Revista Científica ‘‘INGENIAR”: Ingeniería, Tecnología e Investigación. Vol. 8 Núm. (15) 2025. ISSN: 2737-6249

Evaluación del Estado en la Central Termogas Machala a través Machine Learning.

DOI: https://doi.org/10.46296/ig.v8i15.0259

EVALUACIÓN DEL ESTADO EN LA CENTRAL TERMOGAS MACHALA A

TRAVÉS MACHINE LEARNING

CONDITION ASSESSMENT AT TERMOGAS MACHALA POWER PLANT

THROUGH MACHINE LEARNING

1

2

3

Cruz Néstor Xavier ; Quinatoa Carlos Iván ; Porras Jefferson

1

Universidad Técnica de Cotopaxi. Latacunga, Ecuador. Correo: nestor.cruz2@utc.edu.ec .

2

Universidad Técnica de Cotopaxi. Latacunga, Ecuador.

Correo: carlos.quinatoa7864@utc.edu.ec.

3

Universidad Técnica de Cotopaxi. Latacunga, Ecuador.

Correo: jefferson.porras0449@utc.edu.ec.

Resumen

En este estudio se examinó la situación de la Central Termogas Machala. El desafío del proyecto

consiste en superar grandes desafíos para asegurar la continuidad y asegurar un suministro

eficaz de energía eléctrica, así como el uso eficiente de los recursos naturales y la reducción del

impacto ambiental. La central termogas Machala opera en ciclo combinado, dispone de 8

unidades generadoras correspondientes a Machala I y Machala II, con una potencia total de 187

MW. Utilizando la programación en Python y la librería Pyomo para el proceso de optimización,

se pudo examinar las variables de costos de combustible, potencia y energía eléctrica de la

planta. La meta principal es reducir los costos de producción de energía eléctrica y las

limitaciones están vinculadas a los costos de inicio, parada y el equilibrio de potencia. Además,

para solucionar el problema se utiliza GNU Linear Programming Kit (GLPK), ya que el tipo de

programación sugerido es entero lineal mixta. Mediante el análisis efectuado, se pudo determinar

qué generadores térmicos pueden funcionar simultáneamente, elaborar planes de

mantenimiento para la salida programada de estos generadores y determinar la energía total

generada.

Palabras clave: Central Termogas, Ciclo Combinado, Pyomo, Python, Optimización.

Abstract

This study examined the situation of the Termogas Machala power plant. The challenge of the

project is to overcome major challenges to ensure continuity and ensure an efficient supply of

electricity, as well as the efficient use of natural resources and the reduction of environmental

impact. The Machala thermal power plant operates in a combined cycle, has 8 generating units

corresponding to Machala I and Machala II, with a total power of 187 MW. Using Python

programming and the Pyomo library for the optimization process, it was possible to examine the

variables of fuel, power and electric energy costs of the plant. The main goal is to reduce the

electrical energy production costs, and the constraints are linked to the startup, shutdown and

power balance costs. In addition, GNU Linear Programming Kit (GLPK) is used to solve the

problem, since the type of programming suggested is mixed linear integer. Through the analysis

carried out, it was possible to determine which thermal generators can operate simultaneously,

to develop maintenance plans for the scheduled output of these generators and to determine the

total energy generated.

Keywords: Combined cycle, Pyomo, Python, Optimization, Thermogas Plant.

Información del manuscrito:

Fecha de recepción: 14 de febrero de 2025.

Fecha de aceptación: 24 de abril de 2025.

Fecha de publicación: 10 de mayo de 2025.

560

Cruz et al. (2025)

1

. Introducción

concentrated in specific locations,

which we call power generation

plants [4].

