, and remoteness. • Deep learning methods performed best for the 5 and 10 year 2006-based forecasts. • The LIN/EXP model was the best performing method for the 5 year 2011-based forecasts. • Benchmark methods were consistently more accurate for more remote areas and for those with populations <5000. [ABSTRACT FROM AUTHOR]

Predicting electricity demand data is considered an essential task in decisions taking, and establishing new infrastructure in the power generation network. To deliver a high-quality electricity demand prediction, this paper proposes a hybrid combination technique, based on a deep learning model of Convolutional Neural Networks and Echo State Networks, named as CESN. Daily electricity demand data from four sites (Roderick, Rocklea, Hemmant and Carpendale), located in Southeast Queensland, Australia, have been used to develop the proposed hybrid prediction model. The study also analyzes five other machine learning-based models (support vector regression, multilayer perceptron, extreme gradient boosting, deep neural network, and Light Gradient Boosting) to compare and evaluate the outcomes of the proposed deep learning approach. The results obtained in the experimental study showed that the proposed hybrid deep learning model is able to obtain the highest performance compared to other existing models developed for daily electricity demand data forecasting. Based on the statistical approaches utilized in this study, the proposed hybrid approach presents the highest prediction accuracy among the compared models. The obtained results showed that the proposed hybrid deep learning algorithm is an excellent and accurate electricity demand forecasting method, which outperformed the state of the art algorithms that are currently used in this problem. [Display omitted] • A Hybrid deep learning model is proposed for daily electricity demand prediction. • Convolutional Neural Networks and Echo State Networks are integrated into a single forecasting algorithm. • The proposed deep learning-based approach is excellent in electricity demand prediction. • The proposed algorithm has practical use in inter-daily electricity demand prediction. [ABSTRACT FROM AUTHOR]

Photovoltaic (PV) power forecasting (PVPF) plays an important role in the scheduling and operation of modern power systems. Considering the highly varying and complex features of PVPF, this paper constitutes a novel hybrid deep learning forecasting method. A similar day selection method based on the levy-flight beetle antennae search algorithm is proposed to select historical days similar to the forecast day from the real-time massive data. The Pearson correlation coefficient method is utilized to select the main meteorological factors while wavelet packet decomposition is used to decompose and reconstruct the original PV power into a series of sub-signals. A deep learning model taking the sub-signals of similar days as network inputs is established with a group of gated recurrent units (GRU), where the hyperparameters of each GRU network are effectively optimized. The forecasting result of each sub-signal is integrated to obtain the final forecasting PV power value. The simulation regarding the 5-min ahead PVPF is carried out based on a real-world dataset from Alice Spring, Australia. The simulation comparisons indicate that the proposed hybrid deep learning method outperforms other competitive PVPF methods in terms of both forecasting accuracy and computational efficiency. • Application of SDS-based GRU model to PV power forecasting. • A similar day selection method is developed based on the Levy-flight BAS algorithm. • Application of WPD to decompose the PV signal to extract the frequency feature. • Levy-flight BAS is used to optimize the hyperparameters of GRU. [ABSTRACT FROM AUTHOR]

Modern power systems are experiencing the challenge of high uncertainty with the increasing penetration of renewable energy resources and the electrification of heating systems. In this paradigm shift, understanding electricity users' demand is of utmost value to the retailers, aggregators, and policymakers. However, behind-the-meter (BTM) equipment and appliances at the household level are unknown to the other stakeholders mainly due to privacy concerns and tight regulations. In this paper, we seek to identify residential consumers based on their BTM equipment, mainly rooftop photovoltaic (PV) systems and electric heating, using imported/purchased energy data from utility meters. To solve this problem with an interpretable, fast, secure, and maintainable solution, we propose an integrated method called Interpretable Refined Motifs And binary Classification (IRMAC). The proposed method comprises a novel shape-based pattern extraction technique, called Refined Motif (RM) discovery, and a single-neuron classifier. The first part extracts a sub-pattern from the long time series considering the frequency of occurrences, average dissimilarity, and time dynamics while emphasising specific times with annotated distances. The second part identifies users' types with linear complexity while preserving the transparency of the algorithms. With the real data from Australia and Denmark, the proposed method is tested and verified in identifying PV owners and electrical heating system users. The performance of the IRMAC is studied and compared with various state-of-the-art methods. The proposed method reached an accuracy of 96% in identifying rooftop PV users and 94.4% in identifying electric heating users, which is comparable to the best solution based on deep learning, while the speed of the inference model is a thousand times faster. Last but not least, the proposed method is a transparent algorithm, which can tackle the concerns regarding the agnostic decision-making process when policies prohibit some machine learning methods. [ABSTRACT FROM AUTHOR]

