# Energy Flexibility Prediction for Data Center Engagement in Demand Response Programs

^{*}

## Abstract

**:**

## 1. Introduction

- Implementation of an ensemble-based DC energy prediction model that combines a set of individual neural network weak learners to forecast the DC energy demand for the next day and to refine it continuously considering four-hour intervals.
- Definition of energy flexibility in relation to the baseline load and a prediction model to forecast the potential DC energy flexibility to be used in DR programs.
- Implementation of a genetic heuristic to determine the optimal combination of the outcome of individual predictors to minimize the prediction error thus lowering the uncertainty concerning DR participation.

## 2. Related Work

## 3. DC Energy Prediction Model

#### 3.1. Demand Forecasting

- Day-ahead: energy values are forecasted for the next 24 h with a granularity of one hour;
- Intra-day: energy values are forecasted for the next 4 h with a granularity of half an hour;

- Season—the DC may consume/produce different quantities of energy depending on the season. For example, the energy consumption in summer can be higher than the energy consumption in winter especially due to more intensive use of cooling processes. Same reasoning may apply if we consider the renewable energy generation (i.e., solar energy). The possible values for this feature are: Spring, Summer, Autumn and Winter.
- Day of the week—a DC may consume different quantities of energy depending on the day of the week. For example, the energy consumption for Monday may be higher than the energy consumption in a weekend day such as Saturday if the DC is running banking tasks. Possible values for this feature are Sunday, Monday, Tuesday, Wednesday, Thursday, Friday, Saturday.
- Weekend—a DC may consume different quantities of energy depending on whether it is weekend day or not.