In Ecuador, one of the regulatory

entities is the Agency for the

Regulation and Control of Energy

The

application

of

various

technological tools facilitates the

improvement and automation of

numerous manual processes that are

still carried out in an impractical

manner. Similarly, the use of data

obtained by management facilitates

the projection of the country's energy

consumption [5]. Machine Learning

is a field that encompasses various

disciplines of knowledge, including

Deep Learning, which offers a wide

variety of models and algorithms for

and

Non-Renewable

Natural

Resources (ARCERNNR) [1], whose

objective is to regulate various

strategic sectors of the nation.

Among these is the electricity sector,

which

directorates.

Control and Distribution of Electricity

Sector Commercialization

DCDCSE), which oversees the use

has

several

control

The Directorate of

(

of electricity-by-electricity distribution

companies nationwide. One of the

responsibilities of the directorate is

the billing process, which contains a

different purposes.

Time series

represent a challenge that can be

solved by intelligence algorithms.

These models focus on the ability to

train on a volume of data and then

predict values based on the training

data [6].

large amount of data.

Another

relevant factor that the management

considers is energy consumption

trends, this analysis optimizes the

control procedures carried out by the

management [2].

The procedure of estimating energy

intake in Ecuador's electricity sector

poses considerable challenges due

to the volume of data produced

monthly by the electricity distribution

companies [7], as well as the

requirement for accurate analysis

Nowadays, electrical energy is a very

common type of energy worldwide,

since it is used both in the industrial

sector and in most homes [3].

Electrical energy can be generated in

a variety of ways, so it cannot be

categorized as a renewable or non-

renewable energy source. However,

and projections.

Directorate of

Distribution of Electricity Sector

Although the

Control and

most

energy

production

is

561

Revista Científica ‘‘INGENIAR”: Ingeniería, Tecnología e Investigación. Vol. 8 Núm. (15) 2025. ISSN: 2737-6249

Evaluación del Estado en la Central Termogas Machala a través Machine Learning.

Commercialization (DCDCSE) has

access to an enormous volume of

data, manual handling and analysis

of this data is laborious and error [8].

Machine learning, through the

required calculations, can acquire

behavioral patterns and algorithms,

taking into account the PYOMO

Python library to solve optimization

challenges.

The challenge of the project is to take

on major challenges to maintain

continuity and ensure an efficient

supply of electricity, the efficient use

of natural resources and the

reduction of the impact on the

environment. The Machala thermal

power plant operates in a combined

cycle with a total capacity of 187 MW.

Currently, electricity consumption is

The objective of this study is to

establish through an analysis the

state estimation to optimize the

operation of the Machala thermal

power plant through the use of

machine learning. The objective is: to

establish the state estimation in the

Machala thermal power plant [11]; to

carry out the data collection of the

Machala thermal power plant; to

carry out a maintenance planning

and operation tests of the Machala

combined cycle thermal power plant,

by means of a machine learning

system [12].

steadily

increasing,

and

gas

shortages at the Machala thermal

power plant mean that there is not

enough gas to cover the maximum

production demand [9].

In contrast, today, electrical service

from Ecuador's power generation

plants has declined due to

generation shortages [2], lack of

2. Metodología

maintenance

and

facility

The process of construction and

improvement plans. The Machala

combined cycle thermoelectric plant

has the ability to convert thermal

energy from fuel gases into electrical

energy. This term is applied to plants

that use natural gases as fuel and

use gas and steam turbines to

produce electricity [10].

operation

of

the

Machala

Thermoelectric Power Plant began

on July 2, 1996, with the signing of

the

Energy

Development

Corporation (EDC) with the State of

Ecuador for the extraction of natural

gas in the Gulf of Guayaquil.

562

Cruz et al. (2025)

The Machala thermogas production

plant is located in the Bajo Alto sector

of the Tendales parish, Canton El

Guabo, Province of El Oro, as shown

in Figure 1.

The Machala Gas Plant has the

effective power detailed in Table I:

TABLE I. EFFECTIVE POWER

CENTRAL TERMOGAS MACHALA

Effective

Central

Unit

Power MW

64.6

20

6

FA1

FA2

Fig. 1. Machala Thermal Power Plant

Machala I

TM1

TM2

TM3

TM4

TM5

TM6

20

Machala II

19

Total

248.2

B.

