A hesitant fuzzy wind speed forecasting system with novel defuzzification method and multi-objective optimization algorithm

https://doi.org/10.1016/j.eswa.2020.114364Get rights and content

Highlights

  • A novel hesitant fuzzy wind speed forecasting system is proposed for the first time.

  • Multi-fuzzification methods are proposed to deal with the non-determinism problem.

  • The weights of intervals are determined by multi-objective optimization algorithm.

  • A new defuzzification model is developed to obtain accurate and reliable forecasts.

  • The proposed system outperforms comparison models with high accuracy and efficiency.

Abstract

Owing to the nondeterministic nature of wind speed, the conventional fuzzy time series forecasting model has difficulty in establishing a common membership level. Therefore, in this study, the fuzzy series forecasting model was improved based on hesitant fuzzy sets. A hesitant fuzzy wind speed forecasting system with a novel defuzzification method and multiobjective optimization algorithm was developed. First, an advanced decomposition model is employed to extract the effective feature and remove the noise component from the raw wind speed series. Then, the universe of discourse is partitioned into equal and unequal intervals by multifuzzification methods and merged by aggregating hesitant information. A multiobjective intelligent optimization algorithm is applied to determine the optimal weights of different intervals accurately and stably. Furthermore, a novel defuzzification model based on an ordered weighted averaging operator and a regular increasing monotone quantifier is proposed to calculate the final forecasting results. The crucial strengths of the developed system are verifying the possibility of enhancing the performance of wind speed forecasting models by improving conventional fuzzy time series forecasting models and integrating them with decomposition models and artificial-intelligence models. Typical wind speed series datasets with different resolutions were selected to evaluate the performance of the proposed system, and experimental results prove that the proposed system outperforms other comparison models with high forecasting accuracy and computing efficiency.

Introduction

Industrialization consumes a large amount of energy reserves, and this has caused countries worldwide to face the current energy shortage situation. The massive consumption of nonrenewable resources not only results in polluting gases being emitted in the surrounding air, but it also causes serious damage to the ecology and vegetation. Wind power can help alleviate the pressure of the energy crisis and environmental pollution. The share of wind energy in the energy supply of various communities is increasing, and it can provide energy in many modern cities and societies. Accurately forecasting wind speed can help to find appropriate solutions to the optimal operation and planning problems of different parts of the power system (such as the distribution and transmission parts) and to optimize the design of wind farms (Dong et al., 2017, Pearre and Swan, 2018). At the same time, it can significantly reduce the risk of using wind energy in the power system and increase the penetration rate of the renewable energy resource (RES).

The biggest difference between wind energy and other energy sources is the fuzziness, randomness, and uncontrollability of wind speed (Li et al., 2019). Wind power is generated from the wind, which has the characteristics of volatility and intermittency, which makes the dispatch and control of wind farms more difficult.

The objective of this study was to develop an accurate and effective wind speed forecasting system that can better deal with the fuzziness, randomness, and nonlinearity of wind speed series. Owing to the advantages of machine-learning models and fuzzy-logic models, a hesitant fuzzy wind speed forecasting system with a novel aggregate information method and multiobjective optimization algorithm is proposed. It mainly consists of three modules: a data preprocessing module, a fuzzy computation and forecasting module, and a results evaluation module.

In the data preprocessing module, an improved empirical mode decomposition (EMD) powered by the soft sifting stopping criterion is applied to the preprocessing of the wind speed series. It has strong robustness. In the fuzzy computation and forecasting module, the rate of change of the wind speed series is calculated to reduce the impact of data magnitude on forecasting accuracy. Multifuzzification methods (an equal-interval-based method and equal-frequency-based method) are proposed to fuzzify the datasets. This can solve the problem of the nondeterministic nature of fuzzy time series (FTS). Moreover, a multiobjective optimization algorithm is employed to determine the optimal weights of different intervals accurately and stably. Furthermore, a defuzzification approach that integrates a combined midpoint with an ordered weighted averaging (OWA) operator and a regular increasing monotone (RIM) quantifier are proposed to calculate the final forecasting results. In the results evaluation module, the out-of-sample R2 is computed to compare the performance of the proposed system with the comparison models, and the Diebold–Mariano (DM) test is carried out to highlight the difference in accuracy between the proposed system and the benchmark models.

The remainder of this article is organized as follows. Section 2 provides a review of prior work on wind speed forecasting. Section 3 presents the details of the proposed wind speed forecasting system with a novel aggregate information method and multiobjective optimization algorithm. Section 4 introduces the experimental preparation. In Section 5, the experimental results on wind speed forecasting datasets are analyzed. Further comparisons are discussed in Section 6. Finally, the key conclusions are summarized in Section 7.

Section snippets

Literature review

To obtain accurate forecasting results, a variety of approaches have been developed. Table 1 shows the current wind speed forecasting methods. It can be summarized as follows. (1) Physical forecasting approaches have an advantage in long-term wind speed forecasting, but they require a large number of data, and information should be described in detail based on the conditions of the wind farm (Chen et al., 2018). (2) Statistical models are easy to model and lack the ability to process nonlinear

Design of the wind speed forecasting system

A novel wind speed forecasting system is proposed that contains three main modules: a data preprocessing module, a fuzzy computation and forecasting module, and a results evaluation module. The framework of the proposed wind speed forecasting system is shown in Fig. 1.