Particulate matter (PM) or particle pollution include the tiny particles of dust and fly ash particles are expelled from coal-burning power plants. Coal combustion is an extremely prevalent source of air pollution, and resulting PM has substantial impacts on human health, especially in industrial zones. This paper aims to design a novel hybrid deep learning framework based on long short-term memory (LSTM) integrated with a two-stage data filtering technique to accurately predict the air quality indices (i.e., PM 2.5 and PM 10) in a chosen study region 'Miles Airport, Queensland,' that meets the needs of the coal seam gas industry in Australia. The data used to construct the novel hybrid two-stage data filtering technique based on LSTM comprising of six meteorological parameters (i.e., wind direction, wind speed, air temperature, relative humidity, solar radiation, and rainfall) and two environmental factors (i.e., ozone and total suspended particulate). In the first stage, two robust feature selection methods, namely, extreme gradient boosting (XGBoost) and the classification and regression tree (CART) approach, were adopted to explore the most significant predictors. Then in the second stage, the best subset regression (BSR) technique is utilized to determine the best subsets input combinations i.e., C1, C2, and C3 based on several particular metrics. The three BSR-based input combinations were employed in the LSTM model to estimate the PM 2.5 and PM 10. Furthermore, to validate the main hybrid framework, two advanced machine learning (ML) methods (i.e., LightGBM and ridge kernel regression (KRR)) and two traditional ML methods (i.e., Adaptive neuro-fuzzy inference system (ANFIS) and multilayer perceptron neural network (MLP)) were hybridized with the multi-level data filtering strategy by examining the optimal input combinations. Several statistical metrics, graphical tools, and diagnostic analyses evaluated the hybrid models. The outcomes of the PM 2.5 simulation based on 2375 data samples showed that the LSTM-C3 containing all the selected predictors yielded the most promising accuracy, followed by the LightGBM-C3 and MLP-C3 models. On the other hand, the simulation of PM 10 distribution demonstrated that the LSTM-C3 was superior to other models, followed by the KRR-C3 and LightGBM-C3 models. [ABSTRACT FROM AUTHOR]

Modelling workflow for BRF-LSTM and BRF-GRU predictive models. [Display omitted] • Deep learning hybrid predictive model (BRF-LSTM) is built for streamflow forecasting. • Climate indices, rainfall and periodicity are factored for accurate predictions. • Adopts feature algorithm (i.e. Boruta) with long short-term memory and gated recurrent unit. • Six gauging sites tested; BRF-LSTM yields > 98% errors in ± 0.015 m, ~1.30% relative error. • The proposed model is useful for hydrological and strategic water resources planning. Long-term forecasting of any hydrologic phenomena is essential for strategic environmental planning, hydrologic and other forms of structural design, agriculture, and water resources management. Climate mode indices, utilising machine learning methods, are frequently considered as predictor variables in order to forecast several different hydrological variables. In this study, a feature selection algorithm based on two different deep learning models, i.e., long short-term memory and a gated recurrent unit, is applied to improve the forecasting capability of streamflow water levels at six gauging stations in the Murray Darling Basin of Australia. This paper therefore aggregates the significant antecedent lag memory of climate mode indices, rainfall, and the monthly factor based on the periodicity as the predictor variables to attain significantly accurate stream water level forecasts. This novel method identifies an improved relationship between the stream water level and climate mode indices through the aggregation of the significant lagged datasets capturing the historical features to predict the future streamflow water level. The boruta feature selection algorithm (BRF) was then applied in a two phase process before and after attaining the significant lagged inputs to screen the optimum predictor variables. The merits of the forecast models were evaluated through different performance evaluation criteria. The results show that the accumulated significant lagged inputs based on climate mode indices, along with the rainfall and periodicity factors are seen to provide improved forecasting of the SWL over the non-BRF deep learning approaches where no prior feature selection was applied. The hybrid LSTM method (i.e., BRF-LSTM model) achieved a unique advantage in terms of SWL forecasting, particularly attaining over 98% of the predictive errors lying within a band of +/-0.015 m with relatively low relative errors (RRMSE ≈1.30% and RMAE ≈ 0.882%), outperforming all of the benchmark models. It is also found that the periodicity factor has a potential influence on the accuracy of the forecast models for the four monitored study stations. This study concludes that the newly developed hybrid deep learning approaches, coupled with the BRF feature selection, provide improved forecasting performance. The hybrid approach developed in this paper can therefore be used to provide a strong provide predictive response algorithm for the hydrological variables that were influenced by the low-frequency variability of the climate model indices in respect to streamflow water level. [ABSTRACT FROM AUTHOR]

