The accuracy of electricity consumption forecasts is of paramount importance in energy planning, it provides strong support for the effective energy demand management. In this work, we proposed a load forecast through the decomposition of the historical time series in relation to the historical evolution of each hour of the day. The output of these decomposition were served as input to different algorithms of machine learning. We tested our model by five machines learning methods, the achieved results are examined with three of the most commonly used evaluation measures in forecasting. The obtained results were very satisfactory.
Citation: Abdelkarim El khantach, Mohamed Hamlich, Nour eddine Belbounaguia. Short-term load forecasting using machine learning and periodicity decomposition[J]. AIMS Energy, 2019, 7(3): 382-394. doi: 10.3934/energy.2019.3.382
Related Papers:
[1]
Nadwan Majeed Ali, Handri Ammari .
Design of a hybrid wind-solar street lighting system to power LED lights on highway poles. AIMS Energy, 2022, 10(2): 177-190.
doi: 10.3934/energy.2022010
[2]
Mulualem T. Yeshalem, Baseem Khan .
Design of an off-grid hybrid PV/wind power system for remote mobile base station: A case study. AIMS Energy, 2017, 5(1): 96-112.
doi: 10.3934/energy.2017.1.96
[3]
Nadejda Komendantova, Markus Manuel Schwarz, Wolfgang Amann .
Economic and regulatory feasibility of solar PV in the Austrian multi-apartment housing sector. AIMS Energy, 2018, 6(5): 810-831.
doi: 10.3934/energy.2018.5.810
[4]
Fadhil Khadoum Alhousni, Firas Basim Ismail, Paul C. Okonkwo, Hassan Mohamed, Bright O. Okonkwo, Omar A. Al-Shahri .
A review of PV solar energy system operations and applications in Dhofar Oman. AIMS Energy, 2022, 10(4): 858-884.
doi: 10.3934/energy.2022039
[5]
Jin H. Jo, Kadi Ilves, Tyler Barth, Ellen Leszczynski .
Implementation of a large-scale solar photovoltaic system at a higher education institution in Illinois, USA. AIMS Energy, 2017, 5(1): 54-62.
doi: 10.3934/energy.2017.1.54
[6]
Sabir Rustemli, Zeki İlcihan, Gökhan Sahin, Wilfried G. J. H. M. van Sark .
A novel design and simulation of a mechanical coordinate based photovoltaic solar tracking system. AIMS Energy, 2023, 11(5): 753-773.
doi: 10.3934/energy.2023037
[7]
Michael O. Dioha, Atul Kumar .
Rooftop solar PV for urban residential buildings of Nigeria: A preliminary attempt towards potential estimation. AIMS Energy, 2018, 6(5): 710-734.
doi: 10.3934/energy.2018.5.710
[8]
Jin H. Jo, Jamie Cross, Zachary Rose, Evan Daebel, Andrew Verderber, David G. Loomis .
Financing options and economic impact: distributed generation using solar photovoltaic systems in Normal, Illinois. AIMS Energy, 2016, 4(3): 504-516.
doi: 10.3934/energy.2016.3.504
[9]
Rashmi Murali, Sangeeta Malhotra, Debajit Palit, Krishnapada Sasmal .
Socio-technical assessment of solar photovoltaic systems implemented for rural electrification in selected villages of Sundarbans region of India. AIMS Energy, 2015, 3(4): 612-634.
doi: 10.3934/energy.2015.4.612
[10]
Md. Mehadi Hasan Shamim, Sidratul Montaha Silmee, Md. Mamun Sikder .
Optimization and cost-benefit analysis of a grid-connected solar photovoltaic system. AIMS Energy, 2022, 10(3): 434-457.
doi: 10.3934/energy.2022022
Abstract
The accuracy of electricity consumption forecasts is of paramount importance in energy planning, it provides strong support for the effective energy demand management. In this work, we proposed a load forecast through the decomposition of the historical time series in relation to the historical evolution of each hour of the day. The output of these decomposition were served as input to different algorithms of machine learning. We tested our model by five machines learning methods, the achieved results are examined with three of the most commonly used evaluation measures in forecasting. The obtained results were very satisfactory.
