Evolution of a Summer Peak Intelligent Controller (SPIC) for Residential Distribution Networks
Abstract
:1. Introduction
- (1)
- Reviewing literature on peak load management in residential areas on a national and international scale.
- (2)
- Surveying Indian residential consumers regarding summer peak demand statistics.
- (3)
- Implementing the proposed SPIC algorithm in four residential MHEMS prototypes.
- (4)
- Implementing the proposed SPIC algorithm in a real-time TNEB system.
2. PAN INDIA Summer Peak Survey
2.1. Literature Review on Peak Load Management in India
2.2. Consumer Opinion Survey
3. Proposed Architecture and Algorithm
3.1. Multi-Home Energy Management System’s Architecture
- PRDT—power supplied by the residential distribution transformer;
- PH1—power consumed by Home-1;
- PH2—power consumed by Home-2;
- PH3—power consumed by Home-3;
- PH4—power consumed by Home-4;
- PHn—power consumed by Home-n.
3.2. Summer Peak Intelligent Controller (SPIC) Algorithm
4. Experimental Results and Discussion
4.1. Hardware Description
4.2. Analysis of Hardware Prototype
4.2.1. Case 1: Not Considering the Consumer’s Comfort
4.2.2. Case 2: Considering the Consumer’s Comfort
4.3. Analysis of Real-Time TNEB System
4.3.1. Case 1: Without Consideration of the Incentive Program
4.3.2. Case:2 with Consideration of the Incentive Program
5. Conclusions
Author Contributions
Funding
Data Availability Statement
Conflicts of Interest
Nomenclature
MHEMS | Multi-home energy management system |
RDT | Residential distribution transformer |
SPIC | Summer peak intelligent controller |
DSM | Demand side management |
TNEB | Tamil Nadu Electricity Board |
AC | Air conditioner |
HVAC | Heating, ventilation, and air conditioning |
LAN | Local area network |
References
- Central Electricity Authority. All India Installed Capacity (in MW) of Power Stations Installed Capacity (in MW). 2023, p. 7. Available online: https://cea.nic.in/installed-capacity-report/?lang=en (accessed on 13 July 2023).
- CEA INDIA. Central Electricity Authority Executive Summary on Power Sector Government of India Ministry of Power. 2022. Available online: https://cea.nic.in/wp-content/uploads/executive/2023/04/Executive_Summary_Apr_2023-1.pdf (accessed on 30 April 2023).
- Dong, M.; Nassif, A.B.; Li, B. A Data-Driven Residential Transformer Overloading Risk Assessment Method. IEEE Trans. Power Deliv. 2019, 34, 387–396. [Google Scholar] [CrossRef]
- Gomez, R.H.; Dizon, J.L.M.; Teodoro, C.A.S.; Demegillo, J.C.; Florencondia, N.T. Electrical Percent Loading Assessment for the Distribution Transformers Residential-Used of a Barangay. Int. J. Eng. Adv. Technol. 2020, 9, 3020–3024. [Google Scholar] [CrossRef]
- Tulabing, R.S.; Mitchell, B.C.; Covic, G.A.; Boys, J.T. Localized Demand Control of Flexible Devices for Peak Load Management. IEEE Trans. Smart Grid 2023, 14, 217–227. [Google Scholar] [CrossRef]
- Ahmed, M.S.; Mohamed, A.; Khatib, T.; Shareef, H.; Homod, R.Z.; Ali, J.A. Real time optimal schedule controller for home energy management system using new binary backtracking search algorithm. Energy Build. 2017, 138, 215–227. [Google Scholar] [CrossRef]
- Jeddi, B.; Mishra, Y.; Ledwich, G. Distributed load scheduling in residential neighborhoods for coordinated operation of multiple home energy management systems. Appl. Energy 2021, 300, 117353. [Google Scholar] [CrossRef]
- Haider, H.T.; Muhsen, D.H.; Al-Nidawi, Y.M.; Khatib, T.; See, O.H. A novel approach for multi-objective cost-peak optimization for demand response of a residential area in smart grids. Energy 2022, 254, 124360. [Google Scholar] [CrossRef]
- Gholinejad, H.R.; Adabi, J.; Marzband, M. An energy management system structure for Neighborhood Networks. J. Build. Eng. 2021, 41, 102376. [Google Scholar] [CrossRef]
- Dharmarajan, V.; Prakash, N.K. FPGA based Smart Home Energy Management System for Demand Response Applications. Int. J. Innov. Technol. Explor. Eng. IJITEE 2019, 8, 2278–3075. [Google Scholar]
- Popoola, O.; Chipango, M.B. Improved peak load management control technique for nonlinear and dynamic residential energy consumption pattern. Build. Simul. 2021, 14, 195–208. [Google Scholar] [CrossRef]
