Next Article in Journal
Improvement of Ecological Footprint Model in National Nature Reserve Based on Net Primary Production (NPP)
Next Article in Special Issue
Technological Innovation in Biomass Energy for the Sustainable Growth of Textile Industry
Previous Article in Journal
Use of Shared-Mobility Services to Accomplish Emergency Evacuation in Urban Areas via Reduction in Intermediate Trips—Case Study in Xi’an, China
Previous Article in Special Issue
The Nexus Concept Integrating Energy and Resource Efficiency for Policy Assessments: A Comparative Approach from Three Cases
Open AccessTechnical Note

Day-Ahead Forecasting of Hourly Photovoltaic Power Based on Robust Multilayer Perception

by Chao Huang 1,2,3, Longpeng Cao 1,2,3, Nanxin Peng 4, Sijia Li 5, Jing Zhang 1,2,3, Long Wang 1,2,3,*, Xiong Luo 1,3 and Jenq-Haur Wang 6
1
School of Computer and Communication Engineering, University of Science and Technology Beijing (USTB), Beijing 100083, China
2
Key Laboratory of Wind Energy and Solar Energy Technology (Inner Mongolia University of Technology), Ministry of Education, Hohhot 010051, China
3
Beijing Key Laboratory of Knowledge Engineering for Materials Science, Beijing 100083, China
4
School of International Business, Southwestern University of Finance and Economics, Chengdu 611130, China
5
National Internet Finance Association of China, Beijing 100080, China
6
Department of Computer Science and Information Engineering, National Taipei University of Technology, Taipei City 106, Taiwan
*
Author to whom correspondence should be addressed.
Sustainability 2018, 10(12), 4863; https://doi.org/10.3390/su10124863
Received: 13 November 2018 / Revised: 11 December 2018 / Accepted: 14 December 2018 / Published: 19 December 2018
(This article belongs to the Special Issue Sustainable Energy Systems: From Primary to End-Use)

Abstract

Photovoltaic (PV) modules convert renewable and sustainable solar energy into electricity. However, the uncertainty of PV power production brings challenges for the grid operation. To facilitate the management and scheduling of PV power plants, forecasting is an essential technique. In this paper, a robust multilayer perception (MLP) neural network was developed for day-ahead forecasting of hourly PV power. A generic MLP is usually trained by minimizing the mean squared loss. The mean squared error is sensitive to a few particularly large errors that can lead to a poor estimator. To tackle the problem, the pseudo-Huber loss function, which combines the best properties of squared loss and absolute loss, was adopted in this paper. The effectiveness and efficiency of the proposed method was verified by benchmarking against a generic MLP network with real PV data. Numerical experiments illustrated that the proposed method performed better than the generic MLP network in terms of root mean squared error (RMSE) and mean absolute error (MAE).
Keywords: forecasting; multilayer perception; photovoltaic; sustainable energy; pseudo-Huber loss forecasting; multilayer perception; photovoltaic; sustainable energy; pseudo-Huber loss

1. Introduction

Solar energy is considered to be one of the most renewable and sustainable energy resources. Photovoltaics (PV), which convert solar energy into electricity, is the most widely used technique to make use of solar energy. The increasing penetration of PV power, however, brings challenges for the planning and scheduling of the power grid due to the uncertainty of PV power production [1]. Forecasting is an essential technique to alleviate the negative impacts on the grid operation [2] and to facilitate the management of grids including renewable energies [3,4,5].
PV power forecasting methods can be categorized into direct methods and indirect methods. The former produces PV power production as model outputs; while the latter first generates forecasts of solar irradiance, then PV performance models are applied to derive the PV power production based on solar irradiance [6,7]. The presented modeling approaches include numerical weather prediction (NWP) models, statistical models, and artificial intelligence. NWP-based approaches dynamically model the atmospheric states and project their impacts on solar irradiance and PV power production [8]. These approaches are computationally expensive, making them difficult to use for short-term forecasting. In the early stages of PV power forecasting, statistical models such as autoregressive moving average (ARMA) [9] and its variants [10] were frequently used. However, these models are linear, which cannot capture the nonlinear characteristics in PV power production. Recently, artificial intelligence including artificial neural network (ANN) [11,12] and its variants such as extreme learning machine [13], support vector regression (SVR) [14], and Gaussian process regression [15] have been widely applied to renewable energy forecasting. In [16], a wavelet recurrent neural network (WRNN) was proposed for the prediction of energy production in a PV park. As for the day-ahead forecasting of hourly solar irradiance and PV power, ANN topped the methodologies for multi-input multi-output forecasting [17,18,19].
In the aforementioned studies on ANN based forecasting of PV power production, the learnable parameters such as weights and biases of an ANN were usually derived by minimizing the mean squared error on the training dataset. However, the mean squared error can be dominated by a few particularly large errors due to the sudden change in weather patterns that are difficult to predict. This can lead to poor asymptotic relative efficiency of the mean squared error based estimator in terms of estimation theory.
This paper contributes to developing a robust ANN model for day-ahead hourly forecasting of PV power based on a robust loss function, the pseudo-Huber loss, which combines the best properties of squared loss and absolute loss. The pseudo-Huber loss is less sensitive to large errors to train a more robust ANN model. The efficacy of the proposed method was validated on real PV power data by benchmarking against the generic ANN trained on the squared loss and the persistence model. Numerical experimental results showed that the proposed method outperformed the generic ANN model and persistence model in terms of root mean squared error and mean absolute error.

