=Paper=
{{Paper
|id=Vol-3611/paper17
|storemode=property
|title=A novel artificial intelligence technique for enhancing the annual profit of wind farm
|pdfUrl=https://ceur-ws.org/Vol-3611/paper17.pdf
|volume=Vol-3611
|authors=Prasun Bhattacharjee,Rabin K. Jana,Somenath Bhattacharya
|dblpUrl=https://dblp.org/rec/conf/ivus/BhattacharjeeJB22
}}
==A novel artificial intelligence technique for enhancing the annual profit of wind farm==
https://ceur-ws.org/Vol-3611/paper17.pdf
A novel artificial intelligence technique for enhancing the
annual profit of wind farm
Prasun Bhattacharjee1,*,†, Rabin K. Jana2,† and Somenath Bhattacharya3,†
1
Jadavpur University, 188 Raja S.C. Mallick Road, Kolkata 700032, India
2
Indian Institute of Management Raipur, Sejbahar, Chhattisgarh 492015, India
3
Jadavpur University, 188 Raja S.C. Mallick Road, Kolkata 700032, India
Abstract
While climate change is triggering off calamitous aftermaths globally, wind energy offers an apposite alternate to conventional
fossil fuels for abating greenhouse gas emanations. Economic profitability is an important factor for the green transformation
of electricity generation businesses for achieving carbon neutrality as proposed in the Paris agreement of 2015. The current
research aspires to expand the annual profit of wind farms employing an adapted genetic algorithm. A dynamic tactic
for allotting the crossover and mutation factors has been utilized to quantify their proportional proficiency. A randomly
chosen variable wind flow pattern has been employed for calculating the annual profit of wind farms. The research inferences
validate the higher competence of escalating mutation and crossover possibilities tactic for expanding the annual profit of
wind farms with two arbitrarily selected terrain settings.
Keywords
Annual profit maximization, crossover, genetic algorithm, mutation, wind farm
1. Introduction crashed dramatically over the earlier few decades trans-
nationally[5]. Researchers from every corner of the globe
The never-ending release of Green House Gases (GHG) are uninterruptedly endeavoring to boost the profitabil-
into the air is swelling the air temperature and atypical ity of WPG industries to support nations in achieving
meteorological conditions triggering the macro-climate their carbon neutrality goals as quickly as feasible[6].
alteration of the planet[1]. Renewable energy proposes a Genetic Algorithm (GA) was utilized for wind power
proliferating alternative amid the ever-increasing inter- generation site design in Gökçeada islet [7]. Saroha
national trepidation for the constricted provision of fossil and Aggarwal [8] offered a simulation intended for
fuels and their perilous penalties on the atmosphere[2]. WPG guesstimate with GA and Neural Network
Astoundingly, the utilization of renewable power inflated (NN). An NN-empowered technique with Particle
by 3% in 2020, even though the requirement of non- Swarm Optimization (PSO) and GA has been
renewable fuels collapsed throughout the globe due to projected for WPG prognostication [9]. Roy and Das
pandemic-related restrictions[3]. [10] have exercised GA with PSO for WPG expenditure
Accompanied by low GHG production benefit, renew- minimization. A proportional study of GA and Binary
able power solutions like wind energy is necessitated PSO has been presented to curtail the WPG
to stay practicable by propositioning inexpensive gen- expenditure [11]. Although most of the studies focused
eration charge through greater consistency and nom- on reducing the WPG charge, more research needs to
inal cost of maintenance to expedite de-carbonization be aimed at expanding the financial sustainability of
of universal energy techniques to a greater degree wind energy ventures for fulfilling the 2015 Paris
[4]. The Wind Power Generation (WPG) expense has agreement commitments made by various governments
and global entities.
IVUS 2022: 27th International Conference on Information Technology, This research purposes to realize the maximum annual
May 12, 2022, Kaunas, Lithuania profit of WPG farm for a randomly generated wind flow
*
Corresponding author. pattern and two arbitrarily selected layout settings. Be-
†
These authors contributed equally. cause of the intricacy of the WPG process, conventional
$ prasunbhatta@gmail.com (P. Bhattacharjee);
optimization tactics are inept to manage such conditions.
rkjana1@gmail.com (R. K. Jana); snb_ju@yahoo.com
(S. Bhattacharya) Artificial Intelligence (AI) methods have been previously
https://www.researchgate.net/profile/Prasun-Bhattacharjee-2 engaged in miscellaneous technical fields and are apt for
(P. Bhattacharjee); the present optimization situation for their heftiness and
https://www.researchgate.net/profile/Rabin-Jana (R. K. Jana) prompt computing fitness[12, 13, 14, 15, 16].