#### 3.2. Flexibility Forecasting

#### 3.3. Genetic Algorithm Based Ensemble

## 4. Experimental Results

#### 4.1. DC Energy Demand Prediction Results

#### 4.2. DC Flexibility Forecasting Results

## 5. Conclusions

## Author Contributions

## Funding

## Conflicts of Interest

## References

- International Energy Agency. Digitalisation and Energy. Technology Report. November 2017. Available online: https://www.iea.org/reports/digitalisation-and-energy (accessed on 15 December 2019).
- Koronen, C.; Ahman, M.; Nilsson, L.J. Data centres in future European energy systems—Energy efficiency, integration and policy. Energy Effic. J.
**2020**, 13, 129–144. [Google Scholar] [CrossRef] [Green Version] - Cioara, T.; Anghel, I.; Bertoncini, M.; Salomie, I.; Arnone, D.; Mammina, M.; Velivassaki, T.; Antal, M. Optimized flexibility management enacting Data Centres participation in Smart Demand Response programs. Future Gener. Comput. Syst.
**2018**, 78, 330–342. [Google Scholar] [CrossRef] - European Commission. EU Code of Conduct on Data Centre Energy Efficiency. Introductory Guide for All Applicants. Available online: https://e3p.jrc.ec.europa.eu/publications/ict-code-conductintroductory-guide-all-applicants-v312 (accessed on 15 December 2019).
- Huang, P.; Copertaro, B.; Zhang, X.; Shen, J.; Lofgren, I.; Ronnelid, M.; Fahlen, J.; Andersson, D.; Svanfeldt, M. A review of data centers as prosumers in district energy systems: Renewable energy integration and waste heat reuse for district heating. Appl. Energy
**2020**, 258, 114109. [Google Scholar] [CrossRef] - Alapera, I.; Honkapuro, S.; Paananen, J. Data centers as a source of dynamic flexibility in smart girds. Appl. Energy
**2018**, 229, 69–79. [Google Scholar] [CrossRef] - Ponnaganti, P.; Pillai, J.R.; Bak-Jensen, B. Opportunities and challenges of demand response in active distribution networks. WIREs Energy Environ.
**2018**, 7, e271. [Google Scholar] [CrossRef] - COMMUNICATION FROM THE COMMISSION, Delivering the Internal Electricity Market and Making the Most of Public Intervention. Available online: https://ec.europa.eu/energy/sites/ener/files/documents/com_2013_public_intervention_en_0.pdf (accessed on 15 December 2019).
- Cioara, T.; Anghel, I.; Antal, M.; Crisan, S.; Salomie, I. Data center optimization methodology to maximize the usage of locally produced renewable energy. In Proceedings of the 2015 Sustainable Internet and ICT for Sustainability (SustainIT), Madrid, Spain, 14–15 April 2015; pp. 1–8. [Google Scholar] [CrossRef] [Green Version]
- Ogedengbe, E.O.B.; Aderoju, P.A.; Nkwaze, D.C.; Aruwajoye, J.B.; Shitta, M.B. Optimization of energy performance with renewable energy project sizing using multiple objective functions. Energy Rep.
**2019**, 5, 898–908. [Google Scholar] [CrossRef] - Cioara, T.; Anghel, I.; Salomie, I.; Antal, M.; Pop, C.; Bertoncini, M.; Arnone, D.; Pop, F. Exploiting data centres energy flexibility in smart cities: Business scenarios. Inf. Sci.
**2019**, 476, 392–412. [Google Scholar] [CrossRef] - Feuerriegel, S.; Neumann, D. Integration scenarios of Demand Response into electricity markets: Load shifting, financial savings and policy implications. Energy Policy
**2016**, 96, 231–240. [Google Scholar] [CrossRef] [Green Version] - Nicholas Good, K.; Ellis, P.M. Review and classification of barriers and enablers of demand response in the smart grid. Renew. Sustain. Energy Rev.
**2017**, 72, 57–72. [Google Scholar] [CrossRef] [Green Version] - Fallah, S.N.; Deo, R.C.; Shojafar, M.; Conti, M.; Shamshirband, S. Computational Intelligence Approaches for Energy Load Forecasting in Smart Energy Management Grids: State of the Art, Future Challenges, and Research Directions. Energies
**2018**, 11, 596. [Google Scholar] [CrossRef] [Green Version] - Wang, Y.; Chen, Q.; Hong, T.; Kang, C. Review of smart meter data analytics: Applications, methodologies, and challenges. IEEE Trans. Smart Grid
**2018**, 10, 3125–3148. [Google Scholar] [CrossRef] [Green Version] - Zhang, X.; Li, Z.; Ma, L.; Chong, C.; Ni, W. Forecasting the Energy Embodied in Construction Services Based on a Combination of Static and Dynamic Hybrid Input-Output Models. Energies
**2019**, 12, 300. [Google Scholar] [CrossRef] [Green Version] - Miyuru, D.; Wen, Y.; Fan, R. Data center energy consumption modeling: A survey. IEEE Commun. Surv. Tutor.
**2015**, 18, 732–794. [Google Scholar] - Jim, G. Machine Learning Applications for Data Center Optimization. 2014. Available online: https://ai.google/research/pubs/pub42542 (accessed on 15 December 2019).
- Tseng, F.; Wang, X.; Chou, L.; Chao, H.; Leung, V.C.M. Dynamic Resource Prediction and Allocation for Cloud Data Center Using the Multiobjective Genetic Algorithm. IEEE Syst. J.
