Compliance with environmental requirements defined by public agencies and private regulatory, monitoring and financing institutions is a matter of survival for enterprises, including thermoelectric plants (
Dogan & Tugcu, 2015
). CONAMA Resolution 382 (
Ministry of Environment of Brazil, 2007
) which defines maximum limits for atmospheric emissions from stationary sources and CONAMA Resolution 491 which establishes national air quality standards are examples of legislation in force in Brazil (
Ministry of Environment of Brazil, 2018
). The Atmospheric Dispersion Study (ADS) is a prerequisite for obtaining environmental licenses and aims to assess the impact of the new enterprise on air quality in communities located within its radius of influence.
The general objective of the research is to identify the best engine option among those evaluated, developing a decision support model using data from scenario simulations. Specifically, to simulate atmospheric dispersions, the following are intended: to analyze engine data specified in technical catalogs, such as emission rate, exhaust gas temperature, engine power, specific fuel consumption and exhaust gas exit velocity; to evaluate the topography of the project; to calculate the distance from the thermoelectric plant to the nearest communities; to examine local environmental legislation and apply current technical standards.
The guiding question of the research is the following: considering the current environmental legislation, is it possible to improve the evaluation of scenarios with combustion engines to achieve greater efficiency in the decision-making process, to meet the air quality standard in communities located within the radius of influence of the future project?
The basic hypothesis is that the development of a methodology for identifying and recognizing emission rate, power and specific fuel consumption patterns will contribute to improving the decision-making process for selecting the best scenario, resulting in projects that comply with current environmental legislation.
In this context, the following secondary hypotheses are considered: regarding atmospheric emission requirements, gas turbines (Brayton cycle) manufactured more than three decades ago can meet the standard defined in current environmental legislation (
Clementoni et al., 2014
), CONAMA Resolution 491, as long as they undergo retrofitting that includes water injection into the combustion chamber to reduce NOx emissions; and alternative internal combustion engines (Diesel cycle) have post-combustion solutions, treating and filtering combustion gases through a Selective Catalytic Reduction (SCR) system before being released into the atmosphere, in order to comply with the legislation (
Ahmadi et al., 2018
).
The research considered twelve models of combustion engines from different manufacturers. Two software programs were used: “Screen View” to simulate atmospheric dispersions (
Mehdizadeh & Rifai, 2004
) and the “R” program with its graphical interface “RStudio” which enabled the statistical treatment of the research data mass (
Komperda, 2017
).
In thermoelectric plants with combustion engines, these devices are the main sources of pollutant gas emissions. This study was limited to emissions of nitrogen oxides (NOx) because it requires a technological solution with a greater financial impact to ensure an acceptable level of emissions. The solution to control the NOx level also contributes to reducing emissions of other pollutants.
This study was also limited to the twelve combustion engine models presented in
Table 1
, as well as the climatic conditions of the northeastern region of Brazil, where the thermoelectric plant would be installed. However, the methodology presented in this work can be replicated with other engine models and other regions in the world, if the technical operational data of the new engines and the climatic conditions of the new regions are considered in the modeling.
2. Literature Review
According to (
Elzalik et al., 2020
), thermal power plants are characterized by converting the thermal energy released by the combustion of fossil fuels, biomass, or nuclear fission of uranium into electrical energy. The layout of thermal power plants depends on several factors, such as installed power, type of technology, and fuel used (
Favre-Perrod, 2005
).
Dedes et al. (2012)
confirm that atmospheric emission levels are also associated with the aforementioned factors. According to
Nemitallah et al. (2018)
, current models of gas turbines (Brayton cycle) have embedded combustion technology resulting in low emission levels, meeting the limits established in environmental legislation.
According to
Paraschiv et al. (2019)
, Gaussian-type models are widely used in atmospheric dispersion modeling and are based on a Gaussian distribution of the plume in the vertical and horizontal directions under steady-state conditions, particularly for regulatory purposes, with the same process used with Screen View.
Brusca et al. (2016)
confirm that the plume distribution is modified according to distance because of turbulent reflection from the earth’s surface and mixing height. The plume width is determined by stability classes or the travel time from the source exit to the receptor.
As presented by
Alnahdi et al. (2019)
Screen View uses a Gaussian plume model, but for a single source. This program can calculate emission concentrations up to 50 km away from the emission source for worst-case scenarios. It is capable of modeling dispersion pointwise, by area, or volumetrically, calculated through numerical and virtual integration of point sources. Atmospheric stability is calculated from Turner stability classes and uses the power law to correct wind speed at heights above 10 m (
Holmes & Morawska, 2006
).
