Decision support system to evaluate a vandalized and deteriorated oil pipeline transportation system using artificial intelligence techniques. Part 2: analysis of the operational and economic risk

1 Introduction and related works

In 2021, the world economy rebounded considerably from the 2020 outbreak of the COVID-19 pandemic. However, the pandemic continued to be a significant challenge throughout the year, particularly with the emergence of new variants such as Delta in 2021 and Omicron in 2022 (OPEC 2021). The oil industry in Mexico has suffered economic failures due to price fluctuations in recent years and reached historically low sales of USD 36.24/barrel in 2020. However, the problem of the volume of transportation through pipelines increased considerably in 2020, generating downward variations of 8.5 % thousand barrels (TB) against the established daily operating program, reflecting losses for the company. Within the foreign trade of hydrocarbons, for example, the Maya oil mix generated downward variations due to transportation of 75.7 % thousand barrels per day (TBD), which represents significant losses in millions of dollars (MD) (Petroleos Méxicanos 2021). Low profits limit the maintenance of the facilities, some of which have a critical impact on mechanical integrity due to corrosion and vandalism, which have increased considerably from 2016 to 2021 in Mexico. The pipeline transportation systems (gas pipelines, polyducts, oil pipelines), the latter with an operational life of more than 60 years, report a Volume released of 32 (million cubic feet) of crude oil in natural environments only in 2021, being states such as Veracruz, Puebla, Tlaxcala, Tabasco, Campeche, Yucatan of the most affected, in addition to 5289 clandestine intakes to pipelines (fuel theft) detected in the first quarter of 2021 with a nationwide recurrence every 3 h and 3 min by the company surveillance systems. Table 1 represents the information above (IGAVIM Observatotio Ciudadano 2021; Infobae 2018).

The pipeline network has an operating length of over 17,000 km for transporting hydrocarbons, oil, and petrochemical products. The case study is being developed at the Mendoza Distribution Center (MDC) in the states of Veracruz and Puebla with a length of 572 km, pipes of 24/30″Ø, and a variable operating program of 90,000 to 140,000 barrels of oil transported per day (BTD) low demand and high system instability (Pemex Refinación 2011). The process is carried out with a sequential hydrocarbon pumping system that enables the flow of product, powered by six stations, supported by telemetry and supervision, data acquisition and control system (SCADA), 96 isolation valves on the line, 17 check valves, 28 devil traps, and remote-controlled instrumentation kit. The authors presented a systematic literature review (SLR) that deepens the current state-of-the-art with 245 works published between 1994 and 2018 that show an increasing interest in AI in related areas of the oil and gas industry (Cid-Galiot et al. 2021). This study evaluates the effects of corrosion on the mechanical integrity of the PTS using AI to reconfigure and ensure the operational reliability of its process (Cid-Galiot et al. 2022). Part 2 presents a financial risk analysis using Monte Carlo simulation (MCS) that will improve decision-making processes based on corrosive profiles of soils, supply, demand, and inflation.

MCS has proven results for analyzing the behavior of corrosion in pipelines (Caleyo et al. 2002, 2009; Li et al. 2009; Naess et al. 2009; Nesic et al. 2001). They were demonstrating its capacity by estimating corrosive growth scenarios considering current regulations and reliability systems (Hasan et al. 2012; Semiga and Dinovitzer 2012). Working collectively, individually, or in hybrid environments (El-Abbasy et al. 2015; Jahani et al. 2013; Leira et al. 2016; Modiri et al. 2014; Qin and Zhou 2014; Shabarchin and Tesfamariam 2017; Younsi et al. 2013). Modeling the cognitive knowledge of specialists to reduce uncertainty biases, failure probabilities, and material fatigue in pipelines (Abyani and Bahaari 2020; Ossai 2013; Ossai et al. 2015). Analyzing life cycle costs of PTS and its operational profitability concerning time (Jones and Ferrari 2017; Park et al. 2018; Parvizsedghy et al. 2015; Willigers and Bratvold 2009). However, in Mexico, the applications are limited to processes of maintenance, exploration, and oil extraction rather than to economic factors of transportation (Caleyo et al. 2007; Duda et al. 2005; Mudford 2000). The assessment of PTS is developed individually, considering field data, laboratory, and cognitive knowledge, to achieve the best results and to be replicated worldwide under similar conditions. Below, the authors mention the Purpose; Design/methodology/approach; Findings; Practical implications, and Originality/value which generated the research.