Experiment preparation

In this section, detailed information about the experiments is introduced, including a description of the datasets, evaluation criteria, and the experimental arrangement.

Wind speed forecasting with different time resolutions

The wind speed series was first decomposed into several components, and the highest-frequency component was regarded as a random noise impact. Then, to better extract the features of the wind speed series and smooth the original series, the high-frequency component was eliminated, and the series was reconstructed to carry out experiments. Fig. 2 depicts the decomposition and feature selection process of the 10-min dataset at site 1. The figure shows that the reconstructed series generated by

Further discussion

In order to further verify the effectiveness and efficiency of the proposed system, this section discusses the statistical test and the operation time of the system.

Conclusion

The uncertainty, nonlinearity, and fuzzy characteristics embedded in wind speed series lead to unsatisfactory results, and few studies have been conducted on aggregating hesitant fuzzy information during wind speed forecasting. This study improved the conventional FTS forecasting models based on HFSs and developed a wind speed forecasting system based on hesitant fuzzy information for the first time. Multifuzzification methods were proposed to fuzzify the universe of discourse into equal and

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This project is supported by National Natural Science Foundation of China (Grant No. 71671029) and High-level Innovation Team Overseas Training Project of Department of Education of Liaoning province in China (No. 2018LNGXGJWPY-ZD005).

References (63)

  • P. Du et al.

    A novel hybrid model for short-term wind power forecasting

    Applied Soft Computing

    (2019)
  • E. Erdem et al.

    ARMA based approaches for forecasting the tuple of wind speed and direction

    Applied Energy

    (2011)
  • S.-W. Fei et al.

    Wind speed prediction using the hybrid model of wavelet decomposition and artificial bee colony algorithm-based relevance vector machine

    International Journal of Electrical Power & Energy Systems

    (2015)
  • Z. He et al.

    A novel wind speed forecasting model based on moving window and multi-objective particle swarm optimization algorithm

    Applied Mathematical Modelling

    (2019)
  • J. Heng et al.

    Research and application of a combined model based on frequent pattern growth algorithm and multi-objective optimization for solar radiation forecasting

    Applied Energy

    (2017)
  • K. Huarng

    Effective lengths of intervals to improve forecasting in fuzzy time series

    Fuzzy Sets and Systems

    (2001)
  • H. Jahangir et al.

    Short-term wind speed forecasting framework based on stacked denoising auto-encoders with rough ANN

    Sustainable Energy Technologies and Assessments

    (2020)
  • P. Jiang et al.

    Variable weights combined model based on multi-objective optimization for short-term wind speed forecasting

    Applied Soft Computing

    (2019)
  • P. Jiang et al.

    A hybrid forecasting system based on fuzzy time series and multi-objective optimization for wind speed forecasting

    Applied Energy

    (2019)
  • C. Li et al.

    An innovative hybrid system for wind speed forecasting based on fuzzy preprocessing scheme and multi-objective optimization

    Energy

    (2019)
  • H. Li et al.

    Research and application of a combined model based on variable weight for short term wind speed forecasting

    Renewable Energy

    (2018)
  • H. Liu et al.

    A novel two-stage deep learning wind speed forecasting method with adaptive multiple error corrections and bivariate Dirichlet process mixture model

    Energy Conversion and Management

    (2019)
  • M. Liu et al.

    Short-term wind speed forecasting based on the Jaya-SVM model

    International Journal of Electrical Power & Energy Systems

    (2020)
  • X. Liu et al.

    Orness and parameterized RIM quantifier aggregation with OWA operators: A summary

    International Journal of Approximate Reasoning

    (2008)
  • Y. Liu et al.

    Probabilistic spatiotemporal wind speed forecasting based on a variational Bayesian deep learning model

    Applied Energy

    (2020)
  • Z. Liu et al.

    A combined forecasting model for time series: Application to short-term wind speed forecasting

    Applied Energy

    (2020)
  • Z. Liu et al.

    Time-frequency representation based on robust local mean decomposition for multicomponent AM-FM signal analysis

    Mechanical Systems and Signal Processing

    (2017)
  • S. Mirjalili et al.

    Salp Swarm Algorithm: A bio-inspired optimizer for engineering design problems

    Advances in Engineering Software

    (2017)
  • N.S. Pearre et al.

    Statistical approach for improved wind speed forecasting for wind power production

    Sustainable Energy Technologies and Assessments

    (2018)
  • S. Pei et al.

    Wind speed prediction method based on Empirical Wavelet Transform and New Cell Update Long Short-Term Memory network

    Energy Conversion and Management

    (2019)
  • Z. Peng et al.

    A novel deep learning ensemble model with data denoising for short-term wind speed forecasting

    Energy Conversion and Management

    (2020)
  • Cited by (36)

    View all citing articles on Scopus
    View full text