Parametric Cost Model Research Articles

A cost modelling methodology based on machine learning for engineered-to-order products

Recent scientific studies are targeted at applying and assessing the effectiveness of Machine Learning (ML) approaches for cost estimation during the preliminary design phases. To train ML prediction models, comprehensive and structured datasets of historical data are required. This solution is inapplicable when such information is unavailable or sparse due to the lack of structured datasets. For engineered-to-order products, the number of historical records is often limited and strongly influenced by different purchasing or manufacturing strategies, thus requiring complex normalisation of such data.This method overcomes the above limitations by presenting an ML-based cost modelling methodology for the conceptual design that is applicable even when historical data are insufficient to train the prediction algorithms. The training dataset is generated through an analytical and automatic software tool for manufacturing cost estimation. Such a tool, starting from a 3D model of a product, can quickly and autonomously assess the related cost in different scenarios. An extensive and structured training dataset can be easily generated. The proposed methodology was based on CRISP-DM (Cross Industry Standard Process for Data Mining).Cost engineers of an Oil & Gas company used the method to develop parametric cost models for discs and spacers of an axial compressor. The solution guarantees lower error (7% vs 9%) and significant time-saving (minutes instead of hours) than estimations based on other approaches. Cost models are more comprehensive (capable of analysing different scenarios), explainable (not conceived as a black box), and self-learning (can be updated by extending the training dataset).

Seismic retrofit cost model for Italian masonry residential buildings to support territorial-scale risk analysis

Over the last 50 years, earthquakes forced the Italian Government to pay emergency, recovery and reconstruction costs amounting to 180 billion euros, i.e., about 3.6 billion euros per year. These figures underline the poor seismic resilience of Italy and therefore the need for sustainable risk mitigation plans at national level.Within this context, this study aims to provide a seismic retrofit cost model for Italian unreinforced masonry (URM) residential buildings, to support territorial-scale risk analysis. To this end, well-established intervention techniques, suitable for being analyzed on a large scale, were selected. These interventions include strengthening of masonry walls, stiffening and anchoring of floors, improving the wall system connection, and their possible combinations. First, for each of the selected interventions, the related cost items (obtained via a work-breakdown-structure procedure) were identified. Second, their unit cost was either retrieved from Italian price lists or estimated according to market analysis ad hoc conducted. Third, a quantity-survey-method approach was implemented to a set of 445 URM buildings to estimate the retrofit costs per square meter (€/m2) of investigated retrofit interventions. Lastly, based on the estimated cost of retrofitting of these buildings, parametric cost models were developed to inform seismic risk analyses at various territorial scales, depending on the available building information (i.e., exposure data).The cost models here developed and proposed can be used to support the design of cost-effective seismic mitigation measures, i.e., to identify retrofit solutions that are technically effective but also economically sustainable at large territorial scale.

Cost estimation for launch vehicle families considering uncertain market scenarios

When optimizing the staging of a launch vehicle, usually a reference mission with a fixed payload mass is used to determine the specific launch cost (or a surrogate thereof, e.g. total mass or dry mass). While this approach neglects suboptimal loading, it delivers sufficiently good results for the evaluation of an individual launcher if the design mission is well known. However, this approach does not capture the full range of payloads that a launch vehicle is expected to launch during its operational lifetime. In order to estimate the cost and subsequently optimize the staging of such launch vehicle families, a full launch scenario has to be considered. The cost of an individual launch vehicle loses importance and the deciding factor is the total cost associated with launching the specified scenario. However, estimates of the future market inevitably include large uncertainties. These uncertainties and their impact are ideally considered and quantified when assessing the specific recurring cost of a single launch vehicle. The consideration of a whole launch market scenario is especially relevant when evaluating a family of launch vehicles, an option currently being discussed for future European launchers. In this case, the distribution of the payloads onto the different members of the launch vehicle family has to be optimized in order to reach the lowest possible cost. Within this manuscript, a method that allows the determination of the recurring cost of such a launcher family based on a launch market scenario including uncertainties is proposed. Based on the launch numbers derived from the combinatorial optimization, parametric cost models for recurring and non-recurring cost are adapted and applied to three possible future launch vehicle families.

