Simulation-Based Pipeline Route Selection Tool

Decisions that are made during the process of selecting the route of a pipeline are very critical to the project. Unfortunately, the decision is often complicated by numerous variables that must be considered, as well as the uncertainty of estimated costs. When choosing a pipeline route, a project manager must balance the likely capital cost of the pipeline with the risks inherent in the chosen route. Ideally, a project manager would investigate numerous alternatives to fully explore the merits of various pipeline routes (including the level of risk) before making his or her final decision. This article presents a simulation tool that a project manager can use to model the costs associated with various pipeline routes effectively and efficiently. The model is designed to be user friendly by replicating the usual decision-making process as much as possible. The model uses a graphical interface that promotes the rapid analysis of numerous alternatives and provides opportunities to investigate in detail the various aspects of a pipeline route. The model output includes a calculation of the costs of the alternative, a statistical analysis of the risks of the project, and information that the project manager can use with confidence to establish the pipeline target price.

Tools and techniques are very useful for if-then scenarios when assessing various alternatives before a project's stakeholders make the final decision, especially when the stakeholders have little or no experience with similar projects.

For instance, the stakeholders' "gut feelings" for analyzing the risks and uncertainties could be converted using tools and techniques. It must be emphasized, however, that using tools alone cannot lead to correct decisions. Human judgment must be invoked in the decision-making process.

Project selection is the process of evaluating individual projects, to choose the right project, based on an analysis so that the company's objectives will be achieved. It involves a thorough analysis including the most important financial aspects to determine the optimum project among all the alternatives.

The high capital cost of pipelines makes the selection of the pipeline route the single most critical decision of the project. In fact, economic considerations are so important that the American Society of Civil Engineers, in its 1998 publication, Pipeline Route Selection for Rural and Cross-Country Pipelines, devotes an entire chapter to this topic.

Choosing a route is often a process of balancing the goal of minimizing pipeline length (the largest driver of cost) with the additional costs associated with routes crossing challenging terrain or requiring extensive environmental remediation.

To mininize costs, selecting a cross-country petroleum pipeline route selection is governed by the following goals.

• Establishing the shortest possible route, connecting originating, intermediate, and terminal locations;
• Ensuring, as far as practicable, accessibility during operation and maintenance stages;
• Preserving the ecological balance and minimizing environmental damage;
• Avoiding populated areas for public safety reasons;
• Keeping rail, road, river, and canal crossings to a bare minimum;
• Avoiding hilly or rocky terrain;
• Avoiding a route running parallel to high-voltage transmission lines or DC circuits;
• Using an existing right-of-way, if possible;
• Avoiding other obstacles, such as wells, houses, orchards, lakes, or ponds.

The ability to estimate the costs of alternate routes is critical for making correct decisions, because even small adjustments to the route can result in significant changes in capital cost.

The wide variation in terrain that a potential route may encounter requires that a large number of alternatives must be explored before making the final decision. The process of estimating the costs for a given pipelineroute must be both efficient and flexible enough to handle the uncertainty inherent in estimating costs.

Since 1965, many high technology items have been developed to help the project manager and other project team members to be more productive in their jobs. The use of these technologies has allowed many new procedures that were previously not possible.

Nevertheless, in spite of these new technologies, there are still no logical or structured models that will properly assess risks and uncertainties in pipeline route selection. Deterministic decision trees and the expected value (EV) concept have been used by many in the past for decision making problems. However, EV has been criticized for being unsuitable for one-time decisions, though it is suitable for repetitive decisions.

This argument provides the opportunity to develop a tool with which decision makers can evaluate several alternatives easily to determine a one time decision. Simulation techniques are very useful to analyze several alternatives with ease.

This article discusses a pipeline route selection analysis framework based on a special purpose simulation (SPS) concept developed by A. Agbulos and others. The SPS is a user-friendly model that simulates the cost drivers of the pipeline at all stages of project planning. This allows an organization to investigate alternatives and make decisions at the appropriate time.

