Characterizing Time Series Data Diversity for Wind Forecasting

Feng, Cong; Chartan, Erol Kevin; Hodge, Bri‐Mathias; Zhang, Jie

doi:10.1145/3148055.3148065

Cited by 11 publications

(7 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is observed that the load profiles have evident diurnal patterns. This is also proved by a time series analysis that shows all the load time series have the periodicity of 24 (1 day) [21,23]. Moreover, load patterns of the 13 buildings are different, which could be further validated by load statistics shown in Fig.…”

Section: Data Description and Pre-analysismentioning

confidence: 57%

Short-Term Load Forecasting With Different Aggregation Strategies

Feng

Zhang

2018

Volume 2A: 44th Design Automation Conference

Self Cite

View full text Add to dashboard Cite

Effective short-term load forecasting (STLF) plays an important role in demand-side management and power system operations. In this paper, STLF with three aggregation strategies are developed, which are information aggregation (IA), model aggregation (MA), and hierarchy aggregation (HA). The IA, MA, and HA strategies aggregate inputs, models, and forecasts, respectively, at different stages in the forecasting process. To verify the effectiveness of the three aggregation STLF, a set of 10 models based on 4 machine learning algorithms, i.e., artificial neural network, support vector machine, gradient boosting machine, and random forest, are developed in each aggregation group to predict 1-hour-ahead load. Case studies based on 2-year of university campus data with 13 individual buildings showed that: (a) STLF with three aggregation strategies improves forecasting accuracy, compared with benchmarks without aggregation; (b) STLF-IA consistently presents superior behavior than STLF based on weather data and STLF based on individual load data; (c) MA reduces the occurrence of unsatisfactory single-algorithm STLF models, therefore enhancing the STLF robustness; (d) STLF-HA produces the most accurate forecasts in distinctive load pattern scenarios due to calendar effects.

show abstract

Section: Data Description and Pre-analysismentioning

confidence: 57%

Short-Term Load Forecasting With Different Aggregation Strategies

Feng

Zhang

2018

Volume 2A: 44th Design Automation Conference

Self Cite

View full text Add to dashboard Cite

show abstract

“…Feng et al in [7] analyse the characterisation of the time series structure by applying decomposition, linearity analysis of entropy. This approach is used in a subsequent work [8] where the defined characterisations are applied to wind sites in North America, analysing the relationship of the uncertainty in forecastability to spectral entropy using regression approaches.…”

Section: Previous Work In Forecastabiltymentioning

confidence: 99%

Forecastability Measures that Describe the Complexity of a Site for Deep Learning Wind Predictions

Manero

Béjar²

2021

JSFI

View full text Add to dashboard Cite

The application of deep learning to wind time series for multi-step prediction obtains good results at short horizons. The accuracy of a wind forecast is highly dependent on the specific structure of wind in the specific location, as many local features influence wind behaviour. The characterization of the complexity of a site for wind prediction is defined as forecastability or predictability and can be obtained from the inner structure of the meteorological time series observations from a site. We analyze the time series structure searching for properties that have a high correlation with the prediction result, properties that can create measures that have the potential to describe the forecastability of a site. The best measures will show a high correlation with the accuracy of the predictions. In this work, we analyze wind time series from 126,692 wind locations in the US, where we apply several deep learning methods first, and then we verify several forecastability descriptors with the accuracy deep learning results. We require High-Performance Computing (HPC) resources for this task as the deep learning algorithms have sensible resource requirements and are applied to a large set of data. The measures defined and explored in this work are based on several techniques that decompose or transform the wind time-series. By combining several of these measures, we can obtain better predictors of the site complexity, which will allow us to evaluate the future error of a prediction on this site. Forecastability measures can contribute to a wind site multi-dimensional description, becoming a valuable tool for wind resource analysts and wind forecasters.

show abstract

“…To ensure the efficiency of HS-based forecasting, diurnal patterns and hourly similarities in solar data need to be explored and identified. Table I shows the periodicity, trend, and seasonality of the six solar features, which are calculated by a time series characteristic analysis [11], [12]. The periodicity in Table I indicates the frequency an observation occurs in the time domain.…”

Section: B Diurnal Patterns and Hourly Similaritiesmentioning

confidence: 99%

Hourly-Similarity Based Solar Forecasting Using Multi-Model Machine Learning Blending

Feng

Zhang

2018

2018 IEEE Power &Amp; Energy Society General Meeting (PESGM)

Self Cite

View full text Add to dashboard Cite

With the increasing penetration of solar power into power systems, forecasting becomes critical in power system operations. In this paper, an hourly-similarity (HS) based method is developed for 1-hour-ahead (1HA) global horizontal irradiance (GHI) forecasting. This developed method utilizes diurnal patterns, statistical distinctions between different hours, and hourly similarities in solar data to improve the forecasting accuracy. The HS-based method is built by training multiple two-layer multi-model forecasting framework (MMFF) models independently with the same-hour subsets. The final optimal model is a combination of MMFF models with the best-performed blending algorithm at every hour. At the forecasting stage, the most suitable model is selected to perform the forecasting subtask of a certain hour. The HS-based method is validated by 1-year data with six solar features collected by the National Renewable Energy Laboratory (NREL). Results show that the HS-based method outperforms the non-HS (all-in-one) method significantly with the same MMFF architecture, wherein the optimal HSbased method outperforms the best all-in-one method by 10.94% and 7.74% based on the normalized mean absolute error and normalized root mean square error, respectively.

show abstract

Characterizing Time Series Data Diversity for Wind Forecasting

Cited by 11 publications

References 29 publications

Short-Term Load Forecasting With Different Aggregation Strategies

Short-Term Load Forecasting With Different Aggregation Strategies

Forecastability Measures that Describe the Complexity of a Site for Deep Learning Wind Predictions

Hourly-Similarity Based Solar Forecasting Using Multi-Model Machine Learning Blending

Contact Info

Product

Resources

About