Symbols: Am: Module area; C: Cost (IQD); CAh: Battery ampere capacity (A/h); Cwh: Battery watt capacity (W/h); d: Interest rate (%); E/m2: Energy received by PV (kW/m2); EL: Energy consumed by the load (kW); IC: Controller current (A); i: inflation rate (%); Nc: Number of controller; Npv: Number of solar cell; Vb: Battery voltage (V)
Abbreviations: MPW: Maintenance present worth; DOD: Permissible depth of discharge rate of a cell; IOP: Inverter output power; IQD: Iraqi Dinar; LCC: life cycle cost; ML: Maximum load; NC: Number of continuous cloudy days; PR: Performance ratio; SF: Safety factor
Greek letters: η module: PV module efficiency; ηb: Battery efficiency
The Electrical energy that produced from Photovoltaic solar system has become an excellant option to offset the growing demand of electricity around the world [1], reliable and has the potential to reduce the CO2 emissions [2]. Many countries have effectively turned on this type of energy due to the availability of the solar radiation over a wide range of the year. This led many researchers to focus their attention on developing and increasing the efficiency of photovoltaic solar systems to achieve an economic feasibility. In the last years, some of studies in literatures have examined the performance evaluation of photovoltaic solar systems. The evaluation was based on some main indicators that reported in these studies including: the total energy production, performance ratio (PR), and the system efficiency. Renu S. and Sonali G. [3] examined the annual performance of 11.2 kWp in India. The experimental examination showed that the total energy, efficiency and PR were 14.96 MWh, 13.42% and 0.78 respectively. In Italy, the performance analysis of two systems 395.61 kWp and 1042.29 kWp showed that the PR of the two systems was 0.872 and 0.832 respectively [4]. In Morocco, Charaf H. et al.[5] examined experimentally different modules and sizes of PV systems and showed that the Twidell was the best module with PR = 0.8 and an average efficiency of 16%. The performance assessment of 4.5 kWp in Nigeria showed that the energy production of the system was 14.52% from the size of the system [6]. In addition, the performance analysis of 1.72 kWp system in Ireland was presented experimentally by Ayompe L.M. et al.[7]. The analysis showed that the annual total energy of the system was 885.1 kW h/kWp and system efficiency was 14.9%. On the other side, some of studies were focused on an economic feasibility of the PV solar systems depending on two main important indicators include life cycle and pay-back period. In Ireland, the experimental feasibility of 53.625 kWp PV system showed that the life cycle was 5.23 years [8]. In Japan, Yanxue L. et al.[9] simulated Techno economic performances of 5.0 kWp and showed that the net present value of system was over 20 years. As reported in Rodrigues S. et al.[10] study, the investment of 5 kWp system during 25 years in Germany and Italy was 13% higher than the most European countries. The payback period of 67.27 kWp system was recorded 11 years in United Kingdom [11]. In Palestine, the economic evaluation of 24.614 kWp system showed that dynamic pay-back period of system was 10.4 years [12]. The main goal of this work is to study the necessary of using the PV system to provide the required electricity for Iraqi houses. An economic feasibility investigation of 3 kWp PV solar system by calculating the life cycle cost was analyzed and compared with the two alternative sources represented in neighborhood and household generators.
2.