• The LSTM method is proposed to construct the LUBE approach-based prediction interval for wind power forecast. • An innovative fuzzy approach is suggested to adjust the prediction interval coverage probability. • CDOA is applied to achieve the most optimal prediction intervals. This paper proposes a novel deep learning method based on lower-upper-bound-estimation (LUBE) and long short-term memory (LSTM) models to capture the uncertainty effects in the power generation of wind turbines. The LUBE model deploys the prediction interval concept as well as the LSTM approach to make a robust and resilient model. The LSTM approach is applied to construct the optimal prediction intervals with an appropriate upper bound and lower bound with a certain confidence level. In addition, the fuzzy set theory is proposed in the model to let adjust LSTM parameters based on the decision maker's ideas. This approach empowers operators to satisfy both coverage probability and width indicators of prediction intervals to achieve the optimal solution. The collective decision optimization algorithm is introduced to provide more flexibility in tuning the LSTM parameters. The efficiency and quality of the proposed scheme are studied using some datasets gathered from the Australia wind farms. [Display omitted] [ABSTRACT FROM AUTHOR]

This paper presents a new hybrid approach for Global Solar Radiation (GSR) prediction problems, based on deep learning approaches. Predictive models are useful ploys in solar energy industries to optimize the performance of photovoltaic power systems. Specifically, in this work we develop a new 4-phase hybrid CXGBRFR framework, which includes a deep learning Convolutional Neural Network, an Extreme Gradient Boosting with Random Forest Regression, and a Harris Hawks Optimization for initial feature selection. The proposed system has been tested in a problem of daily GSR prediction at six solar farms in Australia. Data from three global climate models (GCM) (CSIRO-BOM ACCESS1-0, MOHC Hadley-GEM2-CC and MRI MRI-CGCM3) have been considered as predictive (input) variables for the proposed approach. The variables from these GCMs contain enough information to obtain an accurate prediction of the GSR at each solar farm. The performance of the proposed approach is compared against different deep and shallows learning approaches: Deep Belief Network, Deep Neural Network, Artificial Neural Network, Extreme Learning Machine and Multivariate Auto-Regressive Spline models. We show that the proposed approach exhibits an excellent performance in GSR prediction, against all alternative approaches in all solar farms considered. [Display omitted] • We develop hybrid deep learning model to predict solar radiation • We optimize the algorithm with Harris Hawks, gradient boosting and random forest • We adopt global climate model and observed predictors as learning model inputs • We build an expert system with relevant knowledge in context of solar application • Our methods benefit solar photovoltaic power monitoring in solar engineering problems. [ABSTRACT FROM AUTHOR]

More timely and accurate air quality forecasting could contribute to better public health protection and air pollution prevention. Particulates are a significant indicator for measuring the degree of air pollution. This paper reports on research to model an early warning tool for coarse particulates when assessing the impact of the 12 satellite-derived and ground-based meteorological pollutants out of 30 pollutants considered using hourly Australian data from January 2018–December 2020. A one-dimensional convolutional neural network (CNN) was integrated with a one-directional fully gated recurrent unit (GRU) to forecast consecutive hours' air quality. The CNN model acts as a spatial feature extractor, whereas the new generation GRU makes it computationally efficient. The resultant hybrid ' CNN-GRU' is then comprehensively benchmarked outperforming an ensemble of six other deep learning models. The proposed model's efficacy is indicated at the four most air polluted Australian postcodes in the testing phase. A detailed error analysis with visual and statistical metrics for air quality forecasting ascertains the proposed model's countermeasure to reduce harm and loss. The practical tool is immensely beneficial and can be widely deployed to the regions of public health concern where air pollution is a significant hazard. [Display omitted] • A novel early warning hybrid PM10 forecasting framework designed for Australia. • 30 satellite and ground-based meteorological pollutants were considered with 7 modeling approaches. • CNN integrated with GRU outperformed the other models. • Early warning AI model show potential in the health, environment sector, and decision-making. [ABSTRACT FROM AUTHOR]