MACHINE LEARNING FOR PARAMETRIC COST ESTIMATION OF AXISYMMETRIC COMPONENTS

AbstractMachine learning (ML) is a well-established research topic in Industry 4.0 is boosting its adoption. ML is also used for manufacturing cost estimation during design. Such approaches are commonly used to estimate the cost of mass-produced parts. Many consolidated historical data are available for training the regression models. Unfortunately, very often, such a database of data is not available.The paper defines an ML approach for parametric cost estimation of axisymmetric components. The data for training the ML model derives from automatic software for analytically estimating the manufacturing cost. With a proper set of simulations, the tool can generate a large amount of data for training. The paper presents the steps for developing a parametric cost model using ML. The approach is based on CRoss Industry Standard Process for Data Mining method. The proposed method was used to develop one cost model (to estimate the total cost that considered raw material and manufacturing cost). The obtained Relative Error is 23.52% ± 1.37%, coherent with E2516 − 11, Standard Classification for Cost Estimate Classification System.

Value-Driven System Design of Utility-Scale Airborne Wind Energy

In the current auction-based electricity market, the design of utility-scale renewable energy systems has traditionally been driven by the levelised cost of energy (LCoE). However, the market is gradually moving towards a subsidy-free era, which will expose the power plant owners to the fluctuating prices of electricity. This paper presents a computational approach to account for the influence of time-varying electricity prices on the design of airborne wind energy (AWE) systems. The framework combines an analytical performance model, providing the power curve of the system, with a wind resource characterisation based on ERA5 reanalysis data. The resulting annual energy production (AEP) model is coupled with a parametric cost model based on reference prototype data from Ampyx Power B.V. extended by scaling laws. Ultimately, an energy price model using real-life data from the ENTSO-E platform maintained by the association of EU transmission system operators was used to estimate the revenue profile. This framework was then used to compare the performance of systems based on multiple economic metrics within a chosen design space. The simulation results confirmed the expected behaviour that the electricity produced at lower wind speeds has a higher value than that produced at higher wind speeds. To account for this electricity price dependency on wind speeds in the design process, we propose an economic metric defined as the levelised profit of energy (LPoE). This approach determines the trade-offs between designing a system that minimises cost and designing a system that maximises value.

Identifying key parameters impacting cost in large‐scale complex space programs using simulation‐based global sensitivity analysis

AbstractCost overruns averaging 45%–50% can occur during the acquisition process of large‐scale complex space programs. The factors that impact the cost overrun are frequently misunderstood and are not identified correctly. This paper investigates the impact of the parameters on the overall cost of a geosynchronous communication satellite program using model‐based global sensitivity analysis. A simulation model with the acquisition data was used to identify the key parameters within the system model that interact with the cost of the program. A system simulation model containing a physics‐based satellite model and a parametric cost model is utilized to conduct variance‐based sensitivity analysis. Data from selected acquisition reports are used to validate the system simulation model. Sobol' analysis is performed on the parameters associated with requirements of the satellite system, operations, and support to maintain the system, including the launch system and ground equipment. The results show that parameters related to the system‐based requirements significantly impact the program cost. These critical parameters, which influence the cost, lay the foundation to quantify the impact of system parameters and their uncertainty on the cost of the system using a simulation‐based model which will aid in the reduction of cost overruns during the design and development of future large‐scale complex engineered systems.

Parametric life cycle assessment modeling of reusable and single-use restaurant food container systems

Single-use packaging generates millions of tonnes of plastic waste per year – a circular economy waste-reduction strategy is to implement reusable container systems for restaurant takeout. We developed a parametric life cycle assessment (LCA) and cost model and used scenario analysis to study customer behavior effects on greenhouse gas (GHG) emissions, primary energy, water use, and cost of a reusable container system in Ann Arbor, Michigan. Under our base case scenario, the primary reusable container has lower impacts across most metrics than comparable single use containers. Benefits are sensitive to excess customer transportation; if 5% of customers make trips solely to return used containers, the reusable system has higher life cycle GHG emissions and primary energy use than single-use. Additionally, if a large fraction (close to 100%) of customers practice excess at-home washing of containers, the life cycle primary energy impacts will be greater than those of most single-use containers.

A specific manufacturing cost model for fast cost estimation in additive manufacturing

Additive manufacturing (AM) has been widely used for high-value-added parts. Therefore, cost models are needed for fast and accurate cost estimation during the AM design phase; otherwise, incorrect cost estimation may cause unexpected cost expenses. This work proposes a parametric cost model for AM using experimental and statistical methods. In this work, specific manufacturing costs (SMC) and build rates (cm3/h) are calculated based on experiments and approximated by regression models. For a fast cost estimation in AM, the build rate for a build task needs to be investigated, and then the SMC can be calculated using the regression model.