As addressed by (
Zhong et al., 2011
), the Screen model code is composed of the Fortran language, which is widely used in scientific computing but is not suitable for graphics development. The Screen was originally designed to calculate ground-level concentrations under the plume centerline in the wind direction. Therefore,
Zhong et al. (2011)
proposed the use of some tools to extract and take advantage of all Screen’s benefits.
According to
Racine (2012)
, the “R” program is a popular open-source tool for statistical analysis and powerful features through its graphical interface “RStudio,” which produces diagrams and graphics with high quality. Typically, data mass is the result of research, simulations, modeling, or field or laboratory measurements (
Capelli et al., 2013
).
Dayal (2015)
adds that the “R” program works with data in objects known as datas frames. A data frame is an object with rows and columns, like a matrix. The rows contain different data observations to be studied or measurements from the experiment, called cases. The columns contain values of different variables, called fields. The values in the body of a matrix can only contain numbers, but the values in the body of a data frame can contain numbers and texts.
Issues related to atmospheric dispersion have important applications of linear regression techniques and analysis of variance (ANOVA) for predicting behavior of variables related to the theme (
Pongrac et al., 2020
). According to
Grömping (2015)
, the “R” program supports a flexible modeling language, using formulas and sharing functionalities with network graph functions.
3. Materials and Methods3.1. Analyzed Combustion Engines
The engines examined in this study are well-established equipment, with numerous practical application cases, manufactured by international and traditional companies.
Table 1
presents the symbols (abbreviations) that will be used throughout this article to represent the studied combustion engines and a brief description of each engine.
Diesel Cycle—Engine High Speed with Catalytic Reduction System
MB
Diesel Cycle—Engine Medium Speed
MB SCR
Diesel Cycle—Engine Medium Speed with Catalytic Reduction System
HE
Diesel Cycle—Engine Medium Speed
HE SCR
Diesel Cycle—Engine Medium Speed with Catalytic Reduction System
WC
Diesel Cycle—Engine Medium Speed
WC SCR
Diesel Cycle—Engine Medium Speed with Catalytic Reduction System
WC T2
Diesel Cycle—Engine Medium Speed
VD
Diesel Cycle—Engine High Speed
TF
Brayton Cycle—Gas Turbine without Water Injection
TF IA
Brayton Cycle—Gas Turbine with Water Injection
3.2. Construction of the Data Set
As presented in the introduction, a data set was structured using some technical and operational parameters, and the results of atmospheric dispersion were obtained through simulations.
Table 2
shows the data set with the parameters considered in this research (
Ni et al., 2020
), namely Tier (classification according to the emission levels established by the United States Environmental Protection Agency—EPA), Power (MWh), Rotation (rpm), Specific net consumption in liters/MWh, Wind speed (wind speed in m/s), distance until the community (distance in meters from the thermal power plant to the nearest community), Exhaust Temperature (temperature of gases leaving the chimneys in degrees Celsius), Exhaust Velocity (velocity of gases leaving the chimneys in m/s), Emission Rate (emission rate in g/kWh), Chimney Height (meters), and Dispersion (simulation result obtained with Screen in mg/Nm3) (
Turner, 2020
).
<xref ref-type="bibr" rid="scirp.140744-"></xref>Table 2. Spreadsheet containing part of the data used in the research (subset of the dataset).
Engines
Tier
Power
Rotation
Consumption
Wind Speed
CommunityDist
ExhaustTemp
SpeedExhau
EmissionRate
ChimHeight
Dispersion
Emis.RateQua
ChimHeightQua
DispersionQua
CA
0
3.25
1.800
254
2.1
1.800
474
35
8.0
15
1.200
Infeasible
Low
Bad
CA
0
3.25
1.800
254
1.5
1.800
474
35
8.0
15
1.150
Infeasible
Low
Bad
CA
0
3.25
1.800
254
3.5
1.800
474
35
8.0
15
1.250
Infeasible
Low
Bad
CA
0
3.25
1.800
254
2.1
1.000
474
35
8.0
25
100
Infeasible
Medium
Good
CA
0
3.25
1.800
254
2.1
1.800
474
35
8.0
25
400
Infeasible
Medium
Bad
CA
0
3.25
1.800
254
2.1
3.000
474
35
8.0
25
1.000
Infeasible
Medium
Bad
3.3. Screen View Software and R Program
In the data set presented above, the dispersion parameter was obtained through simulations using the Screen View software (
Espinosa et al., 2021
). This program was obtained for free by accessing the developer’s website at
https://www.weblakes.com/software/freeware/screen-view/
. It was necessary to provide the following input parameters to perform the simulations: emission rate, considering the maximum power of the thermal power plant, worst-case scenario, chimney height and diameter, velocity and temperature of gases leaving the chimney, and ambient temperature.
