SPECIAL ISSUE PAPER

477

Life cycle cost modelling as an aircraft design support tool P Thokala∗ , J Scanlan, and A Chipperfield Computational Engineering Design Centre, School of Engineering Sciences, University of Southampton, Southampton, UK The manuscript was received on 3 April 2009 and was accepted after revision for publication on 16 September 2009. DOI: 10.1243/09544100JAERO574

Abstract: This article summarizes the work that has been carried out as part of the FLAVIIR project, a 5-year research programme looking at technologies for future unmanned air vehicles (UAVs). This is a UK Engineering and Physical Sciences Research Council funded project sponsored by BAE systems. A framework to estimate the life cycle cost of UAVs is presented. The acquisition costs are estimated using a hierarchical structure and a discrete simulation model is used to estimate the maintenance and operation costs. The architecture to estimate the life cycle cost and the associated models are described. A framework is presented in which the cost models developed can be integrated into the design process to facilitate the comparison between different configurations. It is then demonstrated how this framework can be used to perform trade-off analysis and cost-based optimization. Keywords: life cycle cost, cost modelling

1

INTRODUCTION

In the past, technology has been the dominant driver in the design process, but there has always been a demand for cost reduction in the aircraft industry to satisfy the customer’s needs. There has been a realization by the aircraft producers that cost reduction needs to be tackled at the conceptual design phase because it is widely believed that typically 70% of the total avoidable cost is controllable at this stage [1]. Life cycle cost (LCC) is increasingly being used when making procurement decisions or to assess the competitiveness of a product design. The LCC is concerned with the overall cost of a product from its conception up to, and including, its disposal. Asiedu and Gu [2] divided the total product cost or LCC into four distinctive phases: (a) research and development costs; (b) production and construction costs; ∗ Corresponding

author: Computational Engineering Design Cen-

tre, School of Engineering Sciences, University of Southampton, Southampton, SO17 1BJ, UK. email: [email protected] JAERO574

(c) operations and maintenance costs; (d) retirement and disposal costs. The National Aeronautics and Space Administration (NASA) selected the international space station (ISS) systems with emphasis on near-term costs rather than the total programme costs, which resulted in significant cost overruns for the space station [3]. Thus, LCC is of interest when an estimate is to be used in a trade-off study or a design selection process. Cost is an important factor in the aerospace engineering design and it should have a more directly influential role; for example, cost should be a part of an integrated conceptual design process. Differential product evaluation with regards to cost, technology, reliability, and maintainability is important for a better product design. Because cost is not known in advance of production, a cost estimation system is required. Scanlan et al. [4] have identified the need for detailed and reliable cost information for the optimization of a product design. The challenge is to look into all of the aspects of cost and to link these into the design decision-making process at the conceptual stage so that a design-oriented capability can be used to implement product changes that reduce cost. Proc. IMechE Vol. 224 Part G: J. Aerospace Engineering

478

2

P Thokala, J Scanlan, and A Chipperfield

COST ENGINEERING

Cost engineering can be described as the application of scientific and engineering principles and techniques to the problems of cost estimation and cost control. An overview of the state-of-the-art and future trends in the field of cost estimation is given by Layer [5]. A wide range of studies have been performed on the estimation of manufacturing costs. Ben-Arieh [6] estimates the manufacturing cost focusing on the costs for set-up, machining, and raw material costs. Similarly, Stockton et al. [7] elaborates the development of time estimating models for advanced composite manufacturing processes. Studies have also been performed on estimating the manufacturing costs from design specifications. Rehman and Guenov [8] describe a method for modelling manufacturing costs throughout the design phase of a product’s life cycle, from conceptual to detail design. A framework for estimating manufacturing cost from geometric design data is outlined by Wei and Egbelu [9] and a model designed to estimate process cost directly from the design specifications has been developed by Kulkarni and Bao [10]. However, the efforts of most of these studies are concentrated on a particular aspect or a specific method of manufacturing. Cost is an important attribute of any product and is a prime factor in the engineering design process. Dean is well known for promoting considerations for design to cost within NASA [11]. Curran et al. [1] present a generic model of a cost estimating tool that can be used within the design domain. Bao and Samareh [12] demonstrate the use of process-based manufacturing and assembly cost models in a traditional performance focused multi-disciplinary design and optimization (MDO) process. Gantois and Morris [13] also presented a multi-level MDO process implemented through a hierarchical system with cost at the top level and applied the method to a civil aircraft wing to achieve a minimum cost design. These multidisciplinary design studies are performed by taking only the manufacturing costs into account. However, life cycle cost models are needed when an estimate is to be used in a performance trade-off study of a process or activity. Studies have also been performed on life cycle cost-integrated design (i.e. assessing the trade-off between technologies or materials using a life cycle approach). Marx et al. [14] linked MDO to life cycle analysis by defining high level objective functions that encompass the life cycle needs of the aircraft. Hackney and Neufville [15] described a life cycle model for performing comparisons of emissions, costs, and energy efficiency trade-offs for alternative fuel vehicles. Similarly, Bruening [16] conducted a study to define payoffs in terms of mission capability and system level life cycle costs associated with implementing three different propulsion system development approaches into an unmanned combat Proc. IMechE Vol. 224 Part G: J. Aerospace Engineering