Preliminary routes can be eliminated immediately, or the design of a chosen route can be refined, based on the cost information generated by the model. The analysis framework described in this article allows an organization to make the right decision quickly and effectively by graphically modeling various pipeline routes while considering the variables that affect the project's capital cost.

PIPELINE ROUTE SELECTION PROCESS

The pipeline route selection process follows the traditional project planning phases of concept, feasibility, and preliminary project planning.

The typical route selection of a pipeline progresses through three stages: conceptual, feasibility, and detailed route selection with each subsequent stage requiring information of greater detail and precision for decision-making.

The detailed route selection stage is the lengthiest and often involves significant engineering study, design, and review.

A pipeline project typically begins in the conceptual stage with the sponsor organization exploring the feasibility of constructing a pipeline to deliver the product from a source to a delivery point.

The pipeline may consist of a single pipe of constant diameter; a single pipe of varying diameter; or, in the case of multiple tie-in points, a network of pipes. When there is flexibility in locating the tie-in points for thepipeline, the project team members begin the selection of a route by exploring alternatives for the location of the start and the finish points of the pipeline.

Usually the project team possesses only a general knowledge of the terrain to be traversed by the pipeline. The preliminary tie-in points, the size (pipe diameter and length), and the network configuration for thepipeline are usually determined using rule-of-thumb, single number, cost estimates in conjunction with an hydraulic engineering model.

The next stage, feasibility, uses information and tools such as maps, engineering drawings, schematics, and aerial photographs to begin refining the route choice.

At a macro level, routes are chosen that minimize interference with existing facilities, structures, dwellings, or environmentally sensitive areas.

The tie-in points, the size (pipe diameter and length), and the network configuration for the pipeline are confirmed with the hydraulic engineering model. Alternatives are evaluated using route-specific cost estimates developed from historical costs.

The final stage, the detailed route selection, refines the details of the route by considering the terrain (such as hills, rivers, swamps) that may be crossed. Detailed investigation is carried out by flying, driving, or walking along the potential routes.

Sometimes, a preliminary survey will be carried out to locate potential routes. During this stage, decisions are made about how to overcome obstacles such as roads, railways, rivers, and hills.

Soil conditions are also evaluated to determine their effect on the pipeline design, such as the need for buoyancy control (i.e., the use of weights to prevent the pipe from floating in wet terrain), and the pipe route is adjusted to minimize the final cost.

Detailed cost estimates are developed to assist in deciding on the alternative that minimizes cost and manages risk.

In general, the pipeline route process follows the three stages sequentially with rapid progress through the first two stages if the pipeline is not complex. However, information discovered during the detailed routeselection stage could lead to the reevaluation of the feasibility of a chosen route.

MODELING OF THE PIPELINE ROUTE SELECTION PROCESS

The pipeline route selection model was developed on a Simphony platform, which is an SPS-based environment developed by J. Ruwanpura and others.

SPS based models have been successfully developed in many domains. A few samples include tunnel construction operations, project selection, and waste management.

Because of the powerful features in the simphony environment, it is used to model the pipe route selectionprocess and to calculate the costs associated with each alternative.

Simphony is a Microsoft Windows® based computer system developed to provide an environment for the development and use of SPS tools.

Simphony uses an, "object oriented application framework," which provides a structured approach for building a new template. The services provided under the framework include modeling elements, a discrete event simulation engine, a trace manager, random number sampling, statistical collection and analysis, external database access, a graphical user interface, and report generation.

Purpose of Simulation for Pipeline Route Selection Process

The purpose of using simulation in the pipeline route selection process is to provide a tool for pipeline project managers to explore numerous alternatives before making final decisions about the route. Simulation allows users to explore alternatives quickly and inexpensively, resulting in decisions based on rigorous cost analysis rather than "seat-of-the-pants" decision-making.

The tool explained in this article has the following features and benefits.