The output data from Screen View were the values of atmospheric dispersion in the communities closest to the project. The results obtained with Screen View and the parameters that make up the different evaluated scenarios were organized in the spreadsheet presented in the previous section, forming the research dataset.
The R program was obtained for free by accessing the developer’s page at
https://www.r-project.org/
, as shown in
Figure 3
. The RStudio graphical platform of the R program was also obtained for free by accessing the address
https://posit.co/download/rstudio-desktop/
. The processed data set was loaded into the R program, and the RStudio graphical interface (
Kronthaler & Zöllner, 2021
) was applied, producing statistically significant models and graphics crucial for the research, which are presented in the next section.
These models and graphics were made possible by employing the following “R” analysis packages (
Komperda, 2017
): “GGally”, “readxl”, “dplyr”, “Hmisc”, “corrplot”, “PerformanceAnalytics”, “ggplot2”, “qualityTools”, “lattice”, “MVar.pt”, “ggfortify”, “ExpDes”, “ape”, “psych”, “clValid”, “cluster”, “factoextra”, “plotly”, “MVA”, “ellipse”, “HSAUR2”, “tools”, “maps”, “spam”, “fields”, “CCA”, “mvnormtest”, “HH”, “vegan”, “ggpubr”, “rpart”, “rattle”, “rpart.plot”, “Cairo”, “RColorBrewer”, “car”, “FactoMineR”, “factoextra” (
Parsons et al., 2021
).
A script, a programming code containing instructions for loading the data set into the R program, instructions on constructing graphs, matrices, tables, indicators, and statistical models, was facilitated through programming language in “R”. Additionally, a qualitative evaluation among the variables “chimney height,” “dispersion,” “emission rate,” and the engine models were developed (
Chandra & Shang, 2019
), considering the parameters indicated in
Table 3
.
4. Results and Discussion4.1. Correlation between Variables
Initially, a correlation plot was produced between the variables, as shown in
Figure 1
, where blue coloration indicates positive correlation and red coloration indicates negative correlation (
Beunza et al., 2019
). The significance levels between the variables are the values shown in the quadrants of the plot. Higher significance levels exhibit more intense coloration, while lower intensity colorations indicate significance less than the intended 0.05 (
Kassambara, 2017
). It can be observed in
Figure 1
that the dispersion variable is more sensitive to the variables: “EmissionRate” and “ChimHeight” (Chimney Height).
Figure 1. Significance level between variables.
Figure 2
depicts the relationship “Dispersion × Emission Rate” and the engines (
Kumar, 2021
). According to the data in
Figure 2
, the following engines exhibited technically viable emission rates and dispersion values: HE SCR, MB SCR, TF IA, and WC SCR.
Figure 2. Qualitative Relationship between Dispersion × Emission Rate.
As observed in the graph of
Figure 2
, the blue bars represent the viable engine models considering the variables “Dispersion” and “Emission Rate”.
The models with SCR (
Kurzydym et al., 2022
) and with “IA” (Water Injection System) proved to be viable. It is noticeable that the “Diesel Cycle” engines classified as viable are categorized as “Tier 3”, and the gas turbine “Brayton Cycle” has a demineralized water injection system to reduce NOx emissions (
Khaliq et al., 2019
).
4.2. Linear Regression Models
To quantitatively determine the correlation between the variables that contributed most to minimizing “Dispersion,” a selection of correlation tests (
Musthafa et al., 2022
) was configured, considering all available variables. The summary of the linear regression model is provided in
Table 4
, where it can be noted that the “p-value” was 2.2e−16, with a cutoff line of 0.05 (
Macek et al., 2020
). Thus, the null hypothesis for the model was rejected, concluding that there is a statistically significant difference between the studied variables.
The “Pr(>|t|)” column represents the p-values, and significance levels are denoted as follows:
Summary of the model:
These statistics provide information about the distribution of the residuals, which were the differences between the observed values and the values predicted by the regression model.