Research, development and production costs prediction parametric model for future civil hypersonic aircraft

Assessing the economic viability of new high-speed systems concepts since the early design phases is crucial for the success of future hypersonic vehicles including cruisers, reusable access-to-space and re-entry systems. Besides literature reports few parametric cost models for high-speed vehicles, all of them makes exclusively use of mass as parameter and none of the models moves beyond the vehicle level. This paper describes a new parametric cost estimation model which moves beyond the state-of-the-art methodologies (1) by integrating vehicle design and operational parameters (in addition to the mass) as cost drivers for the prediction of the vehicle life-cycle cost, (2) by introducing prediction margins accounting for the uncertainties on the data-driven correlations, (3) by providing a first estimate of the costs of every on-board subsystem, including combined cycle engines and multi-functional subsystems, (4) by increasing the granularity of the analysis up to technology level, thus providing a valuable support to Technology Roadmaping activities. The parametric cost estimation model has been refined and exploited in the context of the Horizon 2020 STRATOFLY project, where the technological, operational, environmental, and economic viability of a Mach 8 waverider concept have been investigated.

Методичний підхід до розрахунку вартості дослідно-конструкторських ро-біт зі створення космічних апаратів

The aim of this work is to develop a methodological approach to the development of a unified, common for all developers, computer-aided branch procedure for spacecraft development cost estimation based on a spacecraft cost model and practical recommendations on computer-aided branch procedure development. The cost parameters of the development, manufacturing, and operation of new spacecraft and their technical level (perfection) are the determining factors in competitiveness assessment. To decide on the advisability of starting or resuming the development of a new spacecraft, one has to correctly estimate the development cost. Using the standardized calculation method in estimating the development cost for new space hardware is unacceptable for lack of bug-free design and production documentation, which is the end product of any development activity. Parametric methods (the basic methods used to estimate the spacecraft development cost in the USA and Europe) cannot be used in the development of a high-quality procedure for spacecraft development cost estimation for lack of a branch statistical database on spacecraft development labor intensiveness and materials consumption at the State Space Agency of Ukraine. This calls for a nonstandard cost model of spacecraft development. The authors’ cost model is based on a method of componentwise analogy for simple spacecraft components, moving (up and down) along the edges of a weighted oriented tree graph that models the spacecraft technical structure, and fuzzy analysis methods. The tree graph Gi(V(С),D) models the spacecraft technical structure (V, C, and D are the sets of graph vertices and edges and spacecraft components, respectively; to each graph vertex there corresponds a spacecraft component). The paper presents a nonstandard cost model of spacecraft development, which in its essence is close to a nonlinear parametric cost model, and a scientific methodology for the development of an advanced branch procedure for spacecraft development cost calculation with component and stage detailing.

Pathways to Sustainability in Lunar Exploration Architectures

With an ever increasing geopolitical relevance of human lunar exploration, the world’s space agencies show an increasing interest in its implementation in the very near term. In August 2020 the International Space Exploration Coordination Group has published a supplement to its Global Exploration Roadmap that focuses on the envisioned future lunar architecture based on objectives of sustainability and benefit generation. Here we apply quantitative benefit assessment methods and a high-level parametric cost model to the architecture in order to analyze the effectiveness of two methods of increasing sustainability: in situ resource utilization and reusability. We find that both measures are effective, but require very different approaches to the investment strategy. Although the utilization of lunar resources achieves the lowest per-mission cost in the utilization phase, it requires significant upfront investments into the development of specific equipment and its transportation. Depending on the choice of model parameters, the return of investment in transporting additional equipment might be achieved only after more than 30 human missions in the utilization phase, even when neglecting the development cost of surface infrastructure. The investment in even partially reusable vehicles, however, is returned already after the first utilization mission, but it does not achieve the same performance in terms of lowering the cost of utilization missions if not combined with the use of lunar resources. We thus recommend space agencies to carefully consider the timing of investment into resource utilization with respect to the benefit it generates.

GPU Optimization for High-Quality Kinetic Fluid Simulation.

