Renewable energy sources are clean energy systems and very important tools in the fight against carbon emissions. However, Energy demand is very crucial in sizing electrical power generators. Thus, the need for energy demand data for effective sizing of renewable of energy systems. The energy demand data of the study area is first obtained on weekly basis. If the load demand is known, then a suitable size of an energy source will be implemented. Furthermore, to curb carbon emissions, there is the need to increase the increase more renewable sources of energy. Accordingly, the renewable energy source can only be implemented if the load being addressed is known.
In the case of a prosumer, [12] Getting details of what is consuming how much in the load side of an energy system is obtained using The Non-Intrusive Load Monitoring (NILM). This technique helps to get such detailed energy consumption information; However, it is based on the use of low sampling frequency (LSF) and High Sampling frequency (HSF) which the lack of suitable datasets for the training and the test of such techniques delayed their development. This method has adopted high sampling frequency measurement, improved synchronization of load in the off and on grid modes, and aggregate measurement was adopted for measuring energy demand by the load.
According to authors in [20], real-time datasets are necessary for more accurate analysis, simple demand response measure implementation, and customized energy feedback for automated Load demands. To address this demand, the UK's industries produced the REFIT electrical load measurements dataset. Aggregate measurements of loads of the entire house as was also adopted in their research.
Meanwhile most of the authors just presented the need for dataset for efficient and reliable energy supply, discussion about appliances, lighting, and plug loads (ALP) that take-up the energy is also emphasized by [13]. In their argument some houses use heating and cooling appliances that need to be captured in the dataset. A methodology distinguishing non-electrically heated and cooled homes from a broad dataset of whole-house profiles was presented in their research. This only concentrated on the need for measure appliances but not the energy source themselves.
Similarly, identifying electrical appliances based on electrical measurements is of importance in demand-side management applications for the electrical power grid including automated load control, load scheduling and non-intrusive load monitoring (NILM). Therefore Plug-Load Appliance Identification Dataset (PLAID), a labelled dataset containing records of the electrical voltage and current of domestic electrical appliances obtained at a high sampling frequency (30 kHz) was measured by [14]. The dataset contains 1876 records of individually metered appliances from 17 different appliance comprising 330 different makes and models, and collected at 65 different locations in Pittsburgh, Pennsylvania (USA). Additionally, PLAID contains 1314 records of the combined operation of 13 of these appliance types (i.e., measurements obtained when multiple appliances were active simultaneously.
A smart energy system needs to provide intelligent feedback to make intelligent decisions this can be done through the non-intrusive load monitoring (NILM). However, most NILM datasets have real power readings, However, they are not user friendly for sophisticated analysis algorithms. Authors in [15] presented the Almanac of Minutely Power dataset (AMPds) for load disaggregation research. Finally, we results shows that load disaggregation algorithm shows that current, rather than real power, is a more effective measure for NILM.
On the other hand, a review by [16] assessed the state of the art for home energy load profile modeling, categorized the models, suggested future lines of inquiry and uses, listed the difficulties the researchers encountered in creating these models, and suggested potential fixes. Thus, possible solutions to the problems faced by modelers include developing and implementing an anonymization algorithm before any human has access to any measured datasets, implementing a crowd sourcing strategy that addresses citizens' privacy concerns, and increasing funding for public and measurement campaigns that encourage the installation of privacy-proof smart meters.
Furthermore, thorough examination of 42 NILM datasets with the help of comparison tables created to highlight the various characteristics of the datasets already in existence. Additionally, as a contribution to the field of energy disaggregation and load identification, the advantages and disadvantages of the current NILM datasets are emphasized together with a forecast for current difficulties and future research objectives [25]. Consequently, load demand data acquisition techniques can be summarized in Fig. 4.
Generation utilities utilize electric load forecasting tools to reduce overloads and congestion and enhance power flow on the transmission network. Distribution utilities would not be very interested in short-term electric load estimates because their distribution networks are mostly radial and have predictable maximum load needs. Because of this, the distribution systems are sized conservatively and are not greatly affected by transient changes in load. Whether an energy system uses nonrenewable or renewable energy, future load demand must be taken into account while sizing it.
Accurate load forecasting can help both electrical producing and distributing organizations make decisions about electricity, load switching, voltage management, network reconfiguration, and infrastructure development [2]. It also increases the reliability of power networks and decreases the cost of generating. The practice of projecting future electricity consumption is called load forecasting.
The ideal approach for load forecasting varies depending on the use case. For instance, econometric models are frequently used for medium-term load forecasting, and time series analysis is frequently used for short-term load forecasting. Long-term load forecasting, as well as short- and medium-term load forecasting, are all possible with artificial intelligence.