<xref ref-type="bibr" rid="scirp.140744-"></xref>Table 4. Summary of the regression model for the atmospheric dispersion parameter.
Predictor Variable
Estimate
Std. Error
t value
Pr(>|t|)
(Intercept)
1.070.000
1.527.000
0.700
0.485
Tier
59.390
73.620
0.807
0.421
Power
4.020
8.125
0.495
0.622
Rotation
−6.341
444
−0.014
0.989
Consumption
−7.74
1.992
−0.004
0.997
WindSpeed
47.660
36.070
1.322
0.189
CommunityDist
302.1
57
5.301
4.66e−07***
ExhaustTemp
−648.4
3.776
−0.172
0.864
SpeedExhau
−28.770
5.146
−5.591
1.23e−07***
EmissionRate
149.200
34.240
4.356
2.62e−05***
ChimHeight
−35.290
2.431
−14.513
2.0e−16***
The summary statistics:
Adjusting the model and conducting ANOVA tests (
Gott et al., 2019
) and Shapiro-Wilk normality tests (
Ruxton et al., 2015
), it was possible to verify that the model is satisfactory, following a normal distribution, with a “p-value” of 3.408e−05, as shown in the summary of the model in
Table 5
. In this case, there is evidence to reject the null hypothesis of normality, suggesting that the data significantly deviate from a normal distribution.
Summary.aov (anova1):
<xref ref-type="bibr" rid="scirp.140744-"></xref>Table 5. Summary of the ANOVA test with the model.
Predictor Variable
Estimate
Std. Error
t value
Pr(>|t|)
Tier
1
6725.897
6725.897
68.224
Power
1
9519.925
9519.925
96.565
Rotation
1
567.424
567.424
5.756
Consumption
1
44.992
44.992
0.456
WindSpeed
1
2358.908
2358.908
23.927
CommunityDist
1
2065.920
2065.920
20.956
ExhaustTemp
1
61
61
0.001
EmissionRate
1
1587.284
1587.284
16.101
ChimHeight
1
14495.321
14495.321
147.033
Residuals
134
13210.494
98.586
The construction of the linear model using the ‘lm’ function generated the model presented in
Figure 3
(
Real et al., 2022
), where the predicted dispersion values by the model and the actual values are shown.
Figure 4
displays the distribution of residuals, with no noticeable clustering in their distribution. Both graphs demonstrate normality in the linear regression model (
Cheshmehzangi et al., 2022
).
Figure 3. Regression models for dispersion.
Figure 4. Residual distribution for the dispersion regression model.
The research used MANOVA test to examine the relationship between multiple response variables and predictors. Pillai’s trace statistic helped identify important differences in mean vectors, revealing that the variables “EmissionRate”, “ChimHeight” and “Dispersion” are strongly correlated. An individual analysis was also performed to check the normality of the data.
The hypothesis tests conducted by Pillai, Wilks and Hotelling-Lawley (
Mello et al., 2018
), as presented in
Tables 6-8
respectively revealed the presence of systematic differences and highly significant responses for the variables “Emission Rate”, “Chimney Height” and “Dispersion” as follows.
The summary of Pillai’s test is presented in
Table 6
.
<xref ref-type="bibr" rid="scirp.140744-"></xref>Table 6. Summary of Pillai’s correlation.
Predictor Variable
Df
Pillai
Approx F
Num Df
Den Df
Pr(>F)
EmissionRate
1
0.94282
274.118
8
133
<2.2e−16***
ChimHeight
1
0.19420
4.007
8
133
0.0002753***
Dispersion
1
0.32809
8.118
8
133
6.808e−09***
Residuals
140
The summary of the Wilks test is presented in
Table 7
.
<xref ref-type="bibr" rid="scirp.140744-"></xref>Table 7. Summary of the Wilks’ correlation.
Predictor Variable
Df
Pillai
Approx F
Num Df
Den Df
Pr(>F)
EmissionRate
1
0.05718
274.118
8
133
<2.2e−16***
ChimHeight
1
0.80580
4.007
8
133
0.0002753***
Dispersion
1
0.67191
8.118
8
133
6.808e−09***
Residuals
140
The summary of Hotelling-Lawley test is presented in
Table 8
.
<xref ref-type="bibr" rid="scirp.140744-"></xref>Table 8. Summary of the Hotelling-Lawley correlation.