Fluid simulations are often performed using the incompressible Navier-Stokes equations (INSE), leading to sparse linear systems which are difficult to solve efficiently in parallel. Recently, kinetic methods based on the adaptive-central-moment multiple-relaxation-time (ACM-MRT) model [1], [2] have demonstrated impressive capabilities to simulate both laminar and turbulent flows, with quality matching or surpassing that of state-of-the-art INSE solvers. Furthermore, due to its local formulation, this method presents the opportunity for highly scalable implementations on parallel systems such as GPUs. However, an efficient ACM-MRT-based kinetic solver needs to overcome a number of computational challenges, especially when dealing with complex solids inside the fluid domain. In this article, we present multiple novel GPU optimization techniques to efficiently implement high-quality ACM-MRT-based kinetic fluid simulations in domains containing complex solids. Our techniques include a new communication-efficient data layout, a load-balanced immersed-boundary method, a multi-kernel launch method using a simplified formulation of ACM-MRT calculations to enable greater parallelism, and the integration of these techniques into a parametric cost model to enable automated prameter search to achieve optimal execution performance. We also extended our method to multi-GPU systems to enable large-scale simulations. To demonstrate the state-of-the-art performance and high visual quality of our solver, we present extensive experimental results and comparisons to other solvers.

Comparison of Artificial Intelligence Techniques for Project Conceptual Cost Prediction: A Case Study and Comparative Analysis

Developing a reliable parametric cost model at the conceptual stage of the project is crucial for project managers and decision makers. Existing methods, such as probabilistic and statistical algorithms have been developed for project cost prediction. However, these methods are unable to produce accurate results for conceptual cost prediction due to small and unstable data samples. Artificial intelligence (AI) and machine learning (ML) algorithms include numerous models and algorithms for supervised regression applications. Therefore, a comparative analysis for AI models is required to guide practitioners to the appropriate model. The article focuses on investigating 20 AI techniques which are conducted for conceptual cost modeling, such as fuzzy logic model, artificial neural networks, multiple regression analysis, case-based reasoning, hybrid models, such as genetic fuzzy model, and ensemble methods such as scalable boosting trees (XGBoost) and random forest. Field canals improvement projects (FCIPs) are used as an actual case study to analyze the performance of the applied ML models. Out of 20 AI techniques, the results show that the most accurate and suitable method is XGBoost with 9.091% and 0.929 based on mean absolute percentage error and adjusted R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> , respectively. Nonlinear adaptability, handling missing values and outliers, model interpretation, and uncertainty have been discussed for the 20 developed AI models. In addition, this study presents a publicly open dataset for FCIPs to be used for future models' validation and analysis.

A Parametric Cost Model for Estimating Civil Aircraft Line and Base Maintenance

The present work proposes a new parametric cost model in order to assess civil aircraft maintenance cost during the initial phase of aircraft design. The model focuses on defined maintenance activities, i.e. line and base maintenance, providing cost estimating relationships for the two cost items under study. The proposed cost model represents an updated alternative tool compared to the state-of-the-art models for civil aircraft maintenance costs estimation. It is based on current aircraft data provided by the International Air Transport Association and it may be useful to assess maintenance cost for new projects during aircraft conceptual design. The collinearity analysis is included in order to select independent cost drivers. The results of the developed model are in good accordance with other reference cost data provided by an independent source. The latter refer to aircraft models not considered in International Air Transport Association reports and adopting new technologies (such as composite structural materials) within their development. The analysis of the results obtained demonstrates that the application of proper corrective coefficients may lead to a maintenance cost estimation model for innovative aircraft configurations based on present architectures.

Blind validation study of parametric cost estimation tool SEER-H for NASA space missions

One of the primary parametric cost modeling tools used by NASA and the aerospace industry to estimate the development and production cost of future spacecraft hardware is SEER-H by Galorath. To date, no independent validation of this tool for cost estimation of space missions has been reported in the literature. In the present validation study, cost estimators used SEER-H to estimate the cost of twelve different past NASA science missions. The estimators were prevented from knowing the actual cost of the missions in an effort to minimize cognitive biases. The point estimates of SEER-H had an average error of 23%, median error of −0.3%, and a standard deviation of 43%. Nine of the twelve mission's actual costs fell within the 80% confidence interval of SEER's probabilistic estimates. Several factors independent of SEER that may have affected the accuracy of the results have been identified and are discussed; these include: uncertainty in the technical data used for the estimates, the methods used to estimate uncertainty in spacecraft component mass and numbers of prototypes, and the experience of the estimators.