Development

of

the

optimization model

Bender’s algorithm is employed to

solve the proposed optimization

problem, this algorithm facilitates the

solution of mixed integer nonlinear

problems [13].

A. Characteristics of the Machala

Thermogas Power Plant

The Termogas Machala power

station operates with gas obtained

from the Gulf of Guayaquil. Until

early 2011, the plant produced more

than 130 MW of energy that is

The GLPK tool, the GNU Linear

Programming Kit, an open-source

software intended for solving large

linear optimization problems and

supplied

to

the

National

Interconnected System (SNI) and

subsequently distributed to end

users.

mixed

integer

linear

mixed

programming problems, is used [14].

C. Problem master

The Machala thermoelectric plant

has two zones known as Machala 1

and Machala 2, where, as shown in

Figure 1, the 6FA natural gas

production units are located, along

with 6 TM2500 gas production units.

It is stated in equation (1).

^푀^푖^푛^푖^푚^푖^푧^푎^푟푥,휃 ^푓1⁽^ꢀ⁾⁺^ꢁ

subject to

푝(ꢀ) ≤ 0

563

Revista Científica ‘‘INGENIAR”: Ingeniería, Tecnología e Investigación. Vol. 8 Núm. (15) 2025. ISSN: 2737-6249

Evaluación del Estado en la Central Termogas Machala a través Machine Learning.

(

푣−1)

(ꢀ ꢄ

≥ 푓 ꢂ푌⁽^푣⁻¹⁾ꢃ + ∑^ꢅ훾_푘

푣−1)

)

information

acquired

by

the

ꢁ

(

푘

2

푘=!

푘

ꢀ

(1)

subproblem is transmitted to the

main Benders problem to get a better

interpretation of the original function,

as illustrated in Figure 2 [17].

ꢁ

≥ 0

Where:

•

ꢁ: continuous and positive

variable.

Fig. 2. Bender Decomposition information

•

v: iteration index of the algorithm.

exchange coupling process

(

푣−1)

ꢀ

: constant value taken by

푘

the variable x in the interaction v-

1.

(

푣−1)

:

•

훾_푘

cost

sensitivities

associated with the constraints

that set the value of the

complication variables.

푌⁽^푣⁻¹⁾: constant value taken by

the variable when solving the

Benders subproblem at iteration

v-1 [15].

D. End of iterations

The iterative process ends when the

lower and upper dimensions meet at

a point or are close to the limits. In

each iteration, the initial problem

dimensions are updated with the

resolution of the main problem and

When solving the main problem, the

value of the complication variables

푋⁽is obtained, as well as the value

of the cost close to the subproblem

푣)

Benders'

subproblem.

The

ꢁ⁽

The solution of the main

problem incorporates the Bender

Cuts procedure, which are

constraints that iteratively

^푣⁾.

procedure in the algorithm depicted

in the flow chart in Figure 3 [18].

reconstruct the initial problem

function [16].

The solution of the main problem and

the Benders subproblem demand a

coupling process, also known as

information

exchange.

The

564

Cruz et al. (2025)

Fig. 3. Benders decomposition flowchart

with the shutdown of the g-th

thermo-unit (1= is shutdown, 0=

is not shutdown).

F. Restrictions

The constraints are linked to a power

balance linked to the minimum and

maximum power limit of the

thermoelectric

generators.

In

addition, constraints are added by

generator startup and shutdown

costs, as discussed in equations 3

and 4 [20].

E. Function Objective

ꢈíꢅ

푡

ꢈá푥

≤ 퐸 ≤ 퐸

푔

퐸

(3)

푔

The objective function (J) is

represented in equation 2 [19].