Predictor Variable
Df
Pillai
Approx F
Num Df
Den Df
Pr(>F)
EmissionRate
1
164.883
274.118
8
133
<2.2e−16***
ChimHeight
1
0.2410
4.007
8
133
0.0002753***
Dispersion
1
0.4883
8.118
8
133
6.808e−09***
Residuals
140
The principal component analysis reduced the number of factors that satisfactorily predicted the studied model. In the results presented below, it is noted that the first five factors were responsible for approximately 92% of the variances, as shown in the standardized data presented in
Figure 5
.
To assess the quality of the variables composing the principal components, the “COS2” resource was utilized. A high “COS2” value indicates a good representation of the variable in the principal components (PCs). The variables demonstrating the best quality in the composition of “Dim.1” were “Power”, “Rotation”, “Consumption” and “ExhaustTem” (Exhaust Gas Temperature), as shown in
Figure 6
(
Burren & Pietsch, 2021
).
According to
Figure 6
, the contribution of the variables “Power”, “Rotation”, “Consumption” and “ExhaustTem” to the best quality in the components of “Dim.1” have been justified by their relationship with the amount of fuel burned in the process of generating electrical energy through thermoelectric plants, meaning that the greater the power produced, requires greater quantit of burning fuel (Consumption) to keep the rotation of engines resulting in higher exhaust gas temperature (ExhaustTem) and, consequently, a higher Dispersion value.
Figure 5. Contribution of factors.
Figure 6. Quality of variables COS2.
In “Dim.2,” it was observed that the primary contribution to the composition of the PCs is from the variable “ChimHeight” (Chimney Height) which was strongly correlated with “Dispersion.” In “Dim.3,” “WindSpeed” and “Tier” represents the mains contributions to the composition of the PCs. Finally, in “Dim.4,” the variable “CommunityDist”, distance until the community, was the main contribution to the PCs.
The correlations between the variables and the engines models are displayed in
Figure 7
, indicating interaction with “Dispersion” (
Burren & Pietsch, 2021
).
The joint analysis of the correlation between the contribution of PCs, Engines and Dispersion, as presented in
Figure 7
, showed that MB SCR, HE SCR, WC SCR, and TF IA engines exhibited the best results, being feasible from the perspective of emissions and atmospheric dispersion.
Figure 7. Correlation between the contribution of PCs, Engines, and Dispersion.
5. Conclusion
This paper presents a model for selecting combustion engines in thermoelectric
projects prior to the atmospheric dispersion study (ADS). This selection optimizes time and costs, avoiding rework and delays, since these engines are the main sources of pollutant emissions. The research fills a gap in the literature on selection models specific to this context. A project with twelve types of engines was used as the basis for the analysis, employing free software tools, such as Screen View and RStudio, for dispersion modeling and data processing.
The analysis of emission data, through the resources provided by the “R” program associated with Screen View, demonstrated that the best solutions, from the perspective of atmospheric emissions and disperson, considering the analyzed combustion engine models, were WC with SCR system, MB with SCR, HE with SCR and TF IA with injection water system. The WC SCR, MB SCR and HE SCR models, all with Tier 3 classification.
The Gas turbine (TF model) without injection water system was the only “Slightly Feasible” engine solution with atmospheric dispersion value near of limit definided in Resolution CONAMA 491, considering the chimney height of 25 meters. The other engine models only presented viable results with a chimney height of at least 45 meters, being rejected because in the case of alternative combustion engines, the individual power is small, requiring dozens or hundreds of engines to reach the maximum power of the thermoelectric plant and chimneys need to be individualized.
The conventional alternative engine models, without Tier classification or lower than Tier 2, had “Bad” results. The Gas turbine model (Brayton cycle) “TF,” without the water injection system for NOx reduction, also had “Bad” results. However, the turbine model considered in this study is from the 1980s, lacking updated technology for reducing atmospheric emissions, which justifies the “Bad” results.
The results showed that “Dispersion” is strongly impacted by the variable “Emission Rate” and “Chimney Height”, so any solution to reduce the level of Dispersion should consider these two variables. The best solutions must consider the technology embedded in the engines, mainly those related to the combustion process, to minimize the impacts of emissions at the source.
The average annual wind speed in the studied region was 2.1 m/s, a value that contributed to the “Good” results for the Dispersion variable. Wind speed values above 3.0 m/s require a “medium” or “high” chimney height (
Table 2
), even for engines with a “feasible” emission rate. The correlation between these variables and the other was presented in
Figure 1
,
Figure 2
and
Figure 6
.