Problem and aim of the study
The electricity sector in Iraq suffered from deterioration since 1991 till now due to:
1. Old power generation plants, distribution and transmision networks;
2. Economic embargo in 1991;
3. American invasion in 2003;
4. Vandalism and neglecting from Iraqi goverment.
These factors led to severe power shortages to households, with power being available for less than 12 hours per day. To meet these power shortages, Iraqi families were forced to use other alternative sources such as the diesel neighborhood generators together with small gasoline portable generators (household), both of which have added a considerable financial burden on people. The financial burden includes the prices of generators, fuel and maintenance costs. In view of this phenomenon, the main goal of this work was to suggest and analyze the feasibility economic of a 3 kW PV solar system and compare it with the financial burden of the two generators (neighborhood and household). The study was conducted for a period of five years from (2014 to 2019) according to the current Iraqi market price. Questionnaire form was prepared to facilitate the calculation cost of neighborhood and household generators as shown in Table 1. The information of a questionnaire form was taken in the region where the neighborhood generator installed. Numbers of parameters were included in the questionnaire form, some of these parameters were related to the neighborhood generator (number and price of ambers) and other parameters were related to household generator (purchase cost, operation hours and maintenance cost).
Table 1.
Questionnaire form of generators information.
Neighborhood generator is a private diesel generator widely used in Iraq after 2003 to provide the electricity during the duration of national power outage. It is invested by a person and placed in a small neighborhood. The Iraqi government supplies the fuel to the generators at a reasonable price. So, the price of ampere from the neighborhood generator are controlled at 8000 IQD/ampere. According to the questionnaire form, the average of amperes demand of Iraqi family is equal to 4 amperes (equivalent to 1 kWh at 250 V supply). Therefore, the cost to be paid to the neighborhood generators for four amperes in one month is 32,000 IQD with total cost (Cn-g) in five years is 1,920,000 IQD (Cn-g = 4 amperes × 8,000 (IQD/amperes×month) × 60 month = 1,920,000 IQD).
4.
Cost of household generator
Iraqi markets contain different types of portable gasoline generators with different rated power (KVA) and different specifications. Gasoline generator model (Tiger-single phase-3KVA) with fuel capacity 10 liters and 0.5 liters oil were included in this study as shown in Figure 1. Schedule run of portable generator is from 5:00 am until 12:00 pm with an actual operation period 4 hours/day and the operation period depends on the national power outage and neighborhood generator off. The cost of household generator is listed in Table 2 including purchase, operation and maintenance cost. The cost calculation of household generator was investigated based on the data of Table 1.
The average maximum load was calculated in the summer and according to the actual need for household appliances as shown in Table 3. The actual operation period of appliances depends on the schedule of power shortages. Based on the table, the maximum energy demand is 1.11 kW then the maximum hourly load consumption should be 17.73 kWh. Figure 2 represents the actual load profile according to the number of operation hours.
Table 3.
Measured appliances loads in Iraqi house.
A 3 kWp system with twelve solar module (Dusol DS60260W Dubi) made from polycrystalline silicon was proposed in this study according to the calculation of the load and sizing of components. The general specifications of the solar system components are summarized in Table 4 and its photograph shown in Figure 3.
Table 4.
General specifications of the solar system components.
The sizing of the PV solar system components includes: number of solar panel, inverter output power, storage capacity of the battery and MPPT controller.
5.3.1. Number of solar cell
The total peak power and the module number of the proposed PV system can be calculated as [13]:
PPV=ELPSH×ηR×ηin×SF=17.738.7×0.9×0.92×1.2=2,855Wp
(1)
NPV=PPVPmodule=2855250≈12cells
(2)
where: EL: Daily Energy consumption (17.73) kWh, PSH: peak sun hours (8.7) hours [14], ηR and ηin: charge regulator and inverter efficiencies (0.9 and 0.92) [13], SF: Safety factor (1.2) [15] and Pmodule: module power is selected (250) W.
5.3.2. Inverter sizing
The inverter output power (IOP) calculated according to [16]: IOP = (25−30)% greater than maximum load (ML) = 0.25 × 1.11 + 1.11 = 1.38 ≈ 1.5 kW. Therefore, the inverter specifications is selected as 1500 W, 24 V DC, 220 V AC, and 50 Hz.
5.3.3. Sizing of battery
As reported in [12], the watt hour capacity of pattery can be calculated as:
NC: Number of continuous cloudy days of the interested area = 0.5, DOD: Permissible depth of discharge rate of a cell = 0.75 [12], ηb: Battery efficiency = 0.85 [12].