air vehicle. However, these studies used analytical models to estimate the maintenance and operation costs. Simulation models can assist in accurate estimation of the operation and maintenance costs compared to analytical models, as solutions to complex systems cannot be provided using analytical methods. A critical evaluation of simulation studies of maintenance systems is performed by Andijani and Duffuaa [17]. Simulation approaches have been used in studying availability or supportability requirements of different weapon systems. For example, Upadhya and Srinivasan [18] and Cook and DiNicola [19] have modelled the availability of fleets of aircraft and helicopters, respectively. A simulation model that evaluates the effect of operating decisions on the maintenance turntimes is shown by Cobb [20]. Similarly, a simulation model has been developed by Adamides et al. [21] to assess the performance of military aircraft engine maintenance system. An integrated mission-based life cycle cost modelling environment is presented by Orsagh et al. [22] and Sandberg et al. [23] predicted LCC using a simulation model for the conceptual design of jet engine components. 3

EXISTING COST MODELS

After reviewing the current state of the art in cost modelling, a number of limitations are apparent. The following section outlines some of the limitations of existing cost models and techniques for improvement. 3.1

Limitations of existing cost models

1. Most cost models are concerned with a particular element of cost instead of looking at the holistic cost architecture. Similarly, modelling is directed towards a particular stage of the life cycle instead of the whole LCC. This leads to highly product specific cost models rather than generic models. 2. Most cost models give a ‘single number’ as the cost estimate and cannot identify which are the costdriving elements in the product description. It is realistic to have a range of cost estimates rather than a discrete value. 3. Most cost models are not easily auditable and are not transparent. Cost models should show variables and parameters that have the most impact on the design so that comparisons between alternative products can be performed. 4. The integration of cost models in the design process for concurrent design is not solved satisfactorily. Most cost models are used for estimating the costs rather than updating the product design for cost reduction. In addition, automation needs to be considered if multi-disciplinary analysis or trade-off studies need to be performed. JAERO574

Life cycle cost modelling as an aircraft design support tool

Fig. 1

3.2

Cost estimation

Improving cost estimation

Cost estimation has been used extensively for many years in the aircraft industry but there is a need for further research. Figure 1 shows cost estimation in the past and present and an idea for future cost estimation. Cost models should be complete and generic (i.e. they should give the whole picture and should be applicable for a range of cases). They should be transparent (i.e. costs should be traced back to the driving elements in the product description). Also, the cost model needs to be linked to product definition so that any change in product details is reflected in the cost model. In order for the cost model to be relevant, it should be integrated with design tools so that a design decision support tool can be achieved. Statistical analysis should be combined with cost estimation in order to predict the cost estimation uncertainty and where it is attributed. Cost models must be built with adequate visualization tools because cost information delivered in a sub-optimal format makes the cost analysis prone to misinterpretation. Finally, cost models have to be easily understandable and accessible as this assists designers in making modifications to the design for cost reduction in early stages. In this article, all these characteristics are incorporated into the framework estimating the LCC for a generic unmanned air vehicle (UAV).