2. Methodology

In this section, the fundamentals of ANN will first be introduced, followed by the training of neural network based on the pseudo-Huber loss function.

2.1. Multilayer Perception Network

The proposed ANN model is based on a typical network structure, the multilayer perception (MLP), as shown in Figure 1. The MLP consists of the input layer, hidden layers, and output layer, and the model can be mathematically expressed as Equations (1) and (2):
H = f h ( ω h X T + b h )
Y = f y ( ω y H T + b y )
In Equations (1) and (2), X = (x1, x2, …, xM) and Y = (y1, y2, …, yQ) denote the M-dimensional and Q-dimensional model input and output, respectively; ω and b represent the weights and biases of the network, respectively; f denotes the activation function; and the subscripts h and y stand for the hidden layer and output layer, respectively.
The learnable parameters including ω and b are usually obtained by minimizing the mean squared error with optimizing algorithms such as the Adam optimization [21]. The mean squared error is sensitive to a few particularly large errors and can lead to a poor estimator. In this paper, the pseudo-Huber loss was applied to the training of the MLP.

2.2. Pseudo-Huber Loss

The pseudo-Huber loss function is defined in Equation (3) [22]:
L δ ( e ) = δ 2 ( 1 + ( e / δ ) 2 1 )
In Equation (3), δ is a controlling parameter. The pseudo-Huber loss function combines the best properties of squared loss and absolute loss that with small errors e, L δ ( e ) approximates e2/2, which is strongly convex, and with extremely large e, L δ ( e ) approximates a straight line with a slope of δ, which is less steep than the squared loss. This property of the pseudo-Huber loss makes it less sensitive to large errors. This paper took advantage of this property of the pseudo-Huber loss to train the MLP as the sudden change in weather patterns can result in large modeling errors. The objective function based on the pseudo-Huber loss for the training of the MLP is expressed in Equation (4):
L = i = 1 N q = 1 Q δ 2 ( 1 + ( e i , q / δ ) 2 1 )
In Equation (4), e i , q = y ^ i , q y i , q for i = 1, …, N and q = 1, …, Q where i denotes the number of training data point and q indexes the output element, respectively, and y and y ^ denote the observed and modeling PV power, respectively. To improve the forecasting performance, swarm intelligence-based optimization algorithms such as the Jaya algorithm [23,24,25,26] can be adopted to optimize δ.

3. Case Study

3.1. Data Description

The hourly PV power production between 1 January 2012 and 31 December 2017 of a PV plant installed at the Andre Agassi Preparatory Academy Building B (36.19N, 115.16W, elevation of 620 m) in the USA was used to verify the efficacy of the proposed MLP for day-ahead hourly PV power forecasting. The specifications of the PV power plant is shown in Table 1, and the data are available at https://maps.nrel.gov/pvdaq/. The histogram on the hourly power production is shown in Figure 2 where it can be observed that the hourly PV power production showed high variability, which induces great challenges for the forecasting model.
The hourly PV power produced over 2017 is illustrated in Figure 3 (data on some days were not available). It can be seen that the hourly PV power production showed strong seasonality and diurnal cycle. In the case study, day time data from 6:00 am to 19:00 pm including 14 h a day were considered. The input of the MLP model was composed of the PV power production of the last seven days for the hourly forecasting of the upcoming day. Hence, the dimensions of input vector and output vector were M = 7 × 14 and Q = 14, respectively. The total data were divided into two groups, the training dataset consisted of observations from 1 January 2012 to 31 December 2016 and the test dataset included data over the whole year of 2017. The data were not always available for the PV plant, so the real training and test data sizes were 1742 and 309, respectively. To simplify the training of the MLP model, all data were normalized to [0, 1] by dividing the nominal power. The operation of normalization also facilitated the comparison of forecasting performance between the different datasets.