2 Methodology and results

The methodology underpins its success by considering the eight critical issues for the decision support systems discipline shown in Figure 1 (Arnott and Pervan 2008).

1. The relevance of DSS research by connecting the practical part (Kaewpradap et al. 2017; Lee et al. 2015) with the theory (Cid-Galiot et al. 2021; Nyborg 2010);

2. DSS research methods and paradigms, with case study research and design sciences (Din et al. 2015; Kaduková et al. 2014; Lu et al. 2018);

3. The theoretical foundations of DSS research, considering explicit bases in the decision-making judgment through theoretical and practical foundations applied in a case study in Mexico;

4. The role of the information technology (IT) artifact in DSS research is fundamental in all the processes to analyze, interpret, and connect the information;

5. Funding for DSS research, no implicit support from any organization. Compliance with keys six, seven and eight are generated by considering the most significant number of interpretive case studies and the academic rigor to support the research designs. Modules one and two contain results of the Analysis and prediction of pipeline corrosion behavior (Cid-Galiot et al. 2022). Module three relates the financial effects of corrosion on the utility of the company (Jones and Ferrari 2017; Park et al. 2018; Parvizsedghy et al. 2015; Willigers and Bratvold 2009). A three-stage structure is proposed for the DSS. The structure is based on the following three modules:

   • Module 1. Evaluation of the mechanical integrity of the PTS with an FES: FES evaluates, analyzes, and predicts the status of PTS using field data and cognitive knowledge;

   • Module 2. ANN predicts the inhibition capacity of Mexican hydrocarbons in pipeline corrosion: Mexican mixtures of crude oil Istmo, Olmeca, and Maya;

   • Module 3. Operational and economic analysis with MCS: The MCS determines the analysis of the operational and economic risks of PTS under the effects of corrosives to inflation, oil supply, and demand.

Figure 1:

Methodology (based on Arnott and Pervan 2008; Cid-Galiot et al. 2021, 2022).

The technicians use the Matlab toolbox for FES development, for the design of ANN, and Palisade’s @Risk for the design of MCS. Each procedure was developed following international regulations. Integrating the three modules to evaluate the PTS generates a comprehensive qualitative perspective of the current operational reliability of the system and a quantitative one when analyzing the desired module to determine decisions. As an intelligent evaluation tool for operational reliability in pipelines, the DSS approach generates notable advantages over traditional procedures and disadvantages by generating changes in the organization’s culture, as shown in Table 2 below.

3 Development of the DSS

3.1 Operational and economic analysis with MCS

MCS is a mathematical technique that predicts the possible results of an uncertain event, which uses past data to predict a series of future results based on a choice of action. Its principle lies in ergodicity, which describes the statistical behavior of a moving point in a closed system where the computer runs enough simulations to produce the eventual output of different inputs, consisting of input variables, output variables, and a mathematical model. The methodology proposed in this section comprises four phases: Planning, which evaluates and determines the variables with the most significant impact on the study, the scope, and the limitations of the research with the company. The preparation phase interprets the study variables’ behavior through mathematical models that reliably represent the system’s behavior. The evaluation phase develops a risk analysis and validation of alternative solutions to increase the company’s profitability. The improvement phase shows continuous monitoring and evaluation of various scenarios through probability and statistics to identify improvement areas. The methodology was created considering the basic principles of the discounted cash flow (DCF) technique with MCS, considering a wide range of risks and their correlation, and incorporating changes over time shown in Figure 2 (Flores Ríos and Moscoso Escobar 2009).

Figure 2:

MCS methodology (based on Flores Ríos and Moscoso Escobar 2009).

3.1.1 Planning

The company explained the current status of the transportation system and expressed its low indicators of operational reliability. The research team planned an improvement proposal, attaching information from its business management system that manages financial, operational, and maintenance processes (Pemex 2016). Databases of the last study of mechanical integrity of the facilities (Kulikov et al. 2011) studies of ph and soil resistivity as well as cognitive knowledge of specialist personnel (Mecanz 2016). MCS is used to generate risk scenarios. Table 3 shows the uncertainty distributions generated for nine parameters, with data from the Mexican oil industry and the corrosion experts. For the Analysis of the economic viability of the PTS, two uncertain parameters were considered: the effects of the sale price and price inflation (supply and demand). For the variable of the sale price, 1095 daily data of the price of the Mexican mixture of 2016–2019 were considered. For the inflation variable, the monthly data of the same period were considered. Oil and gas industry experts assigned triangular distributions of corrosion due to aquatic, terrestrial, and atmospheric effects through soil resistivity and pH. In corrosion due to atmospheric effects, the maximum value (0.65) is the worst case and indicates a high generation due to gases, dust, or excess moisture in the air. The most likely case is (0.43), and the minimum value (0.20) indicates the best case due to the low generation by atmospheric agents. @Risk is used to fit the best distribution and parameters of the above variables through goodness and fit tests to obtain reliable results between the natural system “y(t) = PTS” and the simulated one “ỹ(t) = MCS” to identify the variability between the fitted outputs “σ(t) = y(t) − ỹ(t).” Table 3 presents a representative sample of the population of actual and estimated variables, with 180 data previously defined by a factorial design of experiments to identify correlation or variance between the variables of the actual and simulated systems.