4

GENERIC AIRCRAFT PRODUCT DEFINITION

The framework developed here has the capability to estimate the costs of any given aircraft. This is achieved by having product definition as input to the cost model so that any change in the design is reflected in the JAERO574

479

estimated cost. After careful consideration, the product definition of an aircraft can be broadly classified into explicit and implicit product definition. Explicit product definition includes the design parameters whose effects on the cost are easily recognizable (i.e. a straightforward relationship between cost and the design parameter can be easily identified). Explicit product definition includes the geometry parameters (i.e. dimensions of the design), material type, power plant, etc. For example, a change of the design dimensions leads to a change in raw-material and manufacturing costs. It can be easily observed that there is an explicit relationship between cost and design dimensions, thus making these design parameters part of the explicit product definition. Implicit product definition on the other hand includes design parameters whose affects on the cost are not easily identifiable (i.e. a straightforward relationship between cost and the design parameter cannot be easily observed). For example, the affect of maximum velocity of the aircraft on the cost is not easily apparent. However, the battle damage is dependent on the speed of the aircraft; higher speed means less chance of getting hit/shot which in turn means lower cost of repair. Thus, the product definition of an aircraft can be broadly classified into following categories. 1. Explicit product definition: (a) aircraft geometry; (b) power plant and systems data; (c) material type; (d) aircraft weights, etc. 2. Implicit product definition: (a) performance specifications (range, endurance, acceleration, turn radius, cruise and maximum speeds, manoeuvrability, etc.); Proc. IMechE Vol. 224 Part G: J. Aerospace Engineering

480

P Thokala, J Scanlan, and A Chipperfield

Fig. 2

Overview of the LCC framework

(b) signature data (e.g. visibility, radar cross-section etc); (c) failure modes, effects and criticality analysis (FMECA) data. Implicit product definition design parameters are dependent on explicit product definition and the dependencies can be modelled using physics-based models. The aerodynamic coefficients are calculated from the aircraft geometry by using aerodynamic analysis (full potential method) and a performance model has been developed to estimate the aircraft performance from the aerodynamic characteristics and explicit design parameters, making use of standard flight dynamics equations. Similarly, signature and FMECA data are estimated using simple signature analysis and reliability analysis models respectively. There is no need to specify the implicit product definition design parameters as inputs at the start of the LCC estimation process as they can be derived from explicit product definition of the aircraft. Thus, estimating the LCC of any given aircraft can be achieved by having only explicit product definition as input to the LCC framework. 5

MODEL ARCHITECTURE

This section gives an overview of a proposed LCC estimation framework. The life cycle cost framework is developed with emphasis on conceptual design of aircraft; it is to be noted that the aim is to compare different aircraft designs rather than to estimate the exact cost or return of the aircraft. This is being achieved by having explicit product definition as input as explained in section 4. The LCC of an aircraft Proc. IMechE Vol. 224 Part G: J. Aerospace Engineering

includes the material and the manufacturing costs along with the costs necessary for operation, maintenance and repair of a fleet of aircraft. The cost during planning and design phases is difficult to estimate and is out of the scope of the present work. A schematic flow sketch describing the LCC framework is shown in Fig. 2. From the aircraft geometry specifications, material type, etc. the raw material and manufacturing costs are estimated by the acquisition cost model using an activity-based costing approach [1]. The simulation model gives an estimate of the cost of maintenance, operation, and repair making use of the aircraft’s implicit product definition, mission details, and logistics data as inputs. These costs, when combined, give the whole life cycle cost of the aircraft. The acquisition cost model is developed using DecisionPro™ [24], while the operation and maintenance costs are estimated with a discrete event simulation model developed using Extend™ [25]. 6

ACQUISITION COST MODEL

The model shown here estimates the acquisition costs of an aircraft given its explicit product specification. It has the capability to estimate the costs of aircraft structures manufactured using metal-based as well as non-metal-based materials. The overview of the model is presented in the context of acquisition costs, concentrating on estimating the manufacturing and material costs. DecisionPro™ has a hierarchical structure and taking advantage of this characteristic, the acquisition cost model is organized in a hierarchical tree structure that reflects the actual physical structure of the JAERO574

Life cycle cost modelling as an aircraft design support tool

aircraft which allows easy and intuitive navigation. The cost of the structure is split into wing, fuselage, and empennage costs, which are further divided into different categories. The wing, fuselage, and empennage costs are the sum of the costs of different structural sets (i.e. spar set, rib set, etc., and the cost of each structural set is estimated by adding the cost of individual structural parts (i.e spars, ribs, etc.)). The cost of each part is estimated using activity-based costing approach, which calculates the raw material and manufacturing resources consumed for each part, making use of its dimensions. The dimensions of the individual structural parts are calculated from the explicit product definition of the aircraft (which includes high level dimensions such as wing span, sweep, chord length, fuselage length, width height, etc.) and the internal structural data (which includes structural spacing information such as stringer pitch, rib pitch, etc. and the specifications of the parts such as rib thickness, stringer type, etc.). The raw material and manufacturing costs for each part are estimated by using individual part dimensions along with the relevant process/material data from the knowledge base existing in the cost model. The costs of individual parts are then added to estimate their respective set costs