• Has flexibility and ease of use: The users are given the option to create a graphical-based model, which is familiar to the users in the pipeline projects. One of the advantages of the SPS models is that people who are experts in the model's domain can use the model with very little training [14]. However, its disadvantage is that the model developers must consider the end users, who may not be experts in the model's domain. The developers, therefore, have the challenge of creating a model that is familiar to all users.
• Simulates the real-life route selection process as closely as possible: Real world systems are so complex that some of these systems are virtually impossible to model and solve mathematically [11]. Numerical and computer-based simulations can be used to imitate the system's behavior. A model is defined as a representation of a system created for the purpose of studying the system. Many simulationists, including Francis Hartman's as outlined in, Don't Park Your Brain Outside: A Practical Guide to Improving Shareholder Value With Smart Project Management [8], assert that it is not necessary to consider all a system's details because a model is a substitute and a simplification of a system; nonetheless, the model should be sufficiently detailed to permit valid conclusions to be drawn for the real system. The model that is explained in this article considers the real project requirements for modeling.
• Allows for both stochastic and deterministic inputs to model cost uncertainty: If the risks and uncertainties need to be included in decisions, both stochastic and deterministic inputs are required. The stochastic inputs may include provisions to deal with potential risks and uncertainties.
• Highlights the cost impacts of the key cost drivers in the pipeline: The key cost drivers have to be identified before a report can be produced. The model outlined in this article highlights the real cost impact of key drivers which are explained later in the article.
• Provides outputs that can be used to make decisions about the pipeline route: The ability to create customer required outputs is a key issue for developing many if-then scenarios and evaluating the outputs related to each scenario during decision making.

The graphical user interface and the report generation features in this tool promote interaction with the model that is familiar to the user in real life construction of pipeline projects. The modeling elements and statistical analysis features support a model that can be used to generate a myriad of alternatives efficiently. Adding more elements to a simulation or changing a few input variables can quickly create a new alternative. The user has the option of using as many elements and input variables as necessary to model the level of detail required for the decision to be made.

Concept and Contents of the Pipeline Route Selection Process Simulation

The modeling sequence simulates the three stages used in the traditional route selection process. The initial conceptual stage of modeling a pipeline route is often used to determine a pipeline's feasibility. It is usually completed quickly with a minimum number of modeling elements and with data comprising rule-of-thumb costs. The configuration of the pipeline, including pipe diameter and approximate lengths, is more important than the specifics of each route.

The next stage (feasibility) adds additional elements to the basic model to capture the major cost drivers inherent in the pipeline configuration determined in the conceptual stage. The routes are modeled using additional modeling elements in the template and data based on a general knowledge of the terrain conditions. The intent of this stage is to confirm the original configuration and make decisions about the route based on information that is available without detailed engineering studies.

In the final stage (detailed route selection stage), the model is refined to capture the costs of the actual physical conditions encountered along a pipeline route. Additional modeling elements are added as needed, and the cost parameters are refined based on engineering studies and design choices.

The pipeline route is adjusted as the detailed engineering design of the pipelines determines the costs of the obstacles that may be encountered. In this stage, the pipeline route is more likely to be refined than radically changed. However, there is always the possibility that information about soil conditions or other issues could require that the pipeline configuration be revised.

The estimated costs used to evaluate pipeline route alternatives often contain a degree of uncertainty; as a result, the model must simulate the range and distribution of the costs.

The Simphony framework supports models that accept stochastic input variables that are used to produce a final cost. These input variables can use a number of distributions, such as normal, uniform, beta, triangular, and exponential, to model the uncertainty in the cost estimates. The input variables with a high degree of certainty can be input as deterministic.

The main cost drivers for a pipeline can be broken into three main areas: engineering, materials, and construction. The costs in these areas are significantly impacted by the length and diameter of the pipe, the terrain (soil conditions and buoyancy control), and the crossings (river, road, railway or foreign pipeline). Table 1 contains examples of the major cost drivers of a pipeline. The model summarizes the cost information into categories that reflect the impacts of the various cost drivers.

There are various customized outputs in this model that are used to make decisions about the selection of apipeline route. The outputs include statistics that provide a mean cost and a statistical analysis of the uncertainty of the final cost. Knowing the cost and the uncertainty of the various components of the pipelineallows the decision-maker to refine the alternatives to minimize cost and manage risk.