When sizing a renewable energy system, load forecasting is just one of numerous elements to take into account. Other costs include the cost of the renewable energy technology, storage costs, and backup power costs. The ideal system size will change based on the particular use case and the presumptions made regarding these variables. Renewable energy systems can be made more reliable and economically viable by using load forecasting and scaling strategies. This is critical as the proportion of renewable energy in the world's energy mix grows. A summary of literatures consulted while conducting this research are presented in Table 1.
Table 1
Summary of data acquisition, load forecasting and sizing literatures
S/N | Ref. | Authors | Title | Methodology | Findings | Limitations |
1 | [12] | | A measurement system for creating datasets of on/off-controlled electrical loads | Non-Intrusive Load Monitoring (NILM) | the high sampling frequency measurement capability (kHz), the control over the turn-on and off time instants in synchronization with the grid voltage sinusoid and the measurement with aggregation scenarios of up to six appliances | • MG in this case is considered Prosumer. • Based on the use of low sampling frequency (LSF) and High Sampling frequency (HSF). • lack of suitable datasets for the training and the test of such techniques |
2 | [17] | D. Murray, L. Stankovic, and V. Stankovic. | An electrical load measurements dataset of United Kingdom households from a two-year longitudinal study | REFIT electrical load measurements dataset | The dataset comprises 1,194,958,790 readings, that represent over 250,000 monitored appliance uses | • To design such services, train and validate models, access to data that resembles what is expected of smart meters, collected in a real-world setting. • Data collected for 20 houses |
3 | [13] | | A method for distinguishing appliance, lighting and plug load profiles from electricity ‘smart meter’ datasets | Appliances, lighting, and plug loads (ALP) | This research addresses these loads with a new method of distinguishing non-electrically heated and cooled homes from a broad dataset of whole-house profiles. | • Require accurate and representative ALP electricity profiles as an input. • Available datasets through electricity smart-metering usually consist only of whole-house electricity load and lack summary household characteristics |
4 | [18] | E. Aguilar Madrid and N. Antonio | Short-Term Electricity Load Forecasting with Machine Learning | Short term load forecasting-based machine learning (ML) models | Extreme Gradient Boosting Regressor (XGBoost) algorithm showed the best results | XGBoost models are based on an ensemble of decision trees |
5 | [19] | A. I. Arvanitidis et al. | Enhanced Short-Term Load Forecasting Using Artificial Neural Networks | Artificial Neural Network (ANN) | This study work has segmented the smart home according to the similarities of their electrical load profiles, | Segmentation is done in proportion of energy usage per hour (%) |
6 | [20] | B. Zhao, X. Zhang, P. Li, K. Wang, M. Xue, and C. Wang, | Optimal sizing, operating strategy and operational experience of a stand-alone microgrid on Dongfushan Island | genetic algorithm (GA)-based method | GA was used to solve the sizing optimization problem with multiple objectives including the minimization of life-cycle cost, the maximization of renewable energy source penetration and the minimization of pollutant emissions | An optimal unit sizing method is presented for stand-alone microgrids. wind turbine generators, solar panels, diesel generators and battery storage units considered |
7 | [21] | U. Akram, M. Khalid, and S. Shafiq | Optimal sizing of a wind/solar/battery hybrid grid-connected microgrid system | two constraint-based iterative search algorithms are proposed for optimal sizing of the wind turbine (WT), solar photovoltaic (PV) and the battery energy storage system (BESS) . | The first algorithm, named as sources sizing algorithm, determines the optimal sizes of RE sources while the second algorithm, called as battery sizing algorithm, determines the optimal capacity of BESS | These algorithms are mainly based upon two key essentials, i.e. maximum reliability and minimum cost |
8 | [22] | F. A. Alturki and E. M. Awwad | Sizing and Cost Minimization of Standalone Hybrid WT/PV/Biomass/Pump-Hydro Storage-Based Energy Systems | whale optimization algorithm (WOA), firefly algorithm (FF) and particle swarm optimization (PSO) | The results indicate that the proposed hybrid WT/PV/biomass/pump-hydro storage energy system is environmentally and economically practical | standalone hybrid wind turbine (WT)/photovoltaic (PV)/biomass/pump-hydro-storage energy system was designed. Provide the load demand with an objective function of minimum cost of energy (COE) |
9 | [20] | B. Zhao, X. Zhang, P. Li, K. Wang, M. Xue, and C. Wang, | Optimal sizing, operating strategy and operational experience of a stand-alone microgrid on Dongfushan Island | genetic algorithm (GA)-based method | GA was used to solve the sizing optimization problem with multiple objectives including the minimization of life-cycle cost, the maximization of renewable energy source penetration and the minimization of pollutant emissions | An optimal unit sizing method is presented for stand-alone microgrids. wind turbine generators, solar panels, diesel generators and battery storage units considered |