Costs

퐶_표_ꢅ_푑× 푌 ≥ 퐶 ꢂ푌 ꢄ 푌^푡⁻¹ꢃ

푡

(4)

푔

^표^ꢅ푑

푔

_J_ꢆ_∑푇 퐶 × 푃 + 퐶 × 푌 + 퐶 ×

푡

퐴

푡

ꢇ

푔

푡=1

푇

푔

푡

(푊 ꢄ 푊^푡⁻¹)

푡

푔

푊

(2)

퐶_표_ꢉ_ꢉ_{_}_ꢊ× 푊 ≥ 퐶

푔

표ꢉꢉ__푑푔

Where:

•

U: binary coupling variable.

퐶_표_ꢅ_{_}_ꢊ: start-up cost of generating

퐶 : total production costs.

푇

units.

푡

푔

푃 : power generated by the g- th

thermal unit at time t.

퐶_표_ꢉ_ꢉ_{_}_ꢊ

:

cost of stopping the

퐴

•

퐶 : fixed cost of starting the g- th

푔

generating unit.

thermal unit.

푡

푔

푌 : binary variable associated with

푡

푔

푌 : binary variable associated

the coupling of the g-th thermal unit

with the coupling of the g-th

thermal unit (1= starts, 0= does

(1= starts, 0= does not start) [21].

not start).

푡

푔

푊 : binary variable associated with

ꢇ

•

퐶 : fixed cost of stopping the g-

푔

the shutdown of the g-th thermal unit

th thermo-unit.

(1= is shutdown, 0= is not shutdown).

푡

푊 : binary variable associated

푔

565

Revista Científica ‘‘INGENIAR”: Ingeniería, Tecnología e Investigación. Vol. 8 Núm. (15) 2025. ISSN: 2737-6249

Evaluación del Estado en la Central Termogas Machala a través Machine Learning.

In general, with respect to the

coupling logic constraints for each

thermal generator, it must be

considered that:

c) If the thermal unit is decoupled in

period (t-1) and coupled in period

(t) then the unit was started in (t).

d) If the thermal unit is decoupled in

period (t-1) and also decoupled

in period (t) then no start-up has

been performed [22].

a) If the thermal unit is coupled in

period (t-1) and also coupled in

period (t), the unit was already

operating in (t-1).

For the programming the data

expressed in tables 2 to 5 are used.

b) If the thermal unit is coupled in

period (t-1) and uncoupled in

period (t), the unit stopped in (t).

TABLE II. DEMAND REQUIRED PERIOD 2015-2022

YEAR

DEMAND [kW]

2

015

016

017

018

019

020

021

022

800000

560000

780000

890000

450000

560000

630000

660000

TABLE III. FUEL CONSUMPTION PER KWH

YEAR/TEAM

2015

2016

2017

2018

2019

2020

2021

2022

6

FA1

FA2

4894.55 4889.3 5362.7 5776.85 4161.13 4193.44 4097.86 2310.39

1343.29 4692.4 4964.54 4976.93 3929.96 1490.87 1569.73 1725.51

TM1

TM2

TM3

TM4

TM5

TM6

1667.91 1475.6 1274.79 621.01 617.99 1064.75

1356.64 1639.63 1145.69 589.72 337.86 553.61 277.54 357.43

1613.96 892.42 884.73 976.86 400.1 571.88 141.83 1065.48

17.6

0

1363.04 1538.02 1266.58 844.65 415.38 450.55 725.99 800.39

1053.41 940.94 987.83 263.06 406.54 802.06 1232.92 744.46

0

536.63 260.63

38.5

45.59

214.01

9.93

69.15

566

Cruz et al. (2025)

TABLE IV. FUEL COST (CTVS/KWH)