THE STRUCTURE AND LAYOUT OF THE SIMULATION TEMPLATE

The objective of the simulation template is to provide a pipeline project manager with an easy method for modeling pipeline route alternatives to determine the route with the minimum capital cost while managing risk. The pipeline route selection template has the following features that assist the user in creating a simulation that produces relevant results effectively and efficiently.

• Modeling Elements-The template contains modeling elements that are readily identifiable as components in an actual pipeline. The elements contain attributes that allow the user to model the cost drivers inherent in the components of a pipeline. The variety of elements allows the user to "build up" a model that appears similar to an actual pipeline and explores the cost effect of each of the components.
• Input Parameters-The input parameters for the elements have been chosen for their similarity to the type of information developed when choosing a route. Some parameters assist the decision-maker in keeping track of the choices that have been made in a specific alternative (e.g., type of buoyancy control) while others are used to calculate the cost of the alternative. Many of these parameters offer the opportunity to model the uncertainty associated with a specific cost estimate using a statistical distribution. Sample inputs were gathered based on the experience of one of the author's of this article.
• Outputs-Each element has its own outputs and statistics to allow the user to focus on the cost impact of each component of the pipeline. There are also summary level outputs and statistics so that alternatives can be compared using total project costs without viewing the entire model.
• Flexibility and Scalability-The pipeline route selection model has three levels in a hierarchical relationship. This structure provides the user with a great deal of flexibility when creating the model. In a very simplepipeline, the model could consist of one of each type of element. Conversely, in a complex pipeline, there could be numerous pipeline segment elements used to model complex configurations and numerous terrain elements used to model every different component of the segment in detail.

The template contains three levels of hierarchy. The pipeline parent element (level "a" in figure 2), which is the highest in the hierarchy, accumulates and reports project information. The pipeline segment elements (level "b") are next and allow the creation of the pipeline network. The final level in the hierarchy comprises the terrain elements (level "c") that model the physical characteristics of the route, each with its own unique attributes.

The single modeling element in level "a" contains global input parameters, such as an alternative name, that are used to identify the alternative that is being modeled. The element also contains input parameters that are used to specify how engineering and operating costs are calculated. Engineering costs can be calculated from the engineering costs in each terrain element, as a percentage of the material costs, or entered as a lump sum input parameter. Operating costs can be calculated as a percentage of material costs or entered as a lump sum input parameter. The results of the model (costs and statistics) are summarized from the other elements lower in the hierarchy and reported as total project outputs. Table 2 shows the list of all the parameters in level "a."

The pipeline segment elements in level "b" specify the pipe's attributes and allow the user to model the configuration of the pipeline network. The pipeline segment element contains input parameters, such as the segment name, that are used to identify the segment that is being modeled. The element also contains input parameters that are used to specify the physical attributes of the pipe, such as pipe diameter, pipe wall thickness, percentage of segment requiring clearing, and clearing cost. The results of the model (costs and statistics) are summarized from the terrain elements lower in the hierarchy and reported as total segment outputs. The outputs include segment costs for engineering, pipe, material, clearing, and annual fuel.

There are several modeling elements in level "c." The pipeline start element is connected to the first pipe segment element of each branch of the pipeline network in this level of the model and creates the entities that are used to trigger the calculations in the model. The pipeline end element is the last pipe segment element in this level of the model and calculates the statistics for the total project and destroys the entities. These elements do not require inputs and do not produce any outputs.

The terrain elements model the physical characteristics of the pipeline route. Each terrain element represents a component of a pipeline that can be a significant cost driver in a project and contains input parameters and produces outputs and statistics unique to that component.

Three of the terrain elements are used to model the main types of crossings encountered when constructing apipeline: foreign pipelines, rivers and roads, or railways. Each element allows the user to name the crossing; specify the type, method, and length of crossing; and input the engineering, material, and construction costs.