The main components analysis demonstrated that the variables with the best quality in the composition of “Dim.1” were “Power”, “Rotation”, “Consumption” and “ExhaustTemp” (Exhaust Gas Temperature), as shown in
Figure 7
. It was justified by their relationship with the amount of fuel burned when the power was greater, requiring greater quantit of fuel (Consumption) to keep the rotation of engines resulting in higher exhaust gas temperature (ExhaustTem) and, consequently, a higher Dispersion value.
The data analyzed also showed that for a “Wind Speed” of 2.1 m/s, average annual in the studied region, “Distance to Community” of 1800 meters, “Exhaust Gas Velocity” of 35 m/s, characteristic of the analyzed engines, alternative combustion engines with NOx emission rate of up to 3.0 g/kWh presented the best results for atmospheric “Dispersion.” This emission rate value can be achieved by reciprocating internal combustion engines with “Tier 3” certification.
Tier 3 engines have SCR system, a feature that impacts the initial project cost by approximately 8% to 10%. This system also impacts operational costs, due to the need for periodic replacement of catalysts and the use of AdBlue (
Silva et al., 2017
). Thus, it is important to assess the financial impacts through Technical and Economic Feasibility Studies (TEFS) of the project.
This study was limited to NOx emissions as they are the most challenging atmospheric pollutants in thermoelectric projects when it comes to Diesel cycle and Brayton cycle combustion engines. Technological solutions to meet NOx emission requirements help to meet other atmospheric pollutants.
Therefore, the model proposed in this research contributed to the pre-selection of combustion engines in a thermoelectric project, according to the analysis of emissions and atmospheric dispersion. This model did not intend to replace the definitive atmospheric dispersion study that must be provided to the environmental agency to obtain the Preliminary License and Installation License for the new project but contributed to cost reduction and time for the elaboration of ADS. Finally, although this study was limited to thermoelectric projects, it can be replicated and extrapolated to other industrial sectors.
Data Availability
Data will be made available on request.
Acknowledgements
Eduardo dos Santos Piropo acknowledges the financial support provided by the Coordination for the Improvement of Higher Education Personnel CAPES.
References
Ahmadi, M. H., Ahmadi, M., Açıkkalp, E., Alhuyi Nazari, M., Arab Pour Yazdi, M.,&Kumar, R. (2018). New Thermodynamic Analysis and Optimization of Performance of an Irreversible Diesel Cycle. Environmental Progress&Sustainable Energy, 37, 1475-1490. >https://doi.org/10.1002/ep.12810
Alnahdi, A., Elkamel, A., Shaik, M. A., Al-Sobhi, S. A.,&Erenay, F. S. (2019). Optimal Production Planning and Pollution Control in Petroleum Refineries Using Mathematical Programming and Dispersion Models. Sustainability, 11, Article No. 3771. >https://doi.org/10.3390/su11143771
Beunza, J., Puertas, E., García-Ovejero, E., Villalba, G., Condes, E., Koleva, G. et al. (2019). Comparison of Machine Learning Algorithms for Clinical Event Prediction (Risk of Coronary Heart Disease). Journal of Biomedical Informatics, 97, Article ID: 103257. >https://doi.org/10.1016/j.jbi.2019.103257
Brusca, S., Famoso, F., Lanzafame, R., Mauro, S., Garrano, A. M. C.,&Monforte, P. (2016). Theoretical and Experimental Study of Gaussian Plume Model in Small Scale System. Energy Procedia, 101, 58-65. >https://doi.org/10.1016/j.egypro.2016.11.008
Burren, A.,&Pietsch, C. (2021). Distress Regulates Different Pathways in the Brain of Common Carp: A Preliminary Study. Animals, 11, Article No. 585. >https://doi.org/10.3390/ani11020585
Capelli, L., Sironi, S., Del Rosso, R.,&Guillot, J. (2013). Measuring Odours in the Environment vs. Dispersion Modelling: A Review. Atmospheric Environment, 79, 731-743. >https://doi.org/10.1016/j.atmosenv.2013.07.029
Chandra, Y.,&Shang, L. (2019). Qualitative Research Using R: A Systematic Approach. Springer.