Table 3:

Input parameters for Monte Carlo simulation.

Parameter					Distribution			N		Min	Max		Mean		Std. Dev.		Unit
Corrosion due to aquatic effects (C-AQ)					Triangular			–		0.60	0.95		0.78		–		%
Corrosion by terrestrial effects (C-TE)					Triangular			–		0.45	0.80		0.62		–		%
Corrosion due to atmospheric effects (C-AT)					Triangular			–		0.20	0.65		0.43		–		%
Preventive maintenance (Pre-MA)					Lognormal			114		4	48		8.86		7.32		Hr
Maintenance under operational conditions (OC-MA)					Lognormal			56		4	24		13.28		5.75		Hr
Corrective maintenance (Co-MA)					Weibull			154		2	36		9.38		5.56		Hr
Proactive maintenance (Pro-MA)					Uniform			88		2	24		14.26		6.43		Hr
Inflation of prices (I–P)					LogLogistic			36		2.54	6.77		4.58		1.42		%
Effects of the sale price (E-SP)					BetaGeneral			1095		21.87	77.21		48.08		17.32		USD
Difference between real and simulated variables: σ = y(t) − ỹ(t)
C-AQ	0.6863	0.6536	0.7016	0.7707		0.6646	0.8246		0.8782			0.7412		0.9438		0.8846
ỹ(t) = C-AQ	0.6862	0.6736	0.7065	0.7719		0.6691	0.8419		0.8912			0.7410		0.9441		0.8800
C-TE	0.4754	0.6121	0.5343	0.743		0.6023	0.7024		0.6356			0.6989		0.7934		0.7464
ỹ(t) = C-TE	0.4786	0.6101	0.5486	0.7497		0.6094	0.7175		0.6378			0.6921		0.7965		0.7461
C-AT	0.5628	0.2197	0.6387	0.4269		0.6323	0.5466		0.5831			0.2737		0.3262		0.4542
ỹ(t) = C-AT	0.5608	0.2342	0.6307	0.4654		0.6321	0.5489		0.5832			0.2737		0.3389		0.4500
Pre-MA	4.31	25.44	39.40	9.52		16.18	10.50		42.29			36.36		19.51		5.23
ỹ(t) = Pre-MA	4.30	26.30	39.41	9.54		16.18	9.56		41.18			35.49		20.47		4.04
OC-MA	4.49	12.23	21.1	22.18		15.21	6.24		10.49			8.23		19.48		14.17
ỹ(t) = OC-MA	4.38	12.54	21.1	21.14		14.27	7.04		12.15			8.25		19.41		13.42
Co-MA	31.04	20.18	10.38	3.52		14.24	6.16		29.42			12.16		5.33		18.59
ỹ(t) = Co-MA	31.46	21.15	11.37	3.34		15.04	7.13		29.02			11.47		5.37		16.14
Pro-MA	2.17	18.03	6.01	11.29		20.56	9.35		5.04			15.19		7.58		23.99
ỹ(t) = Pro-MA	2.6	16.16	6.37	10.43		22.04	10.19		6.43			16.18		8.23		23.41
I–P	3.09	5.02	4.27	5.26		6.05	5.59		2.62			3.12		5.11		3.43
ỹ(t) = I–P	3.41	6.13	4.58	4.34		5.36	5.17		2.47			3.46		5.14		2.53
E-SP	72.34	74.25	58.48	62.18		66.27	33.27		26.42			75.41		22.52		18.02
ỹ(t) = E-SP	70.12	75.53	59.11	63.32		65.64	34.41		26.24			76.17		23.19		19.53
[A]

1. N, number of values; Min, minimum; Max, maximum; Std. Dev., standard deviation.

Illustration [A] represents an enlarged view of the behavior of the fundamental variables “y(t) = PTS” and the simulated variables “ỹ(t) = MCS,” which facilitates the visual identification of their variance “σ(t) = y(t) − ỹ(t).” For example, the variable “y(t) = E-SP” represents data of the real variable of the company’s system, and “ỹ(t) = E-SP” represents data of the system simulated by MCS, the variance calculated between the two variables is represented by “E-SP σ(t) = y(t) − ỹ(t)” with a result of 2.01 %, this being the widest gap identified between the study variables, which guarantees low biases between the data used and reliability in the simulated system.