Fig. 3 JAERO574

481

and the costs of all the structural sets are then added to achieve the overall acquisition cost of the aircraft. The knowledge base in the cost model has formalized manufacturing knowledge so that the information can be reused is an easy manner. It contains libraries of processes and materials modelled as objects to enable a generic and hierarchical costing environment. The cost model makes use of different objects (called ‘components’ in DecisionPro) for estimating the material and manufacturing costs. The use of object-oriented approach makes the model consistent, easy to maintain, and permits reuse of components as well as making it easier for testing and validation. For example, it is easier to analyse a particular manufacturing process than the complete cost model. Also, it allows for controlled access (i.e. modification of individual components without disturbing the actual cost model). Moreover, finally, it results in a consistent model structure and standards for the cost model. An example ‘object’ estimating the cost of a skin set incorporated in the hierarchical structure of the model is shown in Fig. 3. The cost of the skin set is estimated by adding the cost of individual skin panels. The raw material and manufacturing costs of each panel are estimated by calling the corresponding objects or

Hierarchical structure of the acquisition cost model Proc. IMechE Vol. 224 Part G: J. Aerospace Engineering

482

P Thokala, J Scanlan, and A Chipperfield

components. For example, the manufacturing cost of a metal-based skin panel is estimated by calling the forming, machining, and friction-stir welding objects that reside in the process libraries. The object that estimates the friction-stir welding cost is shown in Fig. 4. The cost of friction-stir welding is estimated by using the length of the part welded, welding rate, and the hourly cost. Similarly, the machining cost is estimated using the amount of metal removed from the panel while the forming cost is estimated from the dimensions and curvature of the skin panel. The manufacturing cost of each skin panel is estimated by adding its

Fig. 4

corresponding machining, welding, and forming costs and the cost of raw material for each skin panel is estimated from its material type and its dimensions. The cost of the skin set is calculated by adding the raw material and manufacturing costs of the individual skin panels. Similarly, others structural set costs can be estimated by making use of the relevant objects for different manufacturing processes and all the structural sets in the cost model are then combined to form ‘wing cost’, ‘fuselage cost’, and ‘empennage cost’. The sensitivity of wing cost against its geometry parameters is shown in Fig. 5. The figure shows how

Component estimating the welding cost

Fig. 5 Wing cost sensitivity analysis Proc. IMechE Vol. 224 Part G: J. Aerospace Engineering

JAERO574

Life cycle cost modelling as an aircraft design support tool

the cost changes as each of the geometry parameters are varied from their initial value by −30% to +30%. It is apparent that the span of the wing is the main cost driver closely followed by the root chord. Also, if any dimension increases by a structural part pitch (i.e. rib pitch, stringer pitch, etc. a new part needs to be added to the internal structure that results in a steep increase of the cost). This phenomenon can be observed in the tip chord plot, where there is a sharp increase at around 15% variation. Sensitivity analysis is useful in identifying the important design parameters as the computational expense for optimization increases exponentially with the number of design parameters.

7

SIMULATION MODEL

The structure of the simulation model that is developed to estimate the operating and maintenance costs for a fleet of aircraft is shown in Fig. 6. The model detailed here is a discrete-event simulation model which is capable of estimating the operation and maintenance costs of a fleet of aircraft, taking into account the mission characteristics, aircraft performance, and the logistics data. The simulation model is also equipped with survivability and FMECA analysis. The first step in the model is the preparation of a flying schedule covering each aircraft over the time span to be simulated. The rate at which missions are called, the numbers of aircraft required, and the mission lengths can be generated from statistical distributions, or as a pre-determined schedule. Aircraft are drawn from a ready pool, inspected, and launched on the missions. The preparation phase for the aircraft includes the loading of applicable equipment, ground crew pre-flight inspection, and accomplishment of those maintenance tasks discovered in the preparation