Cheshmehzangi, A., Li, Y., Li, H., Zhang, S., Huang, X., Chen, X. et al. (2022). A Hierarchical Study for Urban Statistical Indicators on the Prevalence of COVID-19 in Chinese City Clusters Based on Multiple Linear Regression (MLR) and Polynomial Best Subset Regression (PBSR) Analysis. Scientific Reports, 12, Article No. 1964. >https://doi.org/10.1038/s41598-022-05859-8
Clementoni, E. M., Cox, T. L.,&Sprague, C. P. (2014). Startup and Operation of a Supercritical Carbon Dioxide Brayton Cycle. Journal of Engineering for Gas Turbines and Power, 136, Article ID: 071701. >https://doi.org/10.1115/1.4026539
Dayal, V. (2015). R and RStudio. In An Introduction to R for Quantitative Economics (pp. 7-14). Springer.
Dedes, E. K., Hudson, D. A.,&Turnock, S. R. (2012). Assessing the Potential of Hybrid Energy Technology to Reduce Exhaust Emissions from Global Shipping. Energy Policy, 40, 204-218. >https://doi.org/10.1016/j.enpol.2011.09.046
Dogan, N. O.,&Tugcu, C. T. (2015). Energy Efficiency in Electricity Production: A Data Envelopment Analysis (DEA) Approach for the G-20 Countries. International Journal of Energy Economics and Policy, 5, 246-252. >https://www.econjournals.com/index.php/ijeep/article/view/1043
Elzalik, M., Rezk, H., Mostafa, R., Thomas, J.,&Shehata, E. G. (2020). An Experimental Investigation on Electrical Performance and Characterization of Thermoelectric Generator. International Journal of Energy Research, 44, 128-143. >https://doi.org/10.1002/er.4873
Espinosa, J. F. M.,&Luna, M. A. G. (2021). Uso del modelo de dispersión de contaminantes Screen View (pp. 1-34). Notas de Campus.
Favre-Perrod, P. (2005). A Vision of Future Energy Networks. In Proceedings of the Inaugural IEEE PES 2005 Conference and Exposition in Africa (pp. 13-17). IEEE. >https://doi.org/10.1109/pesafr.2005.1611778
Gott, R. C., Coffey, P. L.,&Bechtel, M. (2019). Evaluation of Behavior-Based Deterrence of Imported Cabbageworm in Collards, 2017. Arthropod Management Tests, 44, tsz004. >https://doi.org/10.1093/amt/tsz004
Grömping, U. (2015). Using R and RStudio for Data Management, Statistical Analysis and Graphics. Journal of Statistical Software, 68, 1-7.
Holmes, N. S.,&Morawska, L. (2006). A Review of Dispersion Modelling and Its Application to the Dispersion of Particles: An Overview of Different Dispersion Models Available. Atmospheric Environment, 40, 5902-5928. >https://doi.org/10.1016/j.atmosenv.2006.06.003
Kassambara, A. (2017). R Graphics Essentials for Great Data Visualization: 200 Practical Examples You Want to Know for Data Science. STHDA. >https://www.sthda.com/english/articles/32-r-graphics-essentials/134-r-basics-for-data-visualization/
Khaliq, A., Habib, M. A.,&Choudhary, K. (2019). A Thermo-Environmental Evaluation of a Modified Combustion Gas Turbine Plant. Journal of Energy Resources Technology, 141, Article ID: 042004. >https://doi.org/10.1115/1.4041898
Komperda, R. (2017). Likert-Type Survey Data Analysis with R and Rstudio. In Computer-Aided Data Analysis in Chemical Education Research (CADACER): Advances and Avenues (pp. 91-116). American Chemical Society. >https://doi.org/10.1021/bk-2017-1260.ch007
Kronthaler, F.,&Zöllner, S. (2021). Testing Normal Distribution with RStudio. In F. Kronthaler,&S. Zöllner (Eds.), Data Analysis with RStudio (pp. 59-63). Springer. >https://doi.org/10.1007/978-3-662-62518-7_5
Kumar, R. V. (2021). Exploratory Data Analysis Using R&RStudio.