0000-0001-9493-5883 (P. Bhattacharjee); 0000-0001-8564-112X
GA is a prominent AI-aided method emulating the
(R. K. Jana); 0000-0002-3286-5450 (S. Bhattacharya)
© 2022 Copyright for this paper by its authors. Use permitted under Creative Commons License process of organic predilection and ensuing the objective
Attribution 4.0 International (CC BY 4.0).
CEUR
CEUR Workshop Proceedings (CEUR-WS.org)
Workshop
Proceedings
http://ceur-ws.org
ISSN 1613-0073
of eminent computer scientist Alan Turing to form a
‘knowledge machinery’ impending the strategy of genetic
CEUR
ceur-ws.org
Workshop ISSN 1613-0073
Proceedings
development[17]. GA has been applied in the present decision-makers. The terrain settings have been graphi-
research accompanied by a proportional assessment of cally shown in Figs. 2 and 3.
two distinct procedures of choosing the probabilities of
crossover and mutation processes.
2. Problem construction
2.1. Objective function
The power generated by Wind Turbine (WT) can be ex-
pressed as follows.
1
𝑃𝑊 𝑇 = 𝜌𝐴𝜗3 𝐶𝑝 cos 𝜃 (1)
2
where 𝑃𝑊 𝑇 denotes the generated power, 𝜌 signifies the
density of air, 𝐴 represents the cross-sectional area, 𝑣 Figure 2: Layout 1 without obstacle
is the speed of the wind, 𝐶𝑝 is the Betz threshold value
and 𝜃 is the angular error of yaw[11, 18]. The current
research is dedicated to increasing the annual profit of a
WPG farm. The objective function can be formulated as
follows.
𝑓𝐴 = [𝑆𝑉 − 𝐺𝐶 ] × 𝑃𝑦𝑟 (2)
where 𝑓𝐴 denotes the yearly profit, 𝑆𝑉 signifies the mar-
keting value per unit of wind power, 𝐺𝐶 represents the
generation price per unit of wind energy and 𝑃𝑦𝑟 indi-
cates the wind power generated yearly. The generation
charge of wind power has been calculated as per the
function provided by Wilson et al.[19]. The randomly
generated airflow has been presented in Fig.1.
Figure 3: Layout 2 with an obstacle of 500 m x 500 m dimension
3. Optimization algorithm
GA has been employed in the current research to deter-
mine the optimal annual profit of the WPG farm for the
randomly selected wind flow pattern and two different
Figure 1: Considered randomly generated wind flow pat- layout settings. The algorithm has been briefly discussed
tern for evaluating the annual profit of wind farm as follows. GA has been employed in the current research
to determine the optimal annual profit of the WPG farm
for the randomly selected wind flow p attern a nd two
different layout settings. The algorithm has been briefly
2.2. Terrain settings discussed as follows[12].
Two arbitrarily selected terrain situations have been se-
1. Establish the basic factors like populace size, rep-
lected for evaluating the annual profit of the WPG system.
etition number, probabilities for crossover, and
One of the terrains is with no obstacle and another one
mutation.
has an obstacle within it. The presence of obstacles has
2. Organize the populace indiscriminately.
been considered to evaluate its effect on the profitabil-
ity of the wind farm and increase the practicability of 3. Calculate the suitability of all distinct chromo-
the simulation. Although the terrain settings selected somes.
for the current research are square, they can be easily 4. Accomplish the arithmetic crossover technique
modified to any rectangular shape as per the need of the as follows.
a) Choose a numeral arbitrarily between 0 non-linearly modifying method for assigning the pro-
and 1. If it is less than the chance of the portions of crossover and mutation procedures of the
crossover technique, suggest the parental GA-based wind farm design process. The values of di-
element. verse factors associated with the considered optimization
b) Stimulate the crossover activity. process have been exhibited in Table 1.
c) Reconsider the relevance of the descen-
dants. Table 1
d) If the successor is reasonable, adapt it into Values of different factors related to the proposed enhanced
the up-to-date populace. GA
5. Achieve the mutation method as follows.
Factor Deemed Value
a) Elect a numeral arbitrarily between 0 and 1.
If it is less than the chance of the mutation 𝑐1 0.3
tactic, suggest the parental chromosome. 𝑐2 0.4
b) Stimulate the mutation action. 𝑚1 0.04
𝑚2 0.05
c) Reconsider the fitness of the mutated units.
Populace Size 20
d) If the mutated unit is viable, adapt it into Highest Generation Count 50
the fresh populace. Static Crossover Factor 0.3
6. Measure the appropriateness of the novel units Static Mutation Factor 0.04
shaped by crossover and mutation methods.
