A machine-learning surrogate model for optimising photovoltaic-thermal deployment in complex urban morphologies

REFERENCE	CONTEXT	OUTPUT INDICATORS	ALGORITHM OF ML	MODULE TYPE	MODELLING SOFTWARE	VALIDATION STRATEGY	ACCURACY METRIC	BEST ACCURACY	DATA TYPE	VARIABLE
Shin et al. (2022)	Building	power generation	ANN	BIPV	PVsyst, Solar Pro	Train/test	RMSE, MAE, R²	0.92 (R²)	Simulated, Measured	Solar irradiance
Serrano-Luján et al. (2022)	Building	power generation	GE, DE	BIPV	_	Train/test	Erel	1.55	Measured	Ambient Temperature, solar irradiation, relative outdoor humidity, wind speed
Yousif and Kazem (2021)	Building	power generation	ANN	PVT	_	Train/test	MSE, NMSE, MAE, R²	0.11 (NMSE)	Measured	solar irradiance, ambient temperature
Sulaiman et al. (2024)	Building	total active power	NN	PV	_	Train/test	RMSE, MAE, Standard deviation	0.7 RMSE	Measured	Ambient Temperature, Horizontal irradiation
Mohana et al. (2021)	Building	Power generation	LASSO, RF, LR, PR, XGBoost, SVM, NN	PV	_	k-fold cross-validation	MSE	0.9	Measured	External temperature, wind speed, Humidity
Chaibi et al. (2021)	Building	electrical and thermal efficiencies	ANN	PVT	_	Train/test	MAE	0.0078%	Measured	solar irradiance and the module temperature
Patel et al. (2022)	Urban	power generation	LR	PV	_	Train/test	MAPE	1.40%	Measured	Ambient temperature, Relative Humidity
Kassem and Othman (2022)	Building	power generation	MFFNN, CFNN, RBFNN, ENN	PV	Matlab	Train/test	R, MAE, RMSE	0.0021 (RMSE)	Dataset	Ambient temperature, relative Humidity, solar radiation, wind speed
Suanpang and Jamjuntr (2024)	Microgrid	power generation	LGBM, KNN	PV	_	Train/test	R², RMSE, MAE	0.84 R²	Measured	Solar irradiance, ambient temperature
Cao et al. (2022)	Building	electrical efficiency	ANFIS, ANN, LS-SVR	PVT	_	Train/test	AARD, MSE, R²	0.95 R²	Dataset	radiation intensity, Coolant material
Zazoum (2022)	Module	power generation	SVM, GPR	PV	_	Train/test	RMSE, MAE, R²	0.98 R²	Dataset	Panel temperature, ambient temperature, relative humidity,
Zhang et al. (2021)	Module	power generation	NN	PV	_	Train/test	MSE	0.01	Measured	Radiation, Ambient temperature, Humidity, Wind speed, Evaporation
Pham and Tran (2023)	Building	power generation	KNR, LASSO, SVR, RF, ETR, GBR, XGBoost, ANN	PV	_	Train/test	MAE, RMSE	0.60 (RMSE)	Dataset	Radiation, Ambient temperature, Humidity, Wind speed, Evaporation, Rooftop dimension
Gharaee et al. (2024)	Module	electrical efficiency	MLP, RF, SVR	PVT	_	Train/test	RMSE, R²	0.76 (R²)	Measured	mass flow rate, solar radiation, ambient temperature, wind speed, fluid inlet temperature, PVT surface area, pipe inner diameter
Zhou et al. (2020)	power plant	power generation	ELM, GA, SDA	PV	_	Train/test	R², MAE, nRMSE	0.59 (R²)	Dataset	daily maximal, minimal and averaged temperature, daily averaged global horizontal radiation, daily averaged diffusive horizontal radiation
Rojek et al. (2023)	Building	Power generation, CO₂ reduction	ANN	PV	_	Train/test	RMSE	0.01	Dataset	air temperature, wind speed, cloudiness, Current power direction
Alghamdi et al. (2023)	Module	power generation	DNN	PVT	_	Train/test	MSE	3.34E-08	Measured	Cell type
Scott et al. (2023)	Building	power generation	RF, NN, SVM, LR	PV	_	Train/test	RMSE, MAPE	1.76 (RMSE)	Measured	weather and solar generation data
Tripathi et al. (2024)	power plant	power generation	MR, SVMR, GR	PV	_	Train/test	MSE, MAE, R²	0.88 (R²)	Measured	solar radiation, ambient temperature, relative humidity
Kabilan et al. (2021)	Building	power generation	ANN, DT, QSVM	BIPV	_	Train/test	RMSE, MSE, R², MAPE, MAE	0.88 (R²)	Measured	Building orientations
Elsaraiti and Merabet (2022)	power plant	power generation	LSTM, MLP	PV	_	Train/test	MAE, MAPE, RMSE, R²	0.77 (R²)	Dataset	weather and solar generation data
Asiedu et al. (2024)	power plant	power generation	ANN, XGBoost, RF, DT, KNN, LASSO, LR, RR	PV	_	Train/test	RMSE, MAE, R²	0.84 (R²)	Dataset	ambient temperature, module temperature, irradiation

Table 2

Selected parameters and their values.