**2018**, 12, 1688–1699. [Google Scholar] [CrossRef] - Li, Y.; Hu, H.; Wen, Y.; Zhang, J. Learning-based power prediction for data centre operations via deep neural networks. In Proceedings of the 5th International Workshop on Energy Efficient Data Centres (E2DC ’16), Waterloo, ON, Canada, 21 June 2016; ACM: New York, NY, USA, 2016; p. 10. [Google Scholar] [CrossRef]
- Grange, L.; da Costa, G.; Stolf, P. Green IT scheduling for data center powered with renewable energy. Future Gener. Comput. Syst.
**2018**, 86, 99–120. [Google Scholar] [CrossRef] [Green Version] - Ferreira, J.; Callou, G.; Josua, A.; Tutsch, D.; Maciel, P. An Artificial Neural Network Approach to Forecast the Environmental Impact of Data Centers. Information
**2019**, 10, 113. [Google Scholar] [CrossRef] [Green Version] - Hu, Z.; Ma, J.; Yang, L.; Li, X.; Pang, M. Decomposition-Based Dynamic Adaptive Combination Forecasting for Monthly Electricity Demand. Sustainability
**2019**, 11, 1272. [Google Scholar] [CrossRef] [Green Version] - Antal, M.; Cioara, T.; Anghel, I.; Pop, C.; Salomie, I. Transforming Data Centers in Active Thermal Energy Players in Nearby Neighborhoods. Sustainability
**2018**, 10, 939. [Google Scholar] [CrossRef] [Green Version] - Makridakis, S.; Spiliotis, E.; Assimakopoulos, V. Statistical and Machine Learning forecasting methods: Concerns and ways forward. PLoS ONE
**2018**, 13, e0194889. [Google Scholar] [CrossRef] [Green Version] - Marino, D.L.; Amarasinghe, K.; Manic, M. Building energy load forecasting using Deep Neural Networks. In Proceedings of the IECON 2016—42nd Annual Conference of the IEEE Industrial Electronics Society, Florence, Italy, 23–26 October 2016; pp. 7046–7051. [Google Scholar]
- Cheng, Y.; Xu, C.; Mashima, D.; Thing, V.L.; Wu, Y. PowerLSTM: Power demand forecasting using long short-term memory neural network. In Proceedings of the International Conference on Advanced Data Mining and Applications, Singapore, 5–6 November 2017; Springer: Cham, Switzerland, 2017; pp. 727–740. [Google Scholar]
- Mocanu, E.; Nguyen, P.H.; Gibescu, M.; Kling, W.L. Deep learning for estimating building energy consumption Sustainable Energy. Grids Netw.
**2016**, 6, 91–99. [Google Scholar] - Fayaz, M.; Kim, D. A Prediction Methodology of Energy Consumption Based on Deep Extreme Learning Machine and Comparative Analysis in Residential Buildings. Electronics
**2018**, 7, 222. [Google Scholar] [CrossRef] - Liang, Y.; Niu, D.; Hong, W.C. Short term load forecasting based on feature extraction and improved general regression neural network model. Energy
**2019**, 166, 653–663. [Google Scholar] [CrossRef] - Rahman, H.; Selvarasan, I.; Begum, J. Short-Term Forecasting of Total Energy Consumption for India-A Black Box Based Approach. Energies
**2018**, 11, 3442. [Google Scholar] [CrossRef] [Green Version] - Zufferey, T.; Ulbig, A.; Koch, S.; Hug, G. Forecasting of Smart Meter Time Series Based on Neural Networks. Lect. Notes Comput. Sci.
**2017**, 10097, 10–21. [Google Scholar] [CrossRef] - Lee, S.; Jung, S.; Lee, J. Prediction Model Based on an Artificial Neural Network for User-Based Building Energy Consumption in South Korea. Energies
**2019**, 12, 608. [Google Scholar] [CrossRef] [Green Version] - Huang, C.-J.; Kuo, P.-H. A Short-Term Wind Speed Forecasting Model by Using Artificial Neural Networks with Stochastic Optimization for Renewable Energy Systems. Energies
**2018**, 11, 2777. [Google Scholar] [CrossRef] [Green Version] - Kampelis, N.; Tsekeri, E.; Kolokotsa, D.; Kalaitzakis, K.; Isidori, D.; Cristalli, C. Development of Demand Response Energy Management Optimization at Building and District Levels Using Genetic Algorithm and Artificial Neural Network Modelling Power Predictions. Energies
**2018**, 11, 3012. [Google Scholar] [CrossRef] [Green Version] - Fan, C.; Sun, Y.; Zhao, Y.; Song, M.; Wang, J. Deep learning-based feature engineering methods for improved building energy prediction. Appl. Energy
**2019**, 240, 35–45. [Google Scholar] [CrossRef] - Chen, K.; He, Z.; Wang, S.X. Learning-based Data Analytics: Moving Towards Transparent Power Grids. Csee J. Power Energy Syst.