- Luo, F.; Ranzi, G.; Wan, C.; Xu, Z.; Dong, Z.Y. A Multistage Home Energy Management System with Residential Photovoltaic Penetration. IEEE Trans. Ind. Inform. 2019, 15, 116–126. [Google Scholar] [CrossRef]
- Farzaneh, H.; Malehmirchegini, L.; Bejan, A.; Afolabi, T.; Mulumba, A.; Daka, P.P. Artificial intelligence evolution in smart buildings for energy efficiency. Appl. Sci. 2021, 11, 763. [Google Scholar] [CrossRef]
- Corinaldesi, C.; Fleischhacker, A.; Lang, L.; Radl, J.; Schwabeneder, D.; Lettner, G. European case studies for impact of market-driven flexibility management in distribution systems. In Proceedings of the 2019 IEEE International Conference on Communications, Control, and Computing Technologies for Smart Grids (SmartGridComm), Beijing, China, 21–23 October 2019. [Google Scholar] [CrossRef]
- Corinaldesi, C.; Lettner, G.; Auer, H. On the characterization and evaluation of residential on-site E-car-sharing. Energy 2022, 246, 123400. [Google Scholar] [CrossRef]
- Sousa, J.F.B.; Borges, C.L.T.; Mitra, J. PV hosting capacity of LV distribution networks using smart inverters and storage systems: A practical margin. IET Renew. Power Gener. 2020, 14, 1332–1339. [Google Scholar] [CrossRef]
- Rahman, S.; Khan, I.A.; Khan, A.A.; Mallik, A.; Nadeem, M.F. Comprehensive review & impact analysis of integrating projected electric vehicle charging load to the existing low voltage distribution system. Renew. Sustain. Energy Rev. 2022, 153, 111756. [Google Scholar] [CrossRef]
- Singh, A.; Doolla, S.; Banerjee, R. Case Studies of DSM Projects Implemented in India. Available online: https://shaktifoundation.in/wp-content/uploads/2021/12/CaseStudies_DSM_2016.pdf (accessed on 30 April 2023).
- Fanda, R.B. Policy Brief Policy Brief. Pancanaka 2019, 1, 14. Available online: https://kebijakankesehatanindonesia.net/images/2019/policy_brief_wujudkan_keadilan_sosial_dalam_jkn.pdf (accessed on 1 November 2019).
- Jha, I.S.; Sen, S.; Kumar, R. Smart grid development in India—A case study. In Proceedings of the 2014 18th National Power Systems Conference (NPSC), Guwahati, India, 18–20 December 2014. [Google Scholar] [CrossRef]
- Chunekar, A.; Kelkar, M.; Dixit, S.; Prayas, S.D.; Lane, J.M. Demand Side Management in India: An Overview of State Level Initiatives. 2014. Available online: http://www.prayaspune.org/peg (accessed on 1 September 2014).
- Curiel, J.A.R.; Thakur, J. A novel approach for Direct Load Control of residential air conditioners for Demand Side Management in developing regions. Energy 2022, 258, 124763. [Google Scholar] [CrossRef]
- Ramalingam, S.P.; Shanmugam, P.K. Hardware Implementation of a Home Energy Management System Using Remodeled Sperm Swarm Optimization (RMSSO) Algorithm. Energies 2022, 15, 5008. [Google Scholar] [CrossRef]
- Chinnathambi, N.D.; Nagappan, K.; Samuel, C.R.; Tamilarasu, K. Internet of things-based smart residential building energy management system for a grid-connected solar photovoltaic-powered DC residential building. Int. J. Energy Res. 2022, 46, 1497–1517. [Google Scholar] [CrossRef]
- Anjana, M.S.; Ramesh, M.V.; Devidas, A.R.; Athira, K. Fractal IoT. In Proceedings of the 1st ACM International Workshop on Technology Enablers and Innovative Applications for Smart Cities and Communities, New York, NY, USA, 13–14 November 2019; pp. 10–17. [Google Scholar] [CrossRef]
- Sanjeev, P.; Padhy, N.P.; Agarwal, P. Peak energy management using renewable integrated DC microgrid. IEEE Trans. Smart Grid 2018, 9, 4906–4917. [Google Scholar] [CrossRef]
- Thakur, J.; Chakraborty, B. Demand side management in developing nations: A mitigating tool for energy imbalance and peak load management. Energy 2016, 114, 895–912. [Google Scholar] [CrossRef]
- Natarajan, S.; Loganathan, A.K. Analysis of Energy Management Controller in Grid-Connected PV Wind Power System Coupled with Battery Using Whale Optimisation Algorithm. Iran. J. Sci. Technol. Trans. Electr. Eng. 2022, 46, 77–90. [Google Scholar] [CrossRef]
- Kumar, K.K.P.; Soren, N.; Latif, A.; Das, D.C.; Hussain, S.M.S.; Al-Durra, A.; Ustun, T.S. Day-Ahead DSM-Integrated Hybrid-Power-Management-Incorporated CEED of Solar Thermal/Wind/Wave/BESS System Using HFPSO. Sustainability 2022, 14, 1169. [Google Scholar] [CrossRef]
- Power, M.O. Household Electrification Status. 2019. Available online: https://saubhagya.gov.in/ (accessed on 11 October 2017).