3.1.2 Preparation

The morphology of corrosion in pipelines under atmospheric, aquatic, and terrestrial effects can be divided into three classifications (Zhang et al. 2019): (A) Pitting Corrosion: when pitting corrosion in the local region of the pipeline surface, but the rest is not corroded. The limit state function for Pitting Corrosion (PC) mode g_PC(x) is defined by the following equation (Abyani and Bahaari 2020):

(1)gLB (X)=0.8t –d

where t and d denote pipeline wall thickness and defect depth, respectively. (B) Local Burst: It happens in certain parts of the pipeline metal surface and is more destructive than the previous type even though its weight loss is minor than uniform corrosion. The limit state function for Local Burst (LB) defect g_LB(x) is determined by Eq. (2):

(2)gLB (X)={2mfSy(tD)(1−dt1−dtF)−Pα}

In this equation, m_f, S_y, t, d, D, p, and F represent multiplying factor, yield stress, pipeline wall thickness, defect depth, pipeline diameter, the internal pressure acting at the relevant cross-section, Folia’s factor, and L is the corrosion defect length. Employees in the evaluation of the mechanical integrity of the pipeline in part 1 of the investigation (Cid-Galiot et al. 2022).

(c) Uniform Corrosion (rupture): occurs on the pipeline surface, and the metal loss is considerable, so the rest of the pipeline wall thickness cannot afford to transport media. The limit state function for Uniform Corrosion defect is g_UC(x) obtained by Eq. (3).

(3)gUC (X)=Prp – Pα

Each of the three considered limit state functions P(g(x) < 0) is equal to the probability of failure due to corrosion effects, and P(g(x) ≥ 0) = 1 − P(g(x) < 0) equals the reliability of the corroded pipeline. The statistical properties of the influential random variables and their corresponding limit state functions model to simulate their behavior through MCS. The value of reliability-focused maintenance today has optimized pipeline maintenance availability and control processes; models were first developed in Enbridge Liquids Pipelines in 2006 and, in the last three years, have contributed over $200 Million in capital cost avoidance while maintaining or improving the reliable design and operation of the pipeline system (Jones and Ferrari 2017). The interaction of large, complex, and multi-layered systems can then be analyzed using the MCS (or stochastic discrete event simulation), quantifying the entire system’s output with greater accuracy than other estimating tools or methods.

The four maintenance systems’ uncertainty models use probability distribution curves instead of point estimators, such as the mean time between maintenance. The model uses a random number generator to sample these distribution curves that determine if a pipeline will fail, how long it will take to repair, Etc. The model executes hundreds of life cycles, each generating different results. This approach results in convergence to a statistically probable system response over its lifetime. Lausser’s law was used for a serial maintenance reliability system (Department of Defense of the USA 1998) by Eq. (4).

(4)Rsystem=R1x R2 x…x Rn

Finally, the cash flows of the maintenance scenarios and the equivalent uniform annual cost (EUAC) function of each maintenance scenario are calculated using MCS. It was computed using the cost data previously; the equivalent economic value of each scenario was normally computed using the net present value (NPV). However, the EUAC was used for scenarios with different service lives (Parvizsedghy et al. 2015). The EUAC was computed by converting NPV to a constant annual value over the service life of the MCS and was used to address the uncertainties in the Inflation of mooned and the Effects of the sale price. Equations (5), (6) and (7) were used to calculate the probabilistic NPV and EUAC values:

(5)NPV˜=∑t=1nCt˜(PF,I,˜ t)=∑t=1nCt ˜x 1(1+I˜)t

(6)EUAC˜=NPV˜=(AP, I,˜ n)

(7)(AP, I,˜ n)= I˜(1− I)˜n(1− I)˜n −1

where NPV˜ is the probability distribution function of NPV of the cash flow under evaluation, Ct ˜ is the probability distribution function of the maintenance components of the total cost in year t, n is the service life of the pipeline in years, the probability distribution function of the sale price of the forecast barrel of hydrocarbon during the useful life of the pipeline, and (AP, I,˜ n), is the probability distribution function of conversion factor from Present Worth to Equivalent Uniform Annual Worth. The loss of daily barrels of the Istmo and Maya oil mixtures. The amount of loss of crude oil used is directly related to the gross domestic product (GDP) income by sale, extraction, and distribution.