**2018**, 4, 67–82. [Google Scholar] [CrossRef] - Chen, Y.; Tan, H. Short-term prediction of electric demand in building sector via hybrid support vector regression. Appl. Energy
**2017**, 204, 1363–1374. [Google Scholar] [CrossRef] - Kim, M.; Choi, W.; Jeon, Y.; Liu, L. A Hybrid Neural Network Model for Power Demand Forecasting. Energies
**2019**, 12, 931. [Google Scholar] [CrossRef] [Green Version] - Kuo, P.-H.; Huang, C.-J. An Electricity Price Forecasting Model by Hybrid Structured Deep Neural Networks. Sustainability
**2018**, 10, 1280. [Google Scholar] [CrossRef] [Green Version] - Zahid, M.; Ahmed, F.; Javaid, N.; Abbasi, R.A.; Zainab Kazmi, H.S.; Javaid, A.; Bilal, M.; Akbar, M.; Ilahi, M. Electricity Price and Load Forecasting using Enhanced Convolutional Neural Network and Enhanced Support Vector Regression in Smart Grids. Electronics
**2019**, 8, 122. [Google Scholar] [CrossRef] [Green Version] - He, W. Load Forecasting via Deep Neural Networks. Procedia Comput. Sci.
**2017**, 122, 308–314. [Google Scholar] [CrossRef] - Kim, T.Y.; Cho, S.B. Predicting the Household Power Consumption Using CNN-LSTM Hybrid Networks. In Proceedings of the International Conference on Intelligent Data Engineering and Automated Learning—IDEAL 2018, Madrid, Spain, 21–23 November 2018; pp. 481–490. [Google Scholar]
- Bouktif, S.; Fiaz, A.; Ouni, A.; Serhani, M.A. Optimal Deep Learning LSTM Model for Electric Load Forecasting using Feature Selection and Genetic Algorithm: Comparison with Machine Learning Approaches. Energies
**2018**, 11, 1636. [Google Scholar] [CrossRef] [Green Version] - Eseye, A.T.; Zhang, J.; Zheng, D. Short-term photovoltaic solar power forecasting using a hybrid wavelet-PSO- SVM model based on SCADA and meteorological information. Renew. Energy
**2017**, 118, 357–367. [Google Scholar] - Nayab, A.; Ashfaq, T.; Aimal, S.; Rasool, A.; Javaid, N.; Khan, Z.A. Load and Price Forecasting in Smart Grids Using Enhanced Support Vector Machine. In Advances in Internet, Data and Web Technologies. EIDWT 2019. Lecture Notes on Data Engineering and Communications Technologies; Barolli, L., Xhafa, F., Khan, Z., Odhabi, H., Eds.; Springer: Cham, Switzerland, 2019; Volume 29. [Google Scholar]
- Ouyang, T.; He, Y.; Li, H.; Sun, Z.; Baek, S. A Deep Learning Framework for Short-term Power Load Forecasting. Comput. Eng. Financ. Sci.
**2017**. [Google Scholar] - Fu, C.; Li, G.-Q.; Lin, K.-P.; Zhang, H.-J. Short-Term Wind Power Prediction Based on Improved Chicken Algorithm Optimization Support Vector Machine. Sustainability
**2019**, 11, 512. [Google Scholar] [CrossRef] [Green Version] - Data-Driven Baseline Estimation of Residential Buildings for Demand Response. Available online: https://www.mdpi.com/1996-1073/8/9/10239 (accessed on 15 December 2019).
- Rossetto, N. Measuring the Intangible: An Overview of the Methodologies for Calculating Customer Baseline Load in PJM. Florence School of Regulation. Available online: http://cadmus.eui.eu/bitstream/handle/1814/54744/RSC_PB_2018_05_FSR.pdf?sequence=1 (accessed on 15 December 2019).
- EU Smart City Cluster. Available online: https://www.smartcitiescluster.eu/publications (accessed on 15 December 2019).
- Wang, C.; Urgaonkar, B.; Wang, Q.; Kesidis, G.; Sivasubramaniam, A. Data Center Power Cost Optimization via Workload Modulation. In Proceedings of the 2013 IEEE/ACM 6th International Conference on Utility and Cloud Computing, Dresden, Germany, 9–12 December 2013. [Google Scholar]
- Rynkiewicz, J. Asymptotic statistics for multilayer perceptron with ReLU hidden units. Neurocomputing
**2019**, 342, 16–23. [Google Scholar] [CrossRef] - Le, X.-H.; Ho, H.V.; Lee, G.; Jung, S. Application of Long Short-Term Memory (LSTM) Neural Network for Flood Forecasting. Water
**2019**, 11, 1387. [Google Scholar] [CrossRef] [Green Version]