The outputs from the elements consist of a statistical analysis of the engineering, material, construction, and total element costs. The river crossing element also allows the user to specify the type, percentage of crossing requiring river weights, and cost of the buoyancy control to be used, thus resulting in a statistical analysis of the cost of buoyancy control. The plain pipe element is used to model the situation when the pipeline route is traversing the terrain without encountering any obstacles. Any obstacle that the terrain does contain, however, has a large impact on the pipeline's costs. The majority of the costs come from the plain pipe element. A number of the plain pipe elements can be used in the same segment to model the changing conditions along the route, such as sections that involve the following.

• Require topsoil stripping to protect cultivated land.
• Traverse populated areas and require pipe with a thicker wall thickness.
• Cross areas of permafrost requiring special construction techniques, such as minimizing the amount of tree cover removed to ensure the frost does not thaw.
• Require buoyancy control, such as swamp weights, to cross muskeg or swamp.

This element is used to capture the engineering, material, and construction costs of all these situations. A sample of these inputs include percentage of length requiring topsoil stripping, cost of topsoil stripping per meter, percentage of length containing permafrost, cost of permafrost construction ($/m), buoyancy control method, percentage of length requiring buoyancy control, and cost of buoyancy control ($/m).

Two other terrain elements are used to model the main types of flow control facilities used on a pipeline: valve and pump/compressor sites. These elements allow the user to name the site; specify the type; and input the engineering, material, and construction costs.

The outputs from the elements consist of a statistical analysis of the engineering, material, construction, and total element costs. The pump/ compressor site element also allows the user to specify the power and fuel costs for the site, resulting in a statistical analysis of annual fuel costs.

This simulation also provides a custom element that allows the user to model a user-defined component of thepipeline if a new condition is available in the route. This element is used to capture the engineering, material, and construction costs of a user-defined facility.

SAMPLE ANALYSIS USING A HYPOTHETICAL CASE

The model was used to evaluate two pipeline routes for a 24-inch diameter natural gas pipeline. The pipelinewas designed to deliver gas from a gas plant and transport it to the mainline of a natural gas transmission company approximately 80 km away.

The pipeline routes are in an uninhabited area consisting of large tracts of muskeg and forest. Table 3 lists the major characteristics of both alternatives.

The input parameters are representative of the type of cost estimates that a pipeline project manager would be using during the pipeline route selection process. The range of the cost estimates is indicative of the uncertainty present when choosing the pipeline route in this type of environment using estimates based on historical costs. Using a Monte Carlo analysis on the cost estimates provides an indication of the project's risk profile [16]. The pipeline project manager can use this risk profile to enhance his or her decisions about thepipeline route by balancing cost and risk.

Conceptual Stage

At the conceptual stage, both alternatives are very similar. The only difference between the alternatives is that the pipeline route in alternative 2 had been increased from 83 km to 85 km to take advantage of potentially better soil conditions in the final section.

Each model consists of a pipeline parent element, a pipeline segment element, and three plain pipe elements. Figure 3 shows the structure of the model and all inputs at this stage are shown in table 4. At this stage, alternative 1 is the least costly because the greater length of alternative 2 increases the pipe and construction costs. Details of the outputs in all three stages are described later in this article.

Feasibility Stage

Additional components of the alternatives are modeled in the feasibility stage. The clearing requirement is added to the pipeline segment, and three valve installation elements and two river crossing elements are added to the terrain elements.

The percentage of clearing and thecost of valve installation is the same for both alternatives. Alternative 2 uses the open-cut method to cross both rivers, while alternative 1 uses one open-cut crossing and one horizontal directional drilled (HDD) crossing.

All other parameters from the conceptual stage remain the same. See figure 4 for the structure of the model and table 5 for the additional inputs.

At this stage, alternative 1 remains the least costly because the greater cost of the HDD river crossing was not large enough to offset the increased pipe and construction costs from alternative 2's greater length. However, alternative 1's uncertainty has become greater than alternative 2's, because of the wide range of possible costs associated with a HDD river crossing

Detailed Route Selection Stage

At this stage, preliminary engineering studies in the form of route reconnaissance and soil sampling have been completed to determine the buoyancy control requirements for the pipeline. The buoyancy control requirements for the first two plain pipe segments are very similar.