3.1.3 Evaluation

In this phase, risk analysis and validation of alternative solutions to increase the company’s profitability. With the @Risk software, scenarios are analyzed to calculate the fluctuations due to the distribution and sale of hydrocarbons for the periods 2016 to 2019, considering deterministic parameters such as inactive days of the month (out of operation), losses (incidents, spills, or non-transported product), transportation costs and income, inflation and customs taxes on international sales of hydrocarbons. Table 4 shows a quarter of the year 2018 with station 6, representing the behavior of the Maya and Istmo daily hydrocarbon transportation program established for 460,000 barrels (DTP) with effectiveness percentages below 80.40 % by integrating a maximum of 432,158 thousand of barrels per day (FI L-24″ 30″ Ф/TBD) and losses of up to 24.88 % with 244,892 (JL-DPT/TBD), image [A] represents a graphical interpretation of the data.

Table 4:

Fraction of data from the integrated computation and losses in the first quarter of 2018, based on the estimated daily transport program (Pemex 2017).

	January				February				March
Day	IM L-24″Ф	II L-30″Ф	FI L-24″ 30″Ф	JL-DTP	IM L-24″Ф	II L-30″Ф	FI L-24″ 30″Ф	FL-DTP	IM L-24″Ф	II L-30″Ф	FI L-24″ 30″Ф	ML-DTP
1	101,651	254,994	356,646	103,354	163,158	134,642	297,800	162,200	158,515	190,015	348,530	111,470
2	129,765	260,808	390,575	69,425	141,212	288,707	429,919	30,081	103,379	185,041	288,420	171,580
3	152,904	161,865	314,772	145,228	89,175	275,746	364,921	95,079	150,216	274,883	425,099	34,901
4	109,772	260,046	369,822	90,178	172,403	218,314	390,717	69,283	127,263	206,652	333,915	126,085
5	124,068	276,549	400,622	59,378	95,892	195,267	291,159	168,841	203,603	238,105	441,708	18,292
6	201,142	200,465	401,613	58,387	132,374	179,933	312,307	147,693	193,395	244,019	437,414	22,586
7	110,715	245,284	356,006	103,994	83,698	157,566	241,264	218,736	107,373	203,265	310,638	149,362
8	153,371	253,545	406,924	53,076	75,210	139,898	215,108	244,892	240,312	140,698	381,010	78,990
9	112,655	160,789	273,453	186,547	112,880	168,653	281,533	178,467	99,640	158,669	258,309	201,691
10	206,294	210,666	416,970	43,030	180,157	162,774	342,931	117,069	189,212	117,528	306,740	153,260
11	120,259	219,641	339,911	120,089	166,023	229,244	395,267	64,733	144,728	217,696	362,424	97,576
12	182,516	171,602	354,130	105,870	125,030	142,006	267,036	192,964	161,314	167,011	328,325	131,675
13	141,113	291,032	432,158	27,842	162,014	202,505	364,519	95,481	194,876	214,876	409,752	50,248
14	163,209	209,057	372,280	87,720	107,644	260,123	367,767	92,233	157,742	189,077	346,819	113,181
15	124,408	202,824	327,247	132,753	149,922	140,606	290,528	169,472	245,368	150,919	396,287	63,713
16	207,785	180,488	388,289	71,711	126,229	218,343	344,572	115,428	111,279	271,320	382,599	77,401
17	96,220	180,977	277,214	182,786	227,861	140,155	368,016	91,984	201,680	231,105	432,785	27,215
18	175,512	239,455	414,985	45,015	187,514	183,486	371,000	89,000	198,177	180,722	378,899	81,101
19	133,622	297,386	431,027	28,973	209,316	176,035	385,351	74,649	122,323	103,543	225,866	234,134
20	161,284	188,845	350,149	109,851	146,229	156,119	302,348	157,652	200,492	128,704	329,196	130,804
21	103,420	175,843	279,284	180,716	148,694	153,462	302,156	157,844	170,763	126,051	296,814	163,186
22	150,580	268,524	419,126	40,874	117,148	171,239	288,387	171,613	147,422	110,952	258,374	201,626
23	181,576	241,700	423,299	36,701	146,900	211,630	358,530	101,470	130,592	174,340	304,932	155,068
24	139,207	286,596	425,827	34,173	148,250	216,911	365,161	94,839	125,092	148,999	274,091	185,909
25	179,498	220,208	399,731	60,269	195,490	240,918	436,408	23,592	157,305	239,259	396,564	63,436
26	134,753	210,538	345,317	114,683	180,102	195,773	375,875	84,125	171,125	190,498	361,623	98,377
27	182,005	178,144	360,176	99,824	98,137	217,722	315,859	144,141	99,132	115,795	214,927	245,073
28	109,060	276,141	385,229	74,771	114,105	151,883	265,988	194,012	160,701	271,970	432,671	27,329
29	102,046	268,572	370,647	89,353					174,874	232,431	407,305	52,695
30	192,750	209,233	402,013	57,987					246,083	203,900	449,983	10,017
31	156,482	185,777	342,290	117,710					97,486	255,314	352,800	107,200
Integrated total barrels and losses			11,527,732	2,732,268			9,332,427	3,547,573			10,874,819	3,385,181
Percentages vs. DTP			(80.40 %)	(19.60 %)			(75.12 %)	(24.88 %)			(76.26 %)	(23.74 %)
[A]