- Bhalla, A.K.; Abhay, B. Energy Conservation in Building Space Cooling through Optimum Temperature Setting; Bureau of Energy Efficiency, Ministry of Power, Government of India: New Delhi, India, 2018.
Home Loads | Actual Loads in Prototype | Type of Load |
---|---|---|
FAN Load—72 W | DC Fan—3 W | Small Load—Fan Load |
AC Load—1800 W | Incandescent Lamp—100 W | Heavy Load—AC Load |
Other Loads-1—300 to 400 W | Incandescent Lamp—60 W | Medium Loads—WM/VC/RF/RO |
Other Loads-2—200 to 300 W | Incandescent Lamp—40 W | Small Loads—Light/FAN/TV |
List of Homes | AC Power Rating (kW) | List of Homes | AC Power Rating (kW) | List of Homes | AC Power Rating (kW) |
---|---|---|---|---|---|
H1 | 1.8 | H11 | 1.8 | H22 | 1.8 |
H2 | 1.8 | H12 | 1.5 | H24 | 1.5 |
H4 | 1.5 | H13 | 1.8 | H25 | 1.8 |
H6 | 1.5 | H16 | 1.5 | H28 | 1.5 |
H7 | 1.5 | H17 | 1.5 | ||
H9 | 1.8 | H19 | 1.8 |
List of Homes | Level of Priority | AC Power Rating (kW) | List of Homes | Level of Priority | AC Power Rating (kW) |
---|---|---|---|---|---|
H1 | 75% | 1.8 | H16 | 50% | 1.5 |
H2 | 50% | 1.8 | H17 | 75% | 1.5 |
H4 | 50% | 1.5 | H19 | 100% | 1.8 |
H6 | 75% | 1.5 | H21 | 25% | 1.5 |
H7 | 100% | 1.5 | H22 | 50% | 1.8 |
H9 | 50% | 1.8 | H23 | 100% | 1.5 |
H11 | 100% | 1.8 | H24 | 100% | 1.8 |
H12 | 75% | 1.5 | H26 | 25% | 1.8 |
H13 | 100% | 1.8 | H28 | 75% | 1.5 |
H15 | 25% | 1.8 | H30 | 25% | 1.5 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2023 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 (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Parangusam, K.; Lekshmana, R.; Gono, T.; Gono, R. Evolution of a Summer Peak Intelligent Controller (SPIC) for Residential Distribution Networks. Energies 2023, 16, 6681. https://doi.org/10.3390/en16186681
Parangusam K, Lekshmana R, Gono T, Gono R. Evolution of a Summer Peak Intelligent Controller (SPIC) for Residential Distribution Networks. Energies. 2023; 16(18):6681. https://doi.org/10.3390/en16186681
Chicago/Turabian StyleParangusam, Kanakaraj, Ramesh Lekshmana, Tomas Gono, and Radomir Gono. 2023. "Evolution of a Summer Peak Intelligent Controller (SPIC) for Residential Distribution Networks" Energies 16, no. 18: 6681. https://doi.org/10.3390/en16186681
APA StyleParangusam, K., Lekshmana, R., Gono, T., & Gono, R. (2023). Evolution of a Summer Peak Intelligent Controller (SPIC) for Residential Distribution Networks. Energies, 16(18), 6681. https://doi.org/10.3390/en16186681