**Figure 9.**Electrical energy historical data used in forecasting (orange—cooling sub-system & blue—IT servers’ sub-system).

**Figure 10.**Day-ahead energy demand predictions using MLP: (

**a**) IT servers and (

**b**) cooling sub-systems.

**Figure 11.**Day-ahead energy demand predictions using LSTM: (

**a**) IT servers and (

**b**) cooling sub-systems.

**Figure 12.**Day-ahead electrical energy demand predictions using ensemble model: (

**a**) IT servers and (

**b**) cooling sub-systems.

**Figure 13.**Average MAPE values different prediction model configurations: (

**a**) IT servers and (

**b**) cooling sub-systems.

**Figure 14.**Intra-day electrical energy demand predictions using MLP model: (

**a**) IT servers and (

**b**) cooling sub-systems.

**Figure 15.**Intra-day electrical energy demand predictions using LSTM model: (

**a**) IT servers and (

**b**) cooling sub-system.

**Figure 16.**Intra-day electrical energy demand predictions using ensemble model: (

**a**) IT servers and (

**b**) cooling sub-system.

**Figure 17.**Day-ahead and intra-day energy demand prediction results vs. actual monitored values ((

**a**)—detailed results day number 4, (

**b**)—MAE distribution on 5 days of testing data).

Sub-System | Characteristics |
---|---|

Cooling system | $Coefficient\text{}of\text{}Performance=\text{}3.5$ $Maximum\text{}Cooling\text{}Capacity\text{}=\text{}4000\text{}\mathrm{kWh}$ $Minimum\text{}Cooling\text{}Load\text{}=\text{}200\text{}\mathrm{kWh}$ $Maximum\text{}Cooling\text{}Load\text{}=\text{}2000\text{}\mathrm{kWh}$ $PUE\text{}=\text{}1.3$ |

IT servers | $No=9000,\text{}Type=\text{}Servers\text{}HP\text{}360\text{}DL$ $Maximum\text{}Power\text{}Consumption=3000\text{}\mathrm{kWh}$ $Delay\text{}Tolerant\text{}Workload\text{}=\text{}20\%$ |

DC Component | Time Frame | Prediction Model | No. Models | Contextual Features | No. Inputs | No. Neurons on Hidden Layer | No. Outputs |
---|---|---|---|---|---|---|---|

IT servers consumption | Day-ahead | MLP | 1 | isWeekend | 25 | 37 | 24 |

LSTM | 1 | isWeekend | 25 | 47 | 24 | ||

Intra-day | MLP | 6 | partOfDay | 9 | 20 | 8 | |

LSTM | 6 | partOfDay | 9 | 16 | 8 | ||

Cooling consumption | Day-ahead | MLP | 1 | isWeekend | 25 | 37 | 24 |

LSTM | 1 | isWeekend | 25 | 47 | 24 | ||

Intra-day | MLP | 6 | partOfDay | 9 | 20 | 8 | |

LSTM | 6 | partOfDay | 9 | 16 | 8 |

Prediction Model | Time Frame | Best MAPE Value [%] | |
---|---|---|---|

IT Servers Sub-System | Cooling Sub-System | ||

MLP | Day-ahead | 8.68 | 8.68 |

Intra-day | 8.05 | 8.09 | |

LSTM | Day-ahead | 8.37 | 8.50 |

Intra-day | 8.08 | 8.24 | |

Ensemble | Day-ahead | 8.15 | 8.09 |

Intra-day | 7.20 | 6.81 |

DC Sub-System | Prediction Type | Input Features | N | M | $\mathit{N}\mathit{u}\mathit{m}\mathit{b}\mathit{e}\mathit{r}\text{}\mathit{o}\mathit{f}\text{}\mathit{i}\mathit{n}\mathit{p}\mathit{u}\mathit{t}\mathit{s}$ | $\mathit{H}$ | ${\mathit{C}}_{\mathit{F}}$ | $\mathit{O}\mathit{u}\mathit{t}\mathit{p}\mathit{u}\mathit{t}\mathit{s}$ |
---|---|---|---|---|---|---|---|---|

IT servers | Day-ahead | Historical load Historical baseline Current baseline Contextual features | 24 | 24 | 77 | 100 | 5 | 24 |

Cooling system | Day-ahead | Historical load Historical baseline Current baseline Contextual features Server room flexibility | 24 | 24 | 101 | 120 | 5 | 24 |

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Vesa, A.V.; Cioara, T.; Anghel, I.; Antal, M.; Pop, C.; Iancu, B.; Salomie, I.; Dadarlat, V.T.
Energy Flexibility Prediction for Data Center Engagement in Demand Response Programs. *Sustainability* **2020**, *12*, 1417.
https://doi.org/10.3390/su12041417

**AMA Style**

Vesa AV, Cioara T, Anghel I, Antal M, Pop C, Iancu B, Salomie I, Dadarlat VT.
Energy Flexibility Prediction for Data Center Engagement in Demand Response Programs. *Sustainability*. 2020; 12(4):1417.
https://doi.org/10.3390/su12041417

**Chicago/Turabian Style**

Vesa, Andreea Valeria, Tudor Cioara, Ionut Anghel, Marcel Antal, Claudia Pop, Bogdan Iancu, Ioan Salomie, and Vasile Teodor Dadarlat.
2020. "Energy Flexibility Prediction for Data Center Engagement in Demand Response Programs" *Sustainability* 12, no. 4: 1417.
https://doi.org/10.3390/su12041417