1. TBD, thousands of barrels per day; DTP, daily transportation program; IM L-24″ Ф, total integrated Maya hydrocarbon length – 24 inches of diameter pipeline; II L-30″ Ф, total integrated Istmo hydrocarbon length – 30 inches of diameter pipeline; FI L-24″ 30″ Ф, total hydrocarbon integrated length – 24 and 30 inches of diameter pipeline; JL-DTP, loss of product of January from DTP; FL-DTP, loss of product of February from DTP; ML-DTP, loss of product of March from DTP.

The amount expected to be transported each month is calculated, adding the integrated ones, obtaining a value of 14 260,000 barrels; however, the company considers the losses products not transported or spilled to DTP as losses. With the database, the calculation of the respective probability of each of the nine uncertain parameters of the model is made to obtain a representative value of the existence of each possible risk factor. Calculating the partial income of each month with the values of extraction and distribution per barrel of hydrocarbon of 12 and 20 USD, respectively, in 2018, as well as the net profit of the periods to be evaluated, adding change factors such as the sale rate and inflation.

3.1.4 Improvement and results

Shown according to the duration of the investigation and only three simulation scenarios: In January, February, and March, the Spearman correlation graphs are analyzed with 10 runs of the MCS model, each for the three months investigated. Figure 3A shows a positive behavior of the nine uncertain parameters and had high values of 0.68 % and low values of 0.15 % for January; This month has a probability of 90 %, with a maximum profit of USD $ 58,557,474 million and an average of $ 53,079,672 million (Figure 3B).

Figure 3:

Density and correlation graphs of January.

Figure 4A shows a predominantly positive behavior for eight uncertain parameters, however the effect of the sale price showed negative behaviors −0.48 %, which impacted the maximum profit with USD $238,061,134 million and an average of $207,624,443 million (Figure 4B).

Figure 4:

Density and correlation graphs of February.

Figure 5A shows that in March, each of the nine uncertain parameters had a positive and a negative influence, obtaining this month the highest values in the simulations with 0.82 %, however, it also shows the impact of price inflation in March, with a value of −0.39 %. As inflation in Mexico increased from 4.864 % in February to 5.353 % in March, the actual values included in the Analysis were volatile. With 90 % probability, maximum fluctuations of $253,848,660 million and minimums of $198,950,519 million were obtained (Figure 5B).

Figure 5:

Density and correlation graphs of March.

Various MCS runs were developed to validate previously obtained results, but this time through statistical Analysis (Table 5). Representing the January scenario, a maximum and minimum profit of $252 659 900 208,158,000 million dollars, respectively, considered a scenario with little risk of reaching a high income between the 20th and 25th percentile, equivalent to 7,655,100 million dollars (obtained by subtracting the values obtained by the percentiles) concerning other scenarios presented in the study, which minimizes subjectivity in decision making.

Table 5:

Simulation results.