YEAR/TEAM

2015

2016

2017

2018

2019

2020

2021

2022

6

FA1

FA2

0.0355

0.0354

0.0353

0.0355

0.0354

0.0000

0.0356 0.0353 0.0354 0.0354 0.0354 0.0357 0.0359

0.0356 0.0354 0.0355 0.0354 0.0354 0.0358 0.0000

0.0354 0.0353 0.0354 0.0355 0.0354 0.0358 0.0360

0.0351 0.0352 0.0355 0.0355 0.0354 0.0358 0.0360

0.0354 0.0354 0.0356 0.0355 0.0354 0.0358 0.0360

0.0352 0.0352 0.0354 0.0355 0.0354 0.0358 0.0360

0.0354 0.0355 0.0355 0.0355 0.0354 0.0358 0.0360

0.0357 0.0360 0.0358 0.0357 0.0357 0.0358 0.0364

TM1

TM2

TM3

TM4

TM5

TM6

TABLE V. NET ENERGY [KWH]

YEAR/TEA

M

2

016

2017

2018

2019

2020

2021

2022

4

1

9

27093.4 440357.4 488657.1 518791.5 373367.2 371937.4 353363.7

6

FA1

FA2

2

8

86

16881.7 413944.2 437756.8 423601.9 325645.2 127765.2

36 28 71 66 49

8

65

67

59

0

9

6

24360.9 137677.2 117294.7 55594.87 55099.17 94424.80 1527.739

TM1

TM2

TM3

TM4

TM5

TM6

2

9

76

72

01

41

16

27

58781.9 154873.0 111942.1 54688.33 32153.88 48659.95

24353

2

5

18

26438.3 81491.33 79474.25 84667.30

36 63

23

7

18

89

50720.67 13038.41

37 84

35401.02

6

4

2

53224.1 143980.6 118878.2 75467.96 37558.70 40089.36 60928.74

59 66 24 92 57

26964.9 86230.25 94333.77 22936.11 37145.68 67843.25 100771.3

0

1

5

8

4

78

6792.55 48149.96 22075.10 2993.169

65 85 98

23

18

09

18

18768.23 6077.834

82 74

5

5.725097

9

5

3. Resultados y discusión

yielded the results described in Table

VI.

Mixed integer linear programming,

performed in Python-Pyomo with the

equations described in Section 2,

567

Revista Científica ‘‘INGENIAR”: Ingeniería, Tecnología e Investigación. Vol. 8 Núm. (15) 2025. ISSN: 2737-6249

Evaluación del Estado en la Central Termogas Machala a través Machine Learning.

TABLE VI. NELECTRIC POWER FROM EACH THERMAL GENERATOR AT TERMOGAS MACHALA

THERMAL POWER STATION

ENERGÍA ÓPTIMA [kWh]

AÑO/EQUI

T1

T2

T3

T4

T5

T6

T7

T8

PO

4

88657. 466398. 345299. 371937. 353363.

6

FA1

FA2

0

1

8

03

423601.

97

42

0

47

0

76

128321.

46

3

92957.

6

0

1

9

1

8

1

13940.

133805.

17

TM1

TM2

TM3

TM4

TM5

TM6

0

88681.9

0

2

41854. 158781. 154873.

32153.8 48659.9

3

2

92

02

8

6

2611.3

1

79996.9 78130.7

35401.0 50720.6

0

4

8

2

7

14269. 153224. 143980. 118878.

47543.4

2

0

4

6

1

66

27

94333.7

7

16394.

37145.6

8

100771.

35

0

3

9

5036.7 47344.2

30930.3

0

9

1

depending on the demand required,

the generators will be coupled, as

shown in the graph in Figure 4:

In order to optimally satisfy the power

demand of the plant, not all

generators

will

operate

simultaneously.

Therefore,

Fig. 4. Optimal energy Machala Thermogas Power Plant

568

Cruz et al. (2025)

With a total electrical generation of

50 MWh, generators 6FA1, TM2,

-

Period 1:

4

With a total electrical generation of

TM3 and TM5 are coupled with a

demand of 125 MW.

600 MWh, generators TM1, TM2,

TM3, TM4, TM5 and TM6 are

coupled, with a demand of 119 MW.