The selected DC bus voltage is 24, therefore:
Amperehourcapacity (CAh)=Cwh/24≈579.5 Ah
(5)
Number of batteries = 2 batteries of 12 V and 300 Ah connected in series to give 12 hours operation.
5.3.4. Sizing of MPPT controller
Total maximum power of PV (Pmax-tot) = Pmax×NPV = 250 × 12 = 3000 watts
(6)
The maximum power of controller (Pmax-C) and number of controller (NC) can be calculated as [15]:
Pmax-C = Battery voltage (Vb) × Controller current (IC) = 24 × 60 = 1440 watts
(7)
NC=Pmax-tot/Pmax-C= 3000/1440 ≈ 2 controllers
(8)
5.3.5. Life Cycle Cost (LCC) analysis
The LCC of the PV solar system can be classified into three parts: (ⅰ) components cost (ⅱ) installation cost (ⅲ) operation and maintenance cost. The details of calculating the components cost of the proposed PV solar system are represented in Table 5.
Table 5.
Cost calculations of proposed PV solar system.
where: i is the inflation rate and assumed to be 3% and d is the interest rate assumed to be 9%. Therefore, the total life cycle of the PV system can be calculated as [18]:
LCC=CPV+CC+Cinv+Cb+Cins+CMPW= 4,386,366 IQD
(10)
5.3.6. Evaluation of results
The Iraqi government has not given much attention to use of PV solar energy for solving the problems in the electric power sector. In this paper, the cost and an economic feasibility investigation of proposed PV solar system was analyzed by calculating the life cycle cost. A comparison between the proposed system and two alternative sources (generators) that used together in Iraqi house was investigated. According to the results analysis, the total cost of the financial burden in five years for neighborhood generator was found to be (1,920,000 IQD ≈ 1600 USA$) while the cost of household generator was found to be (3,080,000 IQD ≈ 2566.6 USA$). Totally, the cost of the two generators was recorded about (5,000,000 IQD ≈ 4166.6 USA$) in five years. As shown in Figure 4, the cost of household generator represents 61.6% from the total cost of the two generators. The cost of neighborhood generator increases with an increasing in number of amperes that the Iraqi family needed. While, the cost of household generator depend on two parameters (operations hours and maintenance cost). The operation hours of household generator increase in the event of breakdowns of the neighborhood generator causing an increase in the cost of fuel and oil consumption and maintenance. On the other side, the life cycle cost LCC of the proposed PV solar system is found to be (4,386,366 IQD ≈ 3655 USA$). As shown in Figure 5, the cost of PV system is less than the cost of two generators by (613,634 IQD ≈ 511.3 USA$). In the other words, the LCC of two generators is greatly higher than PV solar cost with 12.27%.
According to the economic feasibility analysis of the PV solar system comparing with the two types of generators, the following conclusions can be made:
1. The cost of household generator is greater than the neighborhood generator about 42.6%;
2. The breakdowns in the neighborhood generator increase the operation hours of the household generator then increase the fuel, oil and maintenance costs;
3. According to LCC, the cost of two generators is greatly higher than PV solar cost with 12.27%;
4. The PV solar system is more economic in Iraqi houses under the deterioration in the electricity sector. Furthermore, the PV solar not pollutes the environment as the case of using household and neighborhood generators.
Acknowledgements
The authors would like to thank the staff of Baqubah Technical Institute and Technical Engineering collage for their help.
Conflict of interest
All authors appear no conflicts of interest.
References
[1]
Murthy Balijepalli VSK, Pradhan V, Khaparde SA, et al. (2011) Review of demand response under smart grid paradigm. ISGT2011-India, Kollam, Kerala, 236–243.