This study proposes a new hybrid deep learning (DL) model, the called CSVR, for Global Solar Radiation (GSR) predictions by integrating Convolutional Neural Network (CNN) with Support Vector Regression (SVR) approach. First, the CNN algorithm is used to extract local patterns as well as common features that occur recurrently in time series data at different intervals. Then, the SVR is subsequently adopted to replace the fully connected CNN layers to predict the daily GSR time series data at six solar farms in Queensland, Australia. To develop the hybrid CSVR model, we adopt the most pertinent meteorological variables from Global Climate Model and Scientific Information for Landowners database. From a pool of Global Climate Models variables and ground-based observations, the optimal features are selected through a metaheuristic Feature Selection algorithm, an Atom Search Optimization method. The hyperparameters of the proposed CSVR are optimized by mean of the HyperOpt method, and the overall performance of the objective algorithm is benchmarked against eight alternative DL methods, and some of the other Machine Learning approaches (LSTM, DBN, RBF, BRF, MARS, WKNNR, GPML and M5TREE) methods. The results obtained shows that the proposed CSVR model can offer several predictive advantages over the alternative DL models, as well as the conventional ML models. Specifically, we note that the CSVR model recorded a root mean square error/mean absolute error ranging between ≈ 2.172–3.305 MJ m2/1.624–2.370 MJ m2 over the six tested solar farms compared to ≈ 2.514–3.879 MJ m2/1.939–2.866 MJ m2 from alternative ML and DL algorithms. Consistent with this predicted error, the correlation between the measured and the predicted GSR , including the Willmott's, Nash-Sutcliffe's coefficient and Legates & McCabe's Index was relatively higher for the proposed CSVR model compared to other DL and Machine Learning methods for all of the study sites. Accordingly, this study advocates the merits of CSVR model to provide a viable alternative to accurately predict GSR for renewable energy exploitation, energy demand or other forecasting-based applications. [Display omitted] • This paper presents a hybrid CSVR algorithm for global solar radiation prediction. • The proposed approach combines Convolutional Neural Network with a Support Vector Machine. • An Atom Search-based feature selection is applied to improve the performance of the system. • Experiments are carried out in data at six solar energy farms in Queensland, Australia. • The results show the efficacy of deep learning CSVR for solar energy prediction problems. [ABSTRACT FROM AUTHOR]

• A novel CTSGAN deep learning framework is first proposed and applied to model day-ahead electricity price scenarios. • The CTSGAN point forecasting model can eventually be transformed into a probabilistic forecasting model by enhancing the diversity of random input. • The CTSGAN probabilistic forecasting model can directly yield high-quality forecasting intervals with different coverage probabilities as a multi-objective forecasting model. • PSO-based optimal conditions selection method can increase the forecasting accuracy. Electricity prices in spot markets are volatile and can be affected by various factors, such as generation and demand, system contingencies, local weather patterns, bidding strategies of market participants, and uncertain renewable energy outputs. Because of these factors, electricity price forecasting is challenging. This paper proposes a scenario modeling approach to improve forecasting accuracy, conditioning time series generative adversarial networks on external factors. After data pre-processing and condition selection, a conditional TSGAN or CTSGAN is designed to forecast electricity prices. Wasserstein Distance, weights limitation, and RMSProp optimizer are used to ensure that the CTGAN training process is stable. By changing the dimensionality of random noise input, the point forecasting model can be transformed into a probabilistic forecasting model. For electricity price point forecasting, the proposed CTSGAN model has better accuracy and has better generalization ability than the TSGAN and other deep learning methods. For probabilistic forecasting, the proposed CTSGAN model can significantly improve the continuously ranked probability score and Winkler score. The effectiveness and superiority of the proposed CTSGAN forecasting model are verified by case studies. [ABSTRACT FROM AUTHOR]

Accurate forecasting of wind power generation plays a key role in improving the operation and management of a power system network and thereby its reliability and security. However, predicting wind power is complex due to the existence of high non-linearity in wind speed that eventually relies on prevailing weather conditions. In this paper, a novel hybrid deep learning model is proposed to improve the prediction accuracy of very short-term wind power generation for the Bodangora wind farm located in New South Wales, Australia. The hybrid model consists of convolutional layers, gated recurrent unit (GRU) layers and a fully connected neural network. The convolutional layers have the ability to automatically learn complex features from raw data while the GRU layers are capable of directly learning multiple parallel sequences of input data. The data sets of 5-min intervals from the wind farm are used in case studies to demonstrate the effectiveness of the proposed model against other advanced existing models, including long short-term memory, GRU, autoregressive integrated moving average and support vector machine, which are tuned to optimise outcome. To further evaluate the efficacy of the proposed model, another data set taken from the Capital wind farm in Australia is used. It is observed that the hybrid deep learning model exhibits superior performance in both the data sets over other forecasting models to improve the accuracy of wind power forecasting, numerically for the Bodangora wind farm, up to 1.59 per cent in mean absolute error, 3.73 per cent in root mean square error and 8.13 per cent in mean absolute percentage error. [ABSTRACT FROM AUTHOR]