-

Period 7:

With a total electrical generation of

60 MWh, generators 6FA1, TM1,

-

Period 2:

5

With a total electrical generation of

00 MWh, generators 6FA2, TM2,

TM2 and TM3 are coupled, with a

demand of 125 MW.

8

TM4 and TM6 are coupled with a

demand of 124 MW.

-

Period 8:

With a total electrical generation of

30 MWh, generators 6FA1, 6FA2,

-

Period 3:

6

With a total electrical generation of

60 MWh, generators TM1, TM2,

TM4 and TM5 are coupled with a

demand of 170 MW.

5

TM3, TM4 and TM6 are coupled, with

a demand of 99 MW.

A.

Verification

The results are verified using the

GAMS payment software. Figure 5

shows the information of both the

problem and the SOLVE; 129

-

Period 4:

With a total electrical generation of

80 MWh, generators 6FA1, TM3,

7

TM4 and TM5 are coupled with a

demand of 125 MW.

variables

are

examined,

137

constants are considered and there

is an objective function.

-

Period 5:

With a total electrical generation of

890 MWh, generators 6FA1 and

6FA2 are coupled, with a demand of

130 MW.

-

Period 6:

569

Revista Científica ‘‘INGENIAR”: Ingeniería, Tecnología e Investigación. Vol. 8 Núm. (15) 2025. ISSN: 2737-6249

Evaluación del Estado en la Central Termogas Machala a través Machine Learning.

Fig. 5. Information on the problem posed.

4. Conclusiones

Currently, the Termogas Machala

power plant generates 630 MWh of

electricity through the operation of

thermal generators 6FA1, 6FA2,

TM4 and TM5. In 2022, the TM6

generator is not in operation and,

according to the research carried out,

it will be put into operation with the

expansion project by 2024.

The image in Figure 6 shows the

results obtained from the

The optimization was carried out

based on fuel costs and the energy

produced resulted in an ideal annual

average of 242.25 MWh, which

establishes an appropriate operation

of the Thermogas Power Plant, with

combined cycle. It is important to

note that the average fuel cost is

optimization process, the results

when compared with Table VI are

exactly the same.

Fig. 6. Results obtained

0.036 ctvs/kWh.

According to the results obtained, the

maintenance and tests to be carried

out will depend on the coupling or not

of certain thermal generators

according to the required demand.

For example, in period 6 the

generators 6FA1,TM2, TM3 and TM5

are coupled, while in the same period

they must be tested and the

corresponding maintenance and

tests must be carried out.

By implementing an algorithm using

the Python programming language

570

Cruz et al. (2025)

and the Pyomo library, it was

possible to establish the coupling of

heat generators according to the

necessary economic dispatch and

fuel cost.

Systems, vol. 19, pp. 360–

3

1

.

73, doi:

0.37394/232016.2024.19.31

2024,

[

3]

E. E. Ogar, S. Chaitusaney,

and W. Benjapolakul, “Solar

Power

Plant

Capacity

The use of Bender's decomposition

method is one of the many

alternatives for solving mixed linear

Monitoring Using Random

Forest Machine Learning

Algorithm,” 2024 IEEE/IAS

Industrial and Commercial

Power System Asia (I&CPS

Asia), pp. 67–72, 2024,

integer

optimization

problems,

although it is more frequently used in

education to explain the procedure

necessary to solve such problems.

Other algorithms can be used to

solve these problems, although the

one used in this study is less

complicated and more effective in

solving the problem.

[

Online].

Available:

https://api.semanticscholar.or

g/CorpusID:274468093

[4]

T. Haigler, “Power Plant

Condition

through

Assessment

Engineering,

Materials Science, and NDT

.0 Technology,” e-Journal of

Nondestructive Testing, 2022,

Online]. Available:

4

[

Bibliografía

https://api.semanticscholar.or

g/CorpusID:254775433

[

1]

P. Manke, S. Rungta, and S.

Bharti, “Forecast Load

[5]

G. S. Patil, U. S. Patil, and P.

Demand in Thermal Power

Plant with Machine Learning

Algorithm: A Review,” Electric

P. Shinde, “Enhancing Power

Transformer

Reliability

through Machine Learning-

Based Fault Prediction Using

Dissolved Gas Analysis,”

Power

Components

and

Systems, 2024, [Online].