7. Pick the most prominent result understanding The wake forfeiture is a significant feature for power
the keenness of the choice-maker. generation from WT as it reduces the accessible kinetic
Accompanied by the established system of consider- energy of the wind of the in-line WTs. To curtail the
ing constant values, this research work has applied an disadvantageous outcome of wake damage, a fixed gap
innovative dynamic procedure for assigning the factors is essential to be kept between two in-line WTs for wind
of crossover and mutation. The dynamic crossover prob- farm design. The conditions of the WT have been offered
ability has been computed as follows. in Table 2.
{︃ (︂ )︂(3/2) }︃
𝑅𝑖 Table 2
𝑐𝑖 = 𝑐1 + (𝑐2 − 𝑐1 ) (3) Factors associated to WT
𝑅𝑚𝑎𝑥
Parameter Value
where 𝑐𝑖 is the non-linearly rising crossover possibility.
𝑐1 and 𝑐2 are the bounds of the crossover proportion. Output 1500 W
Blade Radius 38.5 m
𝑅𝑖 is the present recurrence count and 𝑅𝑚𝑎𝑥 represents
Inter-WT Gap 308 m
the uppermost reiteration count. The dynamic mutation Minimum Operational Wind Speed 12 km/hr
probability has been calculated as follows. Maximum Operational Wind Speed 72 km/hr
{︃ (︂ )︂(3/2) }︃ Capital Expenditure per WT USD 750,000
𝑅𝑖 Expense per Sub-Station USD 8,000,000
𝑚𝑖 = 𝑚1 + (𝑚2 − 𝑚1 ) (4)
𝑅𝑚𝑎𝑥 Yearly Operational Expenditure USD 20,000
Interest 3%
where 𝑚𝑖 is the non-linearly growing mutation possibil- Probable Life 20 years
ity. 𝑚1 and 𝑚2 are the bounds of the mutation propor- WT per Sub-Station 30
tion.
The optimal placements of WTs for Layout 1 using the
novel dynamic and conventional static approach for allo-
4. Results and discussion cating the factors of crossover and mutation processes
GAs have been utilized abundantly in the wind farm have been shown graphically in Figs. 4 and 5 respec-
designing process. They recommend a noticeable and tively. This terrain has no obstacle within its boundaries.
acknowledged paradigm when contrasted with other op- The possible locations for placing WTs has been marked
timization processes from the realm of artificial intelli- with circular red marks. The optimal placements of WTs
gence. The purpose of the existing research is to expand for Layout 2 using the novel dynamic and conventional
the annual profit o f wind farms. T he vending charge static approach for allocating the factors of crossover and
of wind energy has been considered as USD 0.033/kWh. mutation processes have been shown graphically in Figs.
Accompanied by the deliberation of the standard static 6 and 7 respectively. This layout has an obstacle of 500 m
method, the current study has considered an innovative x 500 m dimension within its terrain. The optimization
algorithms have been programmed to avoid placing any
WT within the boundaries of the obstacle.
Figure 7: Optimal placement of WTs for layout 2 using the
conventional static approach for allocating the factors of
crossover and mutation processes of GA
Figure 4: Optimal placement of WTs for layout 1 using
the novel dynamic approach for allocating the factors of
crossover and mutation processes of GA Relative assessments of the optimal yearly profits and
quantity of WTs accomplished by all methods of assign-
ing the possibilities of crossover and mutation procedures
of GA for both of the terrain designs have been offered
in Table 3 and Table 4 respectively.
Table 3
Comparison of optimal yearly profit obtained using both
optimization approaches
Optimization Process Layout 1 Layout 2
Static Approach USD 22,149 USD 21,845
Novel Dynamic Approach USD 22,479 USD 22,322
Figure 5: Optimal placement of WTs for layout 1 using the
Table 4
conventional static approach for allocating the factors of
Comparison of optimal count of WTs obtained using both
crossover and mutation processes of GA
optimization approaches
Optimization Process Layout 1 Layout 2
Static Approach 94 93
Novel Dynamic Approach 93 87
The study results validate the preeminence of the pro-
jected novel dynamic approach of assigning crossover
and mutation factors over the established static tactic
for both designs as it achieved the higher annual profit
with lesser WTs as specified in Table 3 and Table 4. The
increased cost-effectiveness of the wind farm can allow
the enhanced sustainability of the WPG ventures and
assist the progression of GHG discharge control for the
power generation businesses.