3.2. Performance Metrics

To compare the forecasting performance by different methods, root mean squared error (RMSE) and mean absolute error (MAE) as defined in Equations (5) and (6) were employed.
R M S E = 1 K k = 1 K ( p ^ k p k )
M A E = 1 K k = 1 K | p ^ k p k |
In Equations (5) and (6), p ^ k and pk for k = 1, …, K are the normalized forecasts and observations of PV power, respectively.

3.3. Numerical Results and Analysis

The forecasting performance of the proposed method based on robust loss function was compared with the generic MLP network trained by minimizing the mean squared error. The persistence model, p ^ d , h = p d 1 , h where the power production at hour h on the targeted day d was assumed to be equal to the observation of power production at hour h on the day indexed by d − 1, was also considered as a benchmark. The RMSE and MAE computed based on the whole test dataset for all of the forecasting methods are provided in Table 2. It can be observed that the proposed method performed better than the generic MLP model and the persistence model in terms of both RMSE and MAE.
To provide insight into the forecasting performance for each hour, Table 3 gives the RMSE and MAE indexed by the hour. From Table 3, the proposed forecasting method outperformed the generic MLP model for each hour with a lower RMSE and MAE. However, the RMSE and MAE were greater in the middle of the day. At early and late hours (6:00 am, 18:00 pm, and 19:00 pm) in the day, the performance of the persistence model was better than the proposed method and the generic MLP. This is because PV power production at early and late hours are usually very small (in winter the power production at these hours can be zero) and there is little difference each day, which fits the persistence model well.
The daily forecasting performance was also investigated as shown in Table 4. The daily forecasting of power production was computed by summing the forecasts of each hour in a day. It was also observable that the proposed method beat the generic MLP-based forecasting method and the persistence model for daily performance.
The forecasting of PV power production by the proposed method compared with the observations and forecasts by the generic MLP model on consecutive days is depicted in Figure 4. It can be observed that the proposed method well captured the evolution of hourly PV power production and the forecasting errors by the proposed method were generally smaller than the generic MLP network.

4. Conclusions

In this paper, a robust MLP network was proposed for day-ahead hourly forecasting of PV power production. The generic MLP network is usually trained by minimizing the mean squared error; however, the mean squared error loss is sensitive to a few particularly large errors, which can result in an estimator with poor asymptotic relative efficiency. To tackle this problem, this paper adopted the pseudo-Huber loss, which combines the best properties of squared loss and absolute loss.
The efficacy of the proposed method was validated by testing on real PV data. The numerical experiments demonstrated that the proposed method outperformed the generic MLP network in terms of both RMSE and MAE for various time scales. This study suggests that the pseudo-Huber loss can be employed to optimize artificial intelligence-based approaches in handling solar data.

Author Contributions

Conceptualization and writing, C.H.; methodology, L.W.; software, L.C. and J.Z.; validation, N.P. and S.L.; funding acquisition, J.-H.W. and X.L.