[2]
Faria P, Vale Z (2011) Demand response in electrical energy supply: An optimal real time pricing approach. Energy 36: 5374–5384. doi: 10.1016/j.energy.2011.06.049
[3]
Moslehi K, Kumar R (2010) A reliability perspective of the smart grid. IEEE Trans Smart Grid 1: 57–64. doi: 10.1109/TSG.2010.2046346
[4]
Chan SC, Tsui KM, Wu HC, et al. (2012) Load/price forecasting and managing demand response for smart grids: Methodologies and challenges. IEEE Signal Process Mag 29: 68–85. doi: 10.1109/MSP.2012.2186531
[5]
Palma, Wilfredo (2016) Book: Time series analysis.
[6]
Raza MQ, Khosravi A (2015) A review on artificial intelligence based load demand forecasting techniques for smart grid and buildings. Renewable Sustainable Energy Rev 50: 1352–1372. doi: 10.1016/j.rser.2015.04.065
Hyde O, Hodnett PF (1997) An adaptable automated procedure for short-term electricity load forecasting. IEEE Trans Power Syst 12: 84–94. doi: 10.1109/59.574927
[9]
Broadwater RR, Sargent A, Yarali A, et al. (1997) Estimating substation peaks from load research data. IEEE Trans Power Delivery, 12: 451–456.
[10]
Huang SR (1997) Short-term load forecasting using threshold autoregressive models. IEE Proc- Gener, Transm Distrib 144: 477–481. doi: 10.1049/ip-gtd:19971144
[11]
El-Keib AA, Ma X, Ma H (1995) Advancement of statistical based modeling techniques for short-term load forecasting. Electr Power Syst Res 35: 51–58. doi: 10.1016/0378-7796(95)00987-6
[12]
Chen J, Wang W, Huang C (1995) Analysis of an adaptive time-series autoregressive moving-average (ARMA) model for short-term load forecasting. Electr Power Syst Res 34: 187–196. doi: 10.1016/0378-7796(95)00977-1
[13]
Barakat EH, Qayyum MA, Hamed MN, et al. (1990) Short-term peak demand forecasting in fast developing utility with inherit dynamic load characteristics. I. Application of classical time-series methods. II. Improved modelling of system dynamic load characteristics. IEEE Trans Power Syst 5: 813–824.
[14]
Taylor JW (2003) Short-term electricity demand forecasting using double seasonal exponential smoothing. J Oper Res Soc 54: 799–805. doi: 10.1057/palgrave.jors.2601589
[15]
Hyndman RJ, Fan S (2010) Density forecasting for long-term peak electricity demand. IEEE Trans Power Syst 25: 1142–1153. doi: 10.1109/TPWRS.2009.2036017
[16]
Badri A, Ameli Z, Birjandi AM (2012) Application of artificial neural networks and fuzzy logic methods for short term load forecasting. Energy Procedia 14: 1883–1888. doi: 10.1016/j.egypro.2011.12.1183
[17]
Li D, Chang C, Chen C, et al. (2012) Forecasting short-term electricity consumption using the adaptive grey-based approach-An Asian case. Omega 40: 767–773. doi: 10.1016/j.omega.2011.07.007
[18]
Yang HY, Ye H, Wang G, et al. (2006) Fuzzy neural very-short-term load forecasting based on chaotic dynamics reconstruction. Chaos Solitons Fractals 29: 462–469. doi: 10.1016/j.chaos.2005.08.095
[19]
Al-kandari AM, Soliman SA, El-hawary ME (2004) Fuzzy short-term electric load forecasting. Int J Electr Power Energy Syst 26: 111–122. doi: 10.1016/S0142-0615(03)00069-3
[20]
Smith M (2000) Modeling and short-term forecasting of new South Wales electricity system load. J Bus Econ Stat 18: 465–478.