Demonstrating set-based design techniques: an unmanned aerial vehicle case study

The Engineered Resilient Systems research program seeks to improve decision making in the Analysis of Alternatives process by leveraging model-based engineering (MBE) early in the design process to develop more resilient systems. Traditional tradespace exploration using point-based design often converges quickly to an initial baseline design concept with subsequent engineering changes to modify the design. However, this process can lead to significant cost growth if the initial concept is not able to meet requirements or if the revised design is not affordable. Enabled by MBE, set-based design (SBD) considers sets of all possible design concepts and down-selects design concepts to converge to a final design using insights into design trade-off analysis, modeling and simulation, and test data. Using a notional unmanned aerial vehicle case study with low-fidelity physics-based models and an open source Excel® add-in called SIPmath©, this research implements an integrated MBE trade-off analytics framework that simultaneously generates numerous SBDs using parametric performance and cost models and evaluates the designs in the value and cost tradespace. In addition, this research explores incorporating resilience quantification and uncertainty into SBD trade-off analysis. Future research is needed to validate the use of SBD with low-fidelity models for tradespace exploration in early system design.

Intelligent methodology for project conceptual cost prediction

Developing a reliable parametric cost model at the conceptual stage of the project is crucial for projects managers and decision makers. Several methodologies exist to develop a conceptual cost model. However, many gaps exist in the current methodologies such as depending only on experts ‘opinions and questionnaire survey to identify the project features, key cost drivers and developing deterministic predictive models without taking uncertainty nature into consideration. The main contribution of this study is developing an intelligent methodology for predicting the project cost at the conceptual stage. The proposed methodology can automatically identify key cost drivers and maintain uncertainty to predicted cost. Field canals improvement projects (FCIPs) are used as a case study to validate the proposed methodology. The selected methodology has applied quantitative approaches to identify the key cost drivers. In addition, the methodology has applied a genetic fuzzy model that automatically generates fuzzy rules to automatically predict the conceptual cost. Moreover, the results show a superior performance of the genetic fuzzy model than the traditional fuzzy model. In addition, this study presents a publicly open dataset for FCIPs to be used for future models validation and analysis.

A multi-objective sequential method for manufacturing cost and structural optimization of modular steel towers

This paper proposes a methodological approach for the multi-objective optimization of steel towers made from prefabricated cylindrical stacks that are typically used in the oil and gas sector. The goal is to support engineers in designing economical products while meeting structural requirements. The multi-objective optimization approach involves the minimization of the weights and costs related to the manufacturing and assembly phases. The method is based on three optimization levels. The first is used in the preliminary design phase when a company receives a request for proposal. Here, minimal information on the order is available, and the time available to formulate an offer is limited. Thus, parametric cost models and simplified 1-D geometries are used in the optimization loop performed by genetic algorithms. The second phase, the embodiment design phase, starts when an offer becomes an order based on the results of the first stage. Simplified shell geometries and advanced parametric cost models are used in the optimization loop, which present a restricted problem domain. In the last phase involving detailed design, a full 3-D computer-aided design model is generated, and specific finite-element method simulations are performed. The cost estimations, given the high levels of detail considered, are analytic and are performed using dedicated software.

Beyond hit-or-miss: a comparative study of synopses for similarity searching

A DBMS optimizer module takes its decisions by modeling the query costs upon the distribution of the data space. Cost modeling of similarity queries, however, requires the representation of distances’ rather than data distributions. Therefore, the finding of a suitable representation (or synopsis) for the distance distribution has a major impact in the optimization of similarity searches. In this study, we evaluate the quality of estimates drawn from five synopses of distinct paradigms regarding two common query criteria. Moreover, we embed the synopses into a new parametric cost model, called Stockpile, for the cost estimation of similarity queries on metric trees. The model uses the synopses estimation for calculating the probability of traversing a metric tree node, which defines the expected number of both disk accesses (I/O costs) and distance calculations (CPU costs). We performed an extensive set of experiments on real-world data sources regarding the estimates of each synopsis (and its parametric variations) by using paired ranking tests. In global terms, three synopses have outperformed their competitors regarding selectivity estimation, whereas two of them have also surpassed the others in the prediction of both I/O and CPU costs with respect to Stockpile model predictions. Additionally, results also indicate the choice of the most suitable synopsis may depend on characteristics of the distance distribution.

Environmental Parametric Cost Model in Oil and Gas EPC Contracts

This study aims at identifying the parameters that govern the environmental costs in oil and gas projects. An initial conceptual model was proposed. Next, the costs of environmental management work packages were estimated, separately and were applied in project control tools (WBS/CBS). Then, an environmental parametric cost model was designed to determine the environmental costs and relevant weighting factors. The suggested model can be considered as an innovative approach to designate the environmental indicators in oil and gas projects. The validity of variables was investigated based on Delphi method. The results indicated that the project environmental management’s weighting factor is 0.87% of total project’s weighting factor.