Funding

This research was supported in part by the Foundation of Key Laboratory of Wind Energy and Solar Energy Technology, Ministry of Education under Grant 2018ZD02, in part by the Fundamental Research Funds for the Central Universities under Grants 06500103 and 06500078, in part by the University of Science and Technology Beijing–National Taipei University of Technology Joint Research Program under Grant TW2018008, in part by the National Natural Science Foundation of China under Grant 71473155, and in part by the National Key Research and Development Program of China under Grant 2018YFC0808306.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Huang, C.; Wang, L.; Lai, L. Data-driven short-term solar irradiance forecasting based on information of neighboring sites. IEEE Trans. Ind. Electron. 2018. [Google Scholar] [CrossRef]
  2. Sobri, S.; Koohi-Kamali, S.; Rahim, N.A. Solar photovoltaic generation forecasting methods: A review. Energy Convers. Manag. 2018, 156, 459–497. [Google Scholar] [CrossRef]
  3. Zhang, Y.; Gatsis, N.; Giannakis, G.B. Robust Energy Management for Microgrids With High-Penetration Renewables. IEEE Trans. Sustain. Energy 2013, 4, 944–953. [Google Scholar] [CrossRef][Green Version]
  4. Hosseinzadeh, M.; Salmasi, F.R. Robust Optimal Power Management System for a Hybrid AC/DC Micro-Grid. IEEE Trans. Sustain. Energy 2015, 6, 675–687. [Google Scholar] [CrossRef]
  5. Sardou, I.G.; Zare, M.; Azad-Farsani, E. Robust energy management of a microgrid with photovoltaic inverters in VAR compensation mode. Int. J. Electr. Power Energy Syst. 2018, 98, 118–132. [Google Scholar] [CrossRef]
  6. Wang, L.; Huang, C. A novel Elite Opposition-based Jaya algorithm for parameter estimation of photovoltaic cell models. Optik 2018, 155, 351–356. [Google Scholar] [CrossRef]
  7. Huang, C.; Wang, L. Simulation study on the degradation process of photovoltaic modules. Energy Convers. Manag. 2018, 165, 236–243. [Google Scholar] [CrossRef]
  8. Andrade, J.R.; Bessa, R.J. Improving Renewable Energy Forecasting With a Grid of Numerical Weather Predictions. IEEE Trans. Sustain. Energy 2017, 8, 1571–1580. [Google Scholar] [CrossRef]
  9. Wu, J.; Chan, C.K. Prediction of hourly solar radiation using a novel hybrid model of ARMA and TDNN. Sol. Energy 2011, 85, 808–817. [Google Scholar]
  10. Li, Y.T.; Su, Y.; Shu, L.J. An ARMAX model for forecasting the power output of a grid connected photovoltaic system. Renew. Energy 2014, 66, 78–89. [Google Scholar] [CrossRef]
  11. Leva, S.; Dolara, A.; Grimaccia, F.; Mussetta, M.; Ogliari, E. Analysis and validation of 24 hours ahead neural network forecasting of photovoltaic output power. Math. Comput. Simul. 2017, 131, 88–100. [Google Scholar] [CrossRef][Green Version]
  12. Capizzi, G.; Napoli, C.; Bonanno, F. Innovative Second-Generation Wavelets Construction With Recurrent Neural Networks for Solar Radiation Forecasting. IEEE Trans. Neural Netw. Learn. Syst. 2012, 23, 1805–1815. [Google Scholar] [CrossRef] [PubMed]
  13. Luo, X.; Sun, J.; Wang, L.; Wang, W.; Zhao, W.; Wu, J.; Wang, J.-H.; Zhang, Z. Short-Term Wind Speed Forecasting via Stacked Extreme Learning Machine With Generalized Correntropy. IEEE Trans. Ind. Inform. 2018, 14, 4963–4971. [Google Scholar] [CrossRef]
  14. Zendehboudi, A.; Baseer, M.A.; Saidur, R. Application of support vector machine models for forecasting solar and wind energy resources: A review. J. Clean. Prod. 2018, 199, 272–285. [Google Scholar] [CrossRef]
  15. Huang, C.; Zhang, Z.J.; Bensoussan, A. Forecasting of daily global solar radiation using wavelet transform-coupled Gaussian process regression: Case study in Spain. In Proceedings of the 2016 IEEE Innovative Smart Grid Technologies-Asia (ISGT-Asia), Melbourne, Australia, 28 November–1 December 2016; pp. 799–804. [Google Scholar]
  16. Capizzi, G.; Sciuto, G.L.; Napoli, C.; Tramontana, E. Advanced and Adaptive Dispatch for Smart Grids by means of Predictive Models. IEEE Trans. Smart Grid 2017, 9, 6684–6691. [Google Scholar] [CrossRef]
