A Practical Approach for Predicting Power in a Small-Scale Off-Grid Photovoltaic System using Machine Learning Algorithms

Table 7

Methods of analysing meteorological data.

Ref.	Paper title	Type of meteorological data	Method	Inference
[28]	Transformer failure diagnosis using fuzzy association rule mining combined with case based reasoning	—	Hybrid	CBR algorithm is employed to discover the past cases which are the most identical to the present case. The exactness of the model is confirmed using past 10 years data.
[69]	A hybrid algorithm for short-term solar power prediction - sunshine state case study	Hourly	Hybrid	Results obtained from the hybrid algorithm are more accurate with fast convergence compared to the classic algorithm.
[70]	A hybrid ensemble model for interval prediction of solar power output in ship onboard power systems	Hourly	Hybrid	The hybrid algorithm gives outcomes with high efficiency considering the meteorological data along with the ship’s swinging as the input parameters.
[71]	A lightweight short-term photovoltaic power prediction for edge computing	Data sampled every 30 minutes	Hybrid	Compared to other standard ML algorithms, the technique employed here is remarkable and is capable of making short-term power predictions.
[72]	A local training strategy-based artificial neural network for predicting the power production of solar photovoltaic systems	Hourly	Intuitive	Various tests were conducted that showed the superiority of the proposed ANN over the benchmark ANN training strategies.
[73]	A practicable copula-based approach for power forecasting of small-scale photovoltaic systems	Daily	Numerical	From the results, it is clear that the mathematical model used here gives satisfactory prediction for cloudy days.
[74]	A solar time based analog ensemble method for regional solar power forecasting	Hourly	Hybrid	The proposed model adapts easily to the changing weather conditions irrespective of the location with high forecasting accuracy, few parameter requirements, data management, etc.,
[75]	Ensemble approach of optimized artificial neural networks for solar photovoltaic power prediction	Daily	Intuitive	Accurate day-ahead power prediction is obtained and is verified against a real case study. The number of hidden neurons in the hidden layer of ANN is optimized using trial and error method of the proposed model.
[76]	Photovoltaic power forecasting with a hybrid deep learning approach	Daily	Hybrid	The proposed hybrid method is compared with three other benchmark methods and is shown to have very small prediction errors.
[77]	Power generation forecast of hybrid PV-wind system	4 hours daily	Numerical	The duration for which data samples are incomplete or missing can be predicted by using the proposed method.
[78]	Prediction of photovoltaic power generation based on general regression and Back propagation neural network	Daily	Numerical	Temperature and irradiance were found to be the key parameters. Back propagation neural network predicted accurate results, but general regression technique was more appropriate for big data sets.
[79]	Probabilistic forecasting of photovoltaic generation: An efficient statistical approach	Daily	Probabilistic	The technique employed here exhibits very high computational efficiency and proves to be remarkably effective.
[80]	Real-time anomaly detection for very short-term load forecasting	Daily	Numerical	A way to detect and replace anomalies/corrupted data is proposed here whose performance surpasses state-of-the-art methods.
[81]	Day-ahead hierarchical probabilistic load forecasting with linear quantile regression and empirical copulas	Daily	Hybrid	A simple linear regression is adopted here for of accurate prediction improvement.
[82]	Direct quantile regression for nonparametric probabilistic forecasting of wind power generation	Hourly	Probabilistic	The proposed linear programming gives a simple solution with high computational efficiency and flexible framework.
[83]	Solar power probabilistic forecasting by using multiple linear regression analysis	Hourly	Numerical	The forecasting result was satisfactory using linear regression
[84]	IOT based online load forecasting using machine learning	Hourly	Numerical	Compared to all the ML algorithms, LR ML algorithm is found to be better as per the parameters is considered in this paper. Forecasting consumption of power for the next-hour is done using the online IoT platform.