Name	January profit	February profit	March profit
Description cell	Output utility	Output utility	Output utility
Minimum	$ 208,158,000.00	$ 187,196,200.00	$ 196,243,500.00
Maximum	$ 252,659,900.00	$ 234,738,900.00	$ 251,509,000.00
Mean	$ 231,515,200.00	$ 208,565,200.00	$ 222,928,800.00
Std. deviation	$ 14,260,660.00	$ 14,730,530.00	$ 19,776,420.00
Variance	2.03366E+14	2.16988E+14	3.91107E+14
Skewness	−0.2147535	0.4429184	0.03336368
Kurtosis	2.151993	2.651084	1.334294
Errors	0	0	0
Mode	$ 213,720,700.00	$ 193,139,000.00	$ 205,914,900.00
5 % Perc	$ 208,158,000.00	$ 187,196,200.00	$ 196,243,500.00
10 % Perc	$ 208,158,000.00	$ 187,196,200.00	$ 196,243,500.00
15 % Perc	$ 214,144,100.00	$ 194,325,400.00	$ 202,064,900.00
20 % Perc	$ 214,144,100.00	$ 194,325,400.00	$ 202,064,900.00
25 % Perc	$ 221,799,200.00	$ 197,081,200.00	$ 204,758,300.00
30 % Perc	$ 221,799,200.00	$ 197,081,200.00	$ 204,758,300.00
35 % Perc	$ 227,244,800.00	$ 201,310,600.00	$ 206,418,500.00
40 % Perc	$ 227,244,800.00	$ 201,310,600.00	$ 206,418,500.00
45 % Perc	$ 228,702,100.00	$ 205,417,700.00	$ 219,777,300.00
50 % Perc	$ 228,702,100.00	$ 205,417,700.00	$ 219,777,300.00
55 % Perc	$ 233,319,400.00	$ 210,634,800.00	$ 229,146,000.00
60 % Perc	$ 233,319,400.00	$ 210,634,800.00	$ 229,146,000.00
65 % Perc	$ 240,371,700.00	$ 211,885,200.00	$ 234,178,800.00
70 % Perc	$ 240,371,700.00	$ 211,885,200.00	$ 234,178,800.00
75 % Perc	$ 243,190,400.00	$ 215,941,400.00	$ 241,138,200.00
80 % Perc	$ 243,190,400.00	$ 215,941,400.00	$ 241,138,200.00
85 % Perc	$ 245,562,700.00	$ 227,120,300.00	$ 244,053,100.00
90 % Perc	$ 245,562,700.00	$ 227,120,300.00	$ 244,053,100.00
95 % Perc	$ 252,659,900.00	$ 234,738,900.00	$ 251,509,000.00

1. Percentile: Is a measure of position used in statistics that indicates, once the data have been ordered from lowest to highest, the value of the variable below which a given percentage of observations in a group is found.

The model shows a clear vision for the company to structure its improvement activities considering risk factors, identifying in the percentiles the behavior of the scenarios for time, and generating various decision alternatives considering risk factors under a comprehensive approach.

4 Discussion

Supply, demand, inflation, and corrosion are changing but predictable phenomena in the face of international procedures and regulations capable of mitigating their effect. The work presented does not stand out for generating a high immediate impact among the attendees but for establishing a comprehensive vision that will help reduce human, natural, and economic losses due to its ability to model, adapt, and evaluate operational, corrosive, and financial parameters, obtained Through field and laboratory experimentation, historical data and cognitive knowledge, which established a new collaborative vision of AI, which allows establishing management and planning of reliable best financial practices, which allows establishing maintenance and operational programs appropriate to the corrosive soil profile, pipeline conditions, cathodic protection, and inhibitors. The research supports its contribution to knowledge by thoroughly exploring previous works that generated a way forward, as shown in Table 6 below.

The application of MCS for life cycle analysis in oil and gas pipelines is not isolated cases due to its precision; however, the financial factor is not always considered (Caleyo et al. 2002; Hasan et al. 2012; Nesic et al. 2001; Younsi et al. 2013) as well as the cognitive knowledge of the experts who have interpreted and analyzed studies. Of mechanical integrity for years to establish parameters of corrosive growth (Abyani and Bahaari 2020; Ossai 2013; Ossai et al. 2015). Operating strategies determined by 30-year financial valuations through stochastic parameters of the sale price, operating costs, and remaining reserves estimate production rates (Willigers and Bratvold 2009); Calculate the cash flow around the life of a pipeline, it is necessary to consider a review of the literature and compilation of historical data, maintenance periods generated, conditions and defects in the pipeline (Parvizsedghy et al. 2015) but adding variables such as inflation, supply and demand establish more realistic estimates, in addition to the generation of financial scenarios identify probabilities of losses and profits for the company (Jones and Ferrari 2017).