Available:

2024

Third

International

https://api.semanticscholar.or

g/CorpusID:266916716

Conference on Power, Control

and Computing Technologies

(

ICPC2T), pp. 72–76, 2024,

[

2]

D. Pila, C. Quinatoa, L.

Camacho, and J. Vaca,

[

Online]. Available:

https://api.semanticscholar.or

g/CorpusID:268701307

“Transient Stability Analysis of

the Ecuadorian Electrical

System: Case of the Southern

[6]

J. I. Aizpurua, S. Mcarthur, B.

G. Stewart, B. Lambert, J. G.

Cross, and V. M. Catterson,

Segment,”

WSEAS

Transactions

on

Power

571

Revista Científica ‘‘INGENIAR”: Ingeniería, Tecnología e Investigación. Vol. 8 Núm. (15) 2025. ISSN: 2737-6249

Evaluación del Estado en la Central Termogas Machala a través Machine Learning.

“Adaptive Power Transformer

Generation Forecasting,” E3S

Lifetime Predictions Through

Web of Conferences, 2021,

Machine

Uncertainty

Nuclear Power Plants,” IEEE

Transactions on Industrial

Electronics, vol. 66, pp. 4726–

Learning

Modeling

and

in

[Online].

https://api.semanticscholar.or

g/CorpusID:241445677

Available:

[

11] C. Quinatoa, D. Albán, X.

Proaño, and L. Camacho,

4737,

2019,

[Online].

“

Development of an Algorithm

for Protection against

Atmospheric Discharges in

Buildings,” Revista

Available:

https://api.semanticscholar.or

g/CorpusID:59600342

[

7]

C. Pacheco et al., “Enhancing

Predictive Maintenance of

Power Transformers Through

Politecnica, vol. 55, no. 1, pp.

51–60, Feb. 2025, doi:

10.33333/rp.vol55n1.05.

Machine

Approaches,” Learning and

Nonlinear Models, 2024,

Online]. Available:

https://api.semanticscholar.or

g/CorpusID:273291560

Learning

[

12] R. Loganantharaj, G. Palm,

and M. Ali, “Intelligent problem

solvingꢀ: methodologies and

[

approachesꢀ:

13th

International Conference on

Industrial and Engineering

[

8]

M. S. Sachit, H. Z. M. Shafri,

A. F. Abdullah, A. S. M. Rafie,

and M. B. A. Gibril, “Global

Spatial Suitability Mapping of

Wind and Solar Systems

Using an Explainable AI-

Based Approach,” ISPRS Int.

J. Geo Inf., vol. 11, p. 422,

Applications

Intelligence

Systems, IEA/AIE 2000, New

Orleans, Louisiana, USA,

of

and

Artificial

Expert

June

19-22,

2000ꢀ:

proceedings,” 2000. [Online].

Available:

https://api.semanticscholar.or

g/CorpusID:61041622

2022, [Online]. Available:

https://api.semanticscholar.or

g/CorpusID:251145904

[

13] S. Boubaker, S. Kamel, N.

Ghazouani, and A. Mellit,

“Assessment of Machine and

Deep Learning Approaches

[

9]

“Predictive

Maintenance

Prediction Of

Beyond

Failures,” 2022. [Online].

Available:

https://api.semanticscholar.or

g/CorpusID:268113948

for

Photovoltaic Systems Using

Infrared Thermography,”

Remote. Sens., vol. 15, p.

686, 2023, [Online].