[21]
Amina M, Kodogiannis VS, Petrounias I, et al. (2012) A hybrid intelligent approach for the prediction of electricity consumption. Int J Electr Power Energy Syst 43: 99–108. doi: 10.1016/j.ijepes.2012.05.027
[22]
Hsu C, Chen C (2003) Applications of improved grey prediction model for power demand forecasting. Energy Convers Manage 44: 2241–2249. doi: 10.1016/S0196-8904(02)00248-0
[23]
Fiot J, Dinuzzo F (2018) Electricity demand forecasting by multi-task learning. IEEE Trans Smart Grid 9: 544–551. doi: 10.1109/TSG.2016.2555788
[24]
Gonzalez-Romera E, Jaramillo-Moran MA, Carmona-Fernandez D (2006) Monthly electric energy demand forecasting based on trend extraction. IEEE Trans Power Syst 21: 1946–1953. doi: 10.1109/TPWRS.2006.883666
[25]
Zahedi G, Azizi S, Bahadori A, et al. (2013) Electricity demand estimation using an adaptive neuro-fuzzy network : A case study from the Ontario province-Canada. Energy 49: 323–328. doi: 10.1016/j.energy.2012.10.019
[26]
Dudek G (2015) Short-Term load forecasting using random forests. IEEE Conf Intell Syst 821–828.
[27]
Chaturvedi DK, Sinha AP, Malik OP (2015) Short term load forecast using fuzzy logic and wavelet transform integrated generalized neural network. Int J Electr Power Energy Syst 67: 230–237. doi: 10.1016/j.ijepes.2014.11.027
[28]
Ryu S, Noh J, Kim H (2016) Deep neural network based demand side short term load forecasting. Energies 10: 1–20. doi: 10.3390/en10010001
[29]
Shevade SK, Keerthi SS, Bhattacharyya C, et al. (2000) Improvements to the SMO algorithm for SVM regression. IEEE Trans Neural Networks 11: 1188–1193. doi: 10.1109/72.870050
[30]
Mashor MY (2000) Hybrid training algorithm for RBF network. Int J Comput Internet Manage 8: 50–65.
Rasmussen CE (2004) Gaussian processes in machine learning. Adv Lect Mach Learn 63–71.
This article has been cited by:
1.
Rômulo de Oliveira Azevêdo, Paulo Rotela Junior, Luiz Célio Souza Rocha, Gianfranco Chicco, Giancarlo Aquila, Rogério Santana Peruchi,
Identification and Analysis of Impact Factors on the Economic Feasibility of Photovoltaic Energy Investments,
2020,
12,
2071-1050,
7173,
10.3390/su12177173
2.
Noor Mahmood, Ahmed Ajmi, Shamsul Sarip, Hazilah Kaidi, Khairur Jamaludin, Hayati Talib,
Modeling the Sustainable Integration of Quality and Energy Management in Power Plants,
2022,
14,
2071-1050,
2460,
10.3390/su14042460
3.
Omar Mohammed Ali, Omar Rafae Alomar, Shaaban Ibrahim Mohamed,
Technical, economical and environmental feasibility study of a photovoltaic system under the climatic condition of north Iraq,
2022,
0143-0750,
1,
10.1080/01430750.2022.2126004
4.
Jehan Rasheed Majeed, Ali Nasser Hussain, Abdulrahman Th. Mohammad,
2023,
Design Off-Grid Solar Water Pumping System as a Replacement to Petrol Pump in Remote Area,
979-8-3503-4440-0,
108,
10.1109/ICESAT58213.2023.10347290
5.
Omar Rafae Alomar, Noor Moneer Basher, Omar Mohammed Ali, Abdulrahman Salih, Nabeel M. Abdulrazzaq, Sarkaft M. Samad,
Energetic, Economic Environmental Analysis for Photovoltaic Grid-Connected Systems under Different Climate Conditions in Iraq,
2025,
27727831,
100180,
10.1016/j.cles.2025.100180
6.
Nurul Hasanah Johari, Wesam Salah Alaloul, Muhammad Ali Musarat,
Recent advancements of life cycle cost analysis of photovoltaic systems: a systematic review,
2025,
0948-3349,
10.1007/s11367-025-02474-3