The accuracy of the MCS is reliable and comparable, but we can conclude that financial models around corrosive effects contain a wide field of research (Leira et al. 2016; Park et al. 2018; Shabarchin and Tesfamariam 2017). Therefore, it is not advisable to compare results, if not to be homologous and persistent in the importance of considering all the factors involved in the Analysis and adding variables of interest. To increase effectiveness and reduce errors, it recommends including more variables that may arise from new incidents in the system and ensuring that the data guarantees representation and consistency.

This work is a pioneer for the Mexican oil and gas industry in the corrosion of pipelines. It is important to note that the main contribution of this research is to generate new knowledge to the current state-of-the-art through models that facilitate and support pipeline assessment procedures through AI in the world, which allows reconfiguring and generating new management methodologies in pipelines, guaranteeing operational reliability in its processes. It enables us to predict the most critical factors’ impact, improve the pipeline’s mechanical integrity, and calculate the associated economic viability.

5 Conclusions

The Analysis of the operational and economic risk of the PTS under corrosive effects through MCS interprets information and knowledge of 564 Km of soils from the latest cathodic protection study, essential to determine corrosive profiles of the PTS, considering supply and demand effects, with 1095 data on the price of the Mexican mix from 2016 to 2019, as well as monthly inflation data for the same period, to generate financial estimates through the discounted cash flow methodology, considering a wide range of risks and their correlation, incorporating changes over time, which is not a contribution fully comparable with the current works in the literature for its Innovation to model dynamic behaviors. The calculation of the net benefit of the research is not established to determine whether the failure in the models is acceptable according to the current state of knowledge but to interpret with variable flexibilities and results since some non-quantifiable criteria may be necessary for assessing the overall risk, such as the cost of protecting life and the natural environment.

Nomenclature

ASTM

American Society for Testing and Materials

AI

Artificial intelligence

ANN

Artificial neural network

API

American Petroleum Institute

BT

Barrels transported

BTD

Barrels transported per day

C

Corrosion

CK

Cognitive knowledge

C-PTS

Corrosion in pipeline transportation systems

CO₂

Sweet corrosion

COVID-19

Coronavirus diseases

C-AQ

Corrosion due to aquatic effects

C-TE

Corrosion by terrestrial effects

C-AT

Corrosion due to atmospheric effects

Co-MA

Corrective maintenance

DCF

Discounted cash flows

DSS

Decision support system

DPT

Daily hydrocarbon transportation program

E-SP

Effects of the sale price

EUAC

The equivalent uniform annual cost

FR

Financial risk

FI L-24″ 30″ Ф

Total hydrocarbon integrated length – 24 and 30 inches of diameter pipeline

FL-DTP

Loss of product of February from DTP

FES

Fuzzy expert system

GDP

Gross domestic product

HD

Historical data

IP

Inflation of prices

IM L-24″ Ф

Total integrated Maya hydrocarbon length – 24 inches of diameter pipeline

II L-30″ Ф

Total integrated Istmo hydrocarbon length – 30 inches of diameter pipeline

JL-DTP

Loss of product of January from DTP

ρ

The density of the liquid in kg m⁻³

PC

Pitting corrosion

PTS

Pipeline transportation system

Pre-MA

Preventive maintenance

Pro-MA

Proactive maintenance

L

Measured metal loss length

LB

Local burst

m

The oil phase, the oil–water mixture

M

Folias factor

ML-DTP

Loss of product of March from DTP

MCS

Monte Carlo simulation

Max

Maximum

MATLAB

MATrix LABoratory

MDC

Mendoza distribution center

MD

Millions of dollars

Min

Minimum

MI

Mechanical integrity

MIC

Microbiological corrosion

N

Number of values

NPV

Net present value

σ

Variance

OPEC

Organization of the Petroleum Exporting Countries

pH

Hydrogen potential

StdDev

Standard deviation

SCADA

Supervisory control and data acquisition

SMYS

Specified minimum yield strength of the material of the pipeline

t

Nominal wall thickness

TB

Thousands of barrels

TBD

Thousands of barrels per day

SLR

Systematic literary review

USD

American dollar

PTS

The pipeline transport system

RS

Regulations and standards

WALL

Pipeline involvement is internal (1) or external (−1)

y(t)

Normalized deterministic variables of the measures MCS

ỹ(t)

Normalized deterministic variables of the estimated MCS