Kurzydym, D., Żmudka, Z., Perrone, D.,&Klimanek, A. (2022). Experimental and Numerical Investigation of Nitrogen Oxides Reduction in Diesel Engine Selective Catalytic Reduction System. Fuel, 313, Article ID: 122971. >https://doi.org/10.1016/j.fuel.2021.122971
Macek, P., Biskup, M., Terek-Derszniak, M., Krol, H., Smok-Kalwat, J., Gozdz, S. et al. (2020). Optimal Cut-Off Values for Anthropometric Measures of Obesity in Screening for Cardiometabolic Disorders in Adults. Scientific Reports, 10, Article ID: 11253. >https://doi.org/10.1038/s41598-020-68265-y
Mehdizadeh, F.,&Rifai, H. S. (2004). Modeling Point Source Plumes at High Altitudes Using a Modified Gaussian Model. Atmospheric Environment, 38, 821-831. >https://doi.org/10.1016/j.atmosenv.2003.10.041
Mello, K. d., Valente, R. A., Randhir, T. O.,&Vettorazzi, C. A. (2018). Impacts of Tropical Forest Cover on Water Quality in Agricultural Watersheds in Southeastern Brazil. Ecological Indicators, 93, 1293-1301. >https://doi.org/10.1016/j.ecolind.2018.06.030
Ministry of Environment of Brazil (2007). CONAMA Resolution 382: Establishes Maximum Emission Limits for Atmospheric Pollutants from Stationary Sources (pp. 131-137). Official Gazette of the Union, No. 1. >https://conama.mma.gov.br
Ministry of Environment of Brazil (2018). CONAMA Resolution 491: Establishes Air Quality Standards (p. 155). Official Gazette of the Union, No. 223, Section 1. >https://conama.mma.gov.br
Musthafa, M. M., Abdullah, F.,&Koivula, M. J. (2022). The Functioning of Different Beetle (Coleoptera) Sampling Methods across Altitudinal Gradients in Peninsular Malaysia. PLOS ONE, 17, e0266076. >https://doi.org/10.1371/journal.pone.0266076
Nemitallah, M. A., Rashwan, S. S., Mansir, I. B., Abdelhafez, A. A.,&Habib, M. A. (2018). Review of Novel Combustion Techniques for Clean Power Production in Gas Turbines. Energy&Fuels, 32, 979-1004. >https://doi.org/10.1021/acs.energyfuels.7b03607
Ni, P., Wang, X.,&Li, H. (2020). A Review on Regulations, Current Status, Effects and Reduction Strategies of Emissions for Marine Diesel Engines. Fuel, 279, Article ID: 118477. >https://doi.org/10.1016/j.fuel.2020.118477
Paraschiv, S., Coman, G.,&Paraschiv, L. S. (2019). Simulation of Plume Dispersion Emitted from Industrial Sources Based on Gaussian Model. AIP Conference Proceedings, 2123, Article ID: 020059. >https://doi.org/10.1063/1.5116986
Parsons, L., Rodriguez, O.,&Holding, D. R. (2021). Improved Taste and Texture in Novel Popcorn Varieties Compared to Conventional Lines. Journal of Sensory Studies, 36, e12687. >https://doi.org/10.1111/joss.12687
Pongrac, P., Fischer, S., Thompson, J. A., Wright, G.,&White, P. J. (2020). Early Responses of Brassica Oleracea Roots to Zinc Supply under Sufficient and Sub-Optimal Phosphorus Supply. Frontiers in Plant Science, 10, Article No. 1645. >https://doi.org/10.3389/fpls.2019.01645
Racine, J. S. (2012). RStudio: A Platform-Independent IDE for R and Sweave.
Real, D., Bennett, R. G., Nazeri, N. K.,&Weaver, D. M. (2022). Critical P, K and S Concentrations in Soil and Shoot Samples for Optimal Tedera Productivity and Nodulation. Agronomy, 12, Article No. 1581. >https://doi.org/10.3390/agronomy12071581
Ruxton, G. D., Wilkinson, D. M.,&Neuhäuser, M. (2015). Advice on Testing the Null Hypothesis That a Sample Is Drawn from a Normal Distribution. Animal Behaviour, 107, 249-252. >https://doi.org/10.1016/j.anbehav.2015.07.006
Silva, G. N. D., Nery, L. A. S. S.,&Júnior, L. T. K. (2017). Economic-Financial Analysis of an Investment of Tanks for ARLA 32 at the Fleet of a Transportation Company. Independent Journal of Management&Production, 8, 583-595.
Turner, D. B. (2020). Workbook of Atmospheric Dispersion Estimates: An Introduction to Dispersion Modeling. CRC Press.
Zhong, S., Zhou, L.,&Wang, Z. (2011). Software for Environmental Impact Assessment of Air Pollution Dispersion Based on Arcgis. Procedia Environmental Sciences, 10, 2792-2797. >https://doi.org/10.1016/j.proenv.2011.09.433