Available:

Fault

Diagnosis

in

1

10] D. Erlapally, Dr. K. Anuradha,

Dr. G. Karuna, V. Srilakshmi,

and K. Adilakshmi, “Survey

Analysis of Solar Power

https://api.semanticscholar.or

g/CorpusID:257674053

572

Cruz et al. (2025)

[

14] Y. Pan et al., “Broken Power

Strand Detection with Aerial

Images: A Machine Learning

based Approach,” 2020 IEEE

International Smart Cities

Conference (ISC2), pp. 1–7,

Module with Maximum Power

Output Predictingalgorithm

Based on Artificial Intelligence

Technology,” 2024 IEEE 52nd

Photovoltaic

Specialist

Conference (PVSC), p. 1536,

2024, [Online]. Available:

https://api.semanticscholar.or

g/CorpusID:274060797

2020, [Online]. Available:

https://api.semanticscholar.or

g/CorpusID:226293664

[

15] H. Moreno, C. Rueda-Ayala,

V. Rueda-Ayala, Á. Ribeiro, C.

Ranz, and D. Andújar,

[19] S. Roy, S. Tufail, M. Tariq, and

A. I. Sarwat, “Photovoltaic

Inverter Failure Mechanism

“

Machine Learning-Powered

Segmentation of Forage

Crops in RGB Imagery

Through Artificial Sward

Images,” Agronomy, 2025,

Online]. Available:

Estimation

Unsupervised

Learning

Assessment,” IEEE Trans

Reliab, vol. 73, pp. 1418–

1432,

Using

Machine

Reliability

and

[

2024,

[Online].

https://api.semanticscholar.or

g/CorpusID:276029075

Available:

https://api.semanticscholar.or

g/CorpusID:268454893

[

16] Dr. P. Deeskow, T. Kamiński,

and U. Steinmetz, “ANOMALY

DETECTION IN POWER

PLANT DATA OF A LARGE

[20] J. R. Vaca González, C.

Quinatoa, J. Ortiz, and L.

Camacho, “Evaluación de

modelos de optimización

convexos para minimizar

pérdidas en el sistema de

COAL

PLANT

FROM

FIRED POWER

EXPERIENCE

THE RECENT

–

PROJECTS,” 2018. [Online].

Available:

https://api.semanticscholar.or

g/CorpusID:89607781

distribución,”

Conectividad, vol. 5, no. 3, pp.

62–78, Jul. 2024, doi:

10.37431/conectividad.v5i3.1

2.

Revista

5

[

17] H. J. Mackenzie., “Short-Term

Forecasting of Wind Power

Plant Generation for System

Stability and Provision of

Ancillary Services,” 2017.

[21] A. and P. V. and C. A. and C.

L. and O. J. Quinatoa Carlos

and Chasi, “Optimization

Model

for

Coordinated

[

Online].

Available:

Multistage Planning of the

Generation-Transmission

https://api.semanticscholar.or

g/CorpusID:37876434

System

with

Demand

Forecasting Using Neural

Networks,” in Proceedings of

[

18] Y. H. Kim, J. Kim, and J. Jeon,

“How to Predict the PV

573

Revista Científica ‘‘INGENIAR”: Ingeniería, Tecnología e Investigación. Vol. 8 Núm. (15) 2025. ISSN: 2737-6249

Evaluación del Estado en la Central Termogas Machala a través Machine Learning.

the

4th

International

Conference on Electronic

Engineering and Renewable

Energy Systems—Volume 1,

A. and M. A. and R. A. and C.

M. Hajji Bekkay and Gagliano,

Ed., Singapore: Springer

Nature Singapore, 2025, pp.

573–581.

[

22] A. G. Ahungwa, C. Ioana, V.

Bouillet, B. Michel, A.

Bombenger, and P. Véras,

Machine Learning-Assisted

Operation Monitoring

“

Analytics on a Hydro Power

Plant,” 2024 6th Global

Power,

Communication Conference

GPECOM), pp. 461–471,

024, [Online]. Available:

Energy

and

(

2

https://api.semanticscholar.or

g/CorpusID:270974092

574