In the third segment, the better soil conditions of alternative 2's pipeline route reduce the amount of buoyancy control required and allow the use of screw anchors rather than the more expensive bolt-on-weights, which marginally impacts the cost. All other parameters from the feasibility stage remain the same. See figure 4 for the structure of the model and table 6 for the additional inputs.

At the final stage, alternative 2 has become the least costly because of the buoyancy control for this particular project. The reduced cost of buoyancy control in alternate 2 has offset an increase of 2 km of pipeline. Additionally, the uncertainty of alternative 1 remains greater, which increases the attractiveness of alternative 2.

Final Results and Analysis

The analysis produced total project costs for each alternative that have a range of plus or minus 5 percent, which are appropriate for the detailed route selection stage of the pipeline project. Although the difference between the mean costs of the alternatives are marginal, the lowest mean cost, lower level of uncertainty, and lower maximum cost of alternative 2 make it a clear choice.

At this point in the simulation, the project engineer could use the cumulative probability curve to specify the target (or expected) price for the project based on the desired level of probability. For example, if the project manager wanted a 80 percent confidence level, then he or she would set the target price for alternative 1 at $40,600,000 and alternative 2 at $40,300,000.

Figure 5 shows the cumulative probability curves for both alternatives (X axis represents the cost and Y-axis represents the cumulative probability of outcome or confidence). Because the costs are close, the project engineer may be willing to invest in some further studies to refine the analysis by confirming the assumptions inherent in the cost estimates. For example, the project engineer could afford to spend up to $300,000 on soil sampling to verify that the buoyancy control method chosen for alternative 2 is practical.

The SPS pipeline route selection tool presented in this article provides a very comprehensive and useful tool that can assist a project manager's to analyze several alternatives to select the most suitable pipeline route during the most critical phases of any pipeline project.

The results of the model provide the project manager with the range of the costs of the pipelines for each alternative and its risk associated to determine the level of confidence for each alternative. The tool enhances the effectiveness of the decisionmaking process in many ways because it provides the following.

• Graphical Interface and User Friendliness-This tool provides a graphical user interface that is familiar to the expert in pipeline projects where he/she requires minimal training to use the tool to analyze outputs based on varied inputs to arrive at decisions;
• Flexibility-Almost any type of pipeline configuration can be modeled in as much or as little detail as required because of the incorporation of a high degree of flexibility within the modeling framework. The model also provides customized summary outputs so that the project manager can quickly assess the merits of each alternative without reviewing all the details.
• Modeling Uncertainty-This tool demands the user to consider real-life uncertainties by allowing the user to either input deterministic or stochastic input parameters based on his or her level of understanding of the project uncertainties. The cost estimates requires the inputs from a comprehensive list of stochastic variables and options that allow a range of costs to be input;
• Range Estimate-The tool allows the user to perform a Monte Carlo simulation to determine the most likely costs and to quantify the uncertainty using the range estimate fundamentals. In addition, this tool allows the user to explore the costs and uncertainties related to each element in the pipeline so that the areas of high cost and/or high risk can be identified.

Future enhancements could include the incorporation of a net present value (NPV) calculation that uses the existing capital, fuel, and operating cost outputs to provide a true lifecycle cost of the pipeline route. Comparison of the NPV of various routes would be especially useful in the conceptual and feasibility stages of the process because the trade-off between larger pipe sizes and additional compression (i.e., higher fuel and operating costs) is often done when the pipeline configuration is being decided.

The proper capturing of risks and uncertainties on pipeline projects will definitely produce better project plans for the stakeholders. This tools summarized in this article have ample evidence to show the power of simulation to assist in decision making.

The University of Calgary's project management and simulation research group has developed many simulation tools to assist in the decision making in any project. These include the decision support simulation tool (integrated framework to model scheduling, estimating, and decision trees); situationbased modeling to identify the triggering situations that impact the productivity of construction and to improve productivity; and various other construction process modeling, such as HDD, drilling, and pipe installations.

Symber : Hirst, Gary; Ruwanpura, Janaka Y. "Simulation-Based Pipeline Route Selection Tool". 30 Januari 2014. http://search.proquest.com/docview/220455202?accountid=31562