Figure 6: Optimal placement of WTs for layout 2 using
the novel dynamic approach for allocating the factors of
crossover and mutation processes of GA 5. Conclusion
Global organizations are continually attempting in the
direction of reduction of carbon trails by efficient appli-
cation of renewable sources like wind power as planned
by the Paris treaty of 2015. This study concentrates on [8] S. Saroha, S. Aggarwal, Multi step ahead forecasting
amplifying the yearly profit of wind farms through an of wind power by genetic algorithm based neural
innovative dynamic approach for allocating the crossover networks, in: 2014 6th IEEE Power India Interna-
and mutation factors. The optimization results confirm tional Conference (PIICON), IEEE, 2014, pp. 1–6.
the enhanced suitability of the novel dynamic technique [9] D. T. Viet, V. V. Phuong, M. Q. Duong, Q. T. Tran,
over the typical static method for improving the WPG Models for short-term wind power forecasting
site designs with the highest yearly profit. The projected based on improved artificial neural network using
method can aid the WPG trades to plan a reasonably particle swarm optimization and genetic algorithms,
feasible wind farm with the realistic deliberation of nu- Energies 13 (2020) 2873.
merous cost-allied factors and flexible airflow circum- [10] C. Roy, D. K. Das, A hybrid genetic algorithm (ga)–
stances. The present research can bring about impec- particle swarm optimization (pso) algorithm for
cable prospects for wind farm design enhancement and demand side management in smart grid consider-
economic sustainability of WPG systems for facilitating ing wind power for cost optimization, Sādhanā 46
the de-carbonization of the global power sector. (2021) 1–12.
[11] P. Bhattacharjee, R. K. Jana, S. Bhattacharya, A com-
parative analysis of genetic algorithm and moth
Acknowledgments flame optimization algorithm for multi-criteria de-
sign optimization of wind turbine generator bear-
The first author acknowledges the financial contribution
ing, ADBU Journal of Engineering Technology 10
of TEQIP section of Jadavpur University for aiding the
(2021).
current research work.
[12] R. Jana, P. Bhattacharjee, A multi-objective genetic
algorithm for design optimisation of simple and
References double harmonic motion cams, International Jour-
nal of Design Engineering 7 (2017) 77–91.
[1] B. Obama, The irreversible momentum of clean [13] A. Duggirala, R. Jana, R. V. Shesu, P. Bhattacharjee,
energy, Science 355 (2017) 126–129. doi:10.1126/ Design optimization of deep groove ball bearings
science.aam6284. using crowding distance particle swarm optimiza-
[2] P. K. Chaurasiya, V. Warudkar, S. Ahmed, Wind tion, Sādhanā 43 (2018) 1–8.
energy development and policy in india: A review, [14] P. Bhattacharjee, R. K. Jana, S. Bhattacharya, Multi-
Energy Strategy Reviews 24 (2019) 342–357. doi:10. objective design optimization of wind turbine blade
1016/j.esr.2019.04.010. bearing (2021). doi:10.5958/2454-762X.2021.
[3] International Energy Agency, The impact of the 00012.3.
covid-19 crisis on clean energy progress, 2020. URL: [15] P. Bhattacharjee, R. K. Jana, S. Bhattacharya, A
https://www.iea.org/. comparative study of dynamic approaches for al-
[4] G. Nicholas, T. Howard, H. Long, J. Wheals, locating crossover and mutation ratios for genetic
R. Dwyer-Joyce, Measurement of roller load, load algorithm-based optimization of wind power gen-
variation, and lubrication in a wind turbine gear- eration cost in jafrabad region in india (2021).
box high speed shaft bearing in the field, Tribology [16] P. Bhattacharjee, R. K. Jana, S. Bhattacharya, De-
International 148 (2020) 106322. sign optimization of simple harmonic and cycloidal
[5] R. Sitharthan, J. Swaminathan, T. Parthasarathy, motion cams (2021).
Exploration of wind energy in india: A short review, [17] A. M. Turing, Computing machinery and intelli-
in: 2018 National Power Engineering Conference gence, in: Parsing the turing test, Springer, 2009, pp.
(NPEC), IEEE, 2018, pp. 1–5. 23–65. doi:10.1093/oso/9780198250791.003.
[6] P. Bhattacharjee, R. K. Jana, S. Bhattacharya, A 0017.
relative analysis of genetic algorithm and binary [18] Z. Wu, H. Wang, Research on active yaw mecha-
particle swarm optimization for finding the optimal nism of small wind turbines, Energy Procedia 16
cost of wind power generation in tirumala area of (2012) 53–57.
india, in: ITM Web of Conferences, volume 40, EDP [19] D. Wilson, S. Rodrigues, C. Segura, I. Loshchilov,
Sciences, 2021, p. 03016. F. Hutter, G. L. Buenfil, A. Kheiri, E. Keedwell,
[7] S. Şişbot, Ö. Turgut, M. Tunç, Ü. Çamdalı, Optimal M. Ocampo-Pineda, E. Özcan, et al., Evolutionary
positioning of wind turbines on gökçeada using computation for wind farm layout optimization,
multi-objective genetic algorithm, Wind Energy: Renewable energy 126 (2018) 681–691.
An International Journal for Progress and Appli-
cations in Wind Power Conversion Technology 13
(2010) 297–306.