VARIABLE PARAMETERS	RANGE OF VARIATION	STEP OF VARIATION	NUMBER OF CASES	DESCRIPTIONS
Horizontal Angle of panel	–20° to 20°	10°	5
Vertical angle of panel	15° above and below the latitude	5°	7
Distance Between Panels	1 to 5 meters	1 meter	5
Panel elevation from roof	10 to 30 centimetres	10 centimetres	3
Urban Block	–	–	3	In three different regions
Dataset variables number	5 × 7 × 5 × 3 × 3 = 1,575

Table 3

Type of algorithms and hyperparameters.

TYPE OF ALGORITHM	HYPERPARAMETER	NUMBER OF SCENARIOS
ANN	Number of Hidden Layers	1,2,3
	Number of Neurons per Layer	1 to 200 in steps of 10
	Activation Function	RELU
RF	Number of Decision Trees	1 to 500
RF	Max_depth	1 to 10 and None

Simulation results of the EP index for training data in different regions.

Simulation results of the EP index for test data in different regions.

Simulation results of the HWP for training data in different regions.

Simulation results of the HWP for test data in different regions.

Simulation results of the CR for training data in different regions.

Simulation results of the CR for test data in different regions.

Table 4

Specifications of ML Model Hyperparameters.

ALGORITHM	TARGET INDICATOR	OPTIMAL HYPERPARAMETER
RF n_estimators and max_depth	HW	none, 45
	Electricity	none, 100
	Number of panels	none, 45
	CO₂ reduction	none, 100
ANN hidden-layer-sizes	HW	(100,200)
	Electricity	(250,250)
	Number of panels	(100,200)
	CO₂ reduction	(100,200,200)

Table 5

Performance of ML Models Based on Data from Region 1.

ALGORITHM	TARGET INDICATOR	LEARNING TIME (s)	VALIDATION
			RMSE		MAE		R-SQUARE
			TEST	VALIDATION	TEST	VALIDATION	TEST	VALIDATION
RF	HW	0.08	41,643.7	2,500.2	3,229.2	576.04	0.99	0.99
	Electricity	0.07	2,098.4	9.473	137.6	3.5	0.99	0.99
	Number of panels	0.05	0	0	0	0	1	1
	CO₂ reduction	0.1	8,100.81	627.38	731.09	231.98	0.99	0.99
ANN	HW	63.26	75,972.91	41,369.06	101,439.63	101,292.02	0.99	0.99
	Electricity	25.64	32,675.6	34,486.9	30,534.08	32,395.7	0.93	0.91
	Number of panels	37.6	51.6	26	20.03	17.4	0.99	0.99
	CO₂ reduction	50.28	14,301.5	7,901.5	6,709.2	5,916.8	0.99	0.99

Table 6

Performance of ML Models Based on Data from Region 2.

ALGORITHM	TARGET INDICATOR	LEARNING TIME (s)	VALIDATION
			RMSE		MAE		R-SQUARE
			TEST	VALIDATION	TEST	VALIDATION	TEST	VALIDATION
RF	HW	0.06	403.85	1,056.83	181.1	369.3	0.99	0.99
	Electricity	0.05	31.3	20.9	16.1	8.7	0.99	0.99
	Number of panels	0.08	0	0	0	0	1	1
	CO₂ reduction	0.09	218.92	263.41	131.79	131.41	0.99	0.99

Table 7

Performance of ML Models Based on Data from Region 16.

ALGORITHM	TARGET INDICATOR	LEARNING TIME (s)	VALIDATION
			RMSE		MAE		R-SQUARE
			TEST	VALIDATION	TEST	VALIDATION	TEST	VALIDATION
RF	HW	0.09	752.53	1,663.24	363.66	597.99	0.99	0.99
	Electricity	0.07	38.27	23.31	17.94	4.5	0.99	0.99
	Number of panels	0.05	0	0	0	0	1	1
	CO₂ reduction	0.06	345.43	425.25	219.88	238.9	0.99	0.99

Comparison of Simulation and Predicted Results in Region 1, 2, 16.

Proposed framework for evaluation of PVT panels in urban context.

References

Authors

Metrics

Articles in this issue

DOI: https://doi.org/10.66506/essp.8-244-26 | Journal eISSN: 2001-7960

Journal RSS Feed

Language: English

Page range: 29 - 29

Submitted on: Mar 20, 2026

Accepted on: May 4, 2026

Published on: Jul 31, 2026

Published by: European Council for an Energy Efficient Economy (eceee)

In partnership with: Paradigm Publishing Services

Keywords:

photovoltaic-thermal (PVT) systems,

machine learning in renewable energy,

energy efficiency optimisation,

AI-driven energy prediction

© 2026 Alireza Nazeri, Ciara Ahern, published by European Council for an Energy Efficient Economy (eceee)
This work is licensed under the Creative Commons Attribution 4.0 License.

Volume 1 (2026): Issue 1

A machine-learning surrogate model for optimising photovoltaic-thermal deployment in complex urban morphologies

Figures & Tables

Table 1

Table 2

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Table 3

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Table 4

Table 5

Table 6

Table 7

Figure 13

Figure 14

Figure 15

Paradigm

My account