  17. Mellit, A.; Pavan, A.M. A 24-h forecast of solar irradiance using artificial neural network: Application for performance prediction of a grid-connected PV plant at Trieste, Italy. Sol. Energy 2010, 84, 807–821. [Google Scholar] [CrossRef]
  18. Ehsan, R.M.; Simon, S.P.; Venkateswaran, P. Day-ahead forecasting of solar photovoltaic output power using multilayer perceptron. Neural Comput. Appl. 2017, 28, 3981–3992. [Google Scholar] [CrossRef]
  19. Gigoni, L.; Betti, A.; Crisostomi, E.; Franco, A.; Tucci, M.; Bizzarri, F.; Mucci, D. Day-Ahead Hourly Forecasting of Power Generation From Photovoltaic Plants. IEEE Trans. Sustain. Energy 2018, 9, 831–842. [Google Scholar] [CrossRef]
  20. Huang, C.; Bensoussan, A.; Edesess, M.; Tsui, K.L. Improvement in artificial neural network-based estimation of grid connected photovoltaic power output. Renew. Energy 2016, 97, 838–848. [Google Scholar] [CrossRef]
  21. Kingma, D.P.; Ba, J. Adam: A method for stochastic optimization. arXiv, 2014; arXiv:1412.6980. [Google Scholar]
  22. Barron, J.T. A more general robust loss function. arXiv, 2017; arXiv:1701.03077. [Google Scholar]
  23. Wang, L.; Zhang, Z.J.; Huang, C.; Tsui, K.L. A GPU-accelerated parallel Jaya algorithm for efficiently estimating Li-ion battery model parameters. Appl. Soft Comput. 2018, 65, 12–20. [Google Scholar] [CrossRef]
  24. Huang, C.; Wang, L.; Yeung, R.S.C.; Zhang, Z.J.; Chung, H.S.H.; Bensoussan, A. A Prediction Model-Guided Jaya Algorithm for the PV System Maximum Power Point Tracking. IEEE Trans. Sustain. Energy 2018, 9, 45–55. [Google Scholar] [CrossRef]
  25. Huang, C.; Wang, L.; Long, H.; Luo, X.; Wang, J.-H. A Hybrid Global Maximum Power Point Tracking Method for Photovoltaic Arrays under Partial Shading Conditions. Optik 2018, 180, 665–674. [Google Scholar] [CrossRef]
  26. Wang, L.; Huang, C.; Huang, L.M. Parameter estimation of the soil water retention curve model with Jaya algorithm. Comput. Electron. Agric. 2018, 151, 349–353. [Google Scholar] [CrossRef]
Figure 1. The structure of an MLP [20].
Figure 1. The structure of an MLP [20].
Sustainability 10 04863 g001
Figure 2. Histogram on the hourly power production.
Figure 2. Histogram on the hourly power production.
Sustainability 10 04863 g002
Figure 3. PV power production over 2017.
Figure 3. PV power production over 2017.
Sustainability 10 04863 g003
Figure 4. Forecasting performance comparison: (a) forecasting values; (b) forecasting errors.
Figure 4. Forecasting performance comparison: (a) forecasting values; (b) forecasting errors.
Sustainability 10 04863 g004
Table 1. Specifications of the PV power plant.
Table 1. Specifications of the PV power plant.
Nominal DC power68.48 kW
PV module typeNU-U240F1
PV module manufacturerSharp
Inverter type50 kW
Inverter manufacturerSatCon Technology
Table 2. RMSE and MAE on the test dataset.
Table 2. RMSE and MAE on the test dataset.
MethodRMSEMAE
Robust-MLP0.07750.0439
Generic-MLP0.07880.0459
Persistence0.09780.0474
Table 3. RMSE and MAE for each hour in the test dataset.
Table 3. RMSE and MAE for each hour in the test dataset.
HourRMSEMAE
Robust-MLPGeneric-MLPPersistenceRobust-MLPGeneric-MLPPersistence
6:000.00890.01100.00860.00530.00710.0032
7:000.02770.02970.03190.01670.01910.0139
8:000.05660.05720.06300.03910.03980.0339
9:000.07900.08080.09610.05300.05510.0538
10:000.09950.10130.12210.06490.06700.0686
11:000.09900.10120.12450.06570.06870.0725
12:000.10720.10860.14070.07240.07540.0849
13:000.11780.11970.15460.07680.07930.0937
14:000.11120.11250.14460.07290.07530.0879
15:000.10200.10380.12450.06900.07190.0775
16:000.07030.07110.08810.05200.05390.0492
17:000.02890.02990.03510.02000.02070.0193
18:000.00990.01180.00930.00600.00740.0042
19:000.00250.00290.00190.00130.00160.0007
Table 4. Daily forecasting performance on the test dataset.
Table 4. Daily forecasting performance on the test dataset.
MethodRMSEMAE
Robust-MLP0.65080.4370
Generic-MLP0.66350.4511
Persistence0.79880.4990
Back to TopTop