Fig. 6 JAERO574

483

routine. Aircraft are then launched and flown on their specified missions, when all of the required aircraft have successfully completed the pre-flight check. In the course of the mission, system failures are experienced and the aircraft receive combat damage. The systems can fail due to battle damage, unreliability, or both. Survivability and reliability analysis are performed to determine the systems that experienced battle damage or reliability failure respectively. The mission outcome then depends upon the ability of the aircraft to withstand both the damage mechanisms and system failures. The capability of an aircraft to avoid or withstand hostile environments is known aircraft survivability and the ease with which an aircraft is killed in a hostile environment is measured by the probability that the aircraft is killed, PK . The inability of an aircraft to avoid the hostile mission environment (e.g. guns, approaching missiles, exploding warheads, etc.) is measured by PH , the probability the aircraft is hit by a damage causing mechanism, and is referred to as the susceptibility of the aircraft. The inability of an aircraft to withstand the damage caused by the hostile environment is referred to as the vulnerability of the aircraft. Vulnerability can be measured by the conditional probability the aircraft is killed given that it is hit, PK|H . The probability of kill of the aircraft is given by the joint probability the aircraft is hit and it is killed given the hit (i.e. the product of the probability of hit (the susceptibility) PH and the conditional probability of kill given a hit (the vulnerability) PK|H ). Thus Probability of kill = Susceptibility × Vulnerability or PK = PH PK|H

Overview of the simulation model Proc. IMechE Vol. 224 Part G: J. Aerospace Engineering

484

P Thokala, J Scanlan, and A Chipperfield

Similarly, the probability of a specific aircraft system kill given a hit on the aircraft is known as system kill probability (Pk|Hi ). It is the product of the probability that the system is hit (given the hit on the aircraft) Ph|Hi and the probability the system is killed given a hit on the system Pk|hi . During the mission simulation the capability of the aircraft or the systems to survive the hostile environment is measured by these probabilities. These probabilities are dependent upon the aircraft performance, survivability equipment, and weapons carried by the aircraft, tactics implemented during the mission, and the threat scenario. The susceptibility and vulnerability probabilities are assessed for a given aircraft in the mission-threat scenario to determine the probability of survival of the aircraft in that selected scenario. The probability of battle damage for a baseline aircraft during different missions is estimated using a lethality prediction tool, AGILE™ [26] and the battle damage rate for other aircraft is then obtained by comparing the aircraft’s performance against that of the baseline aircraft’s performance. The battle damage probabilities for the individual aircraft systems are modelled using discrete probability distributions, and are classified into critical, major, and minor according to the level of damage. Thus, aircraft combat survivability techniques are used in the simulation model to assess the battle damage. Similarly, reliability of an aircraft or aircraft system is also estimated using probability distributions. The time to failure due to unreliability is modelled as a Weibull distribution and the probability measures are used to predict the systems that fail during the course of a mission. The mission outcome is dependent upon the ability of the aircraft to withstand battle damage and system unreliability. If the failure or damage is non-aborting, the aircraft continues on the mission. However, if the failure or damage is mission aborting, the aircraft returns to base. Similarly, if the combination of systems that suffered critical damage corresponds to the classified critical systems according to the critical component analysis, then that particular aircraft is considered destroyed and is subtracted from the total available aircraft. All the events occurred during the mission are logged. When missions are completed, aircraft are recovered and serviced. A post flight inspection of the aircraft is performed and required scheduled and unscheduled maintenance tasks are performed to return the aircraft to a ready status. The exact level of maintenance needed is directly linked to the survivability analysis. The systems with no damage would be evaluated as no cause for repair. If a system is found critically damaged, then that particular system is considered beyond repair and it has to be replaced. Similarly, if the damage is major, high level of repair is necessary and systems that suffered minor damage need just a simple repair. In the event of repair, personnel, equipment, spares, Proc. IMechE Vol. 224 Part G: J. Aerospace Engineering

and the time required to accomplish the tasks are estimated from the model logistics data. The maintenance and repair costs are estimated using the resources expended for the repair and these resources include the labour costs, spares cost, and the cost of new systems, in the case of system replacements.

8

LCC FRAMEWORK

A framework capable of calculating the whole life cycle cost has been developed. A three-dimensional geometry of the aircraft is built using its product definition, which acts as a sanity check to make sure the aircraft design is realistic and one that is intended by the user before proceeding to the phase of cost estimation. This parametric geometry is based on conventional aircraft configuration comprising of wing, fuselage, and an empennage and is used as input by the acquisition cost model to estimate the cost of aircraft. Although this parametric representation does not consider rotorcrafts, biplanes, or other unusual configurations, it is still flexible enough to represent conventional, canard, and blended body wing (BWB) configurations. The acquisition cost model then uses the product definition and details of the internal structure to infer a manufacturing sequence and to estimate the process times and material costs. It should be noted that an extremely sparse or dense internal structure can lead to structurally unsound aircraft. Predefined manufacturing process sequences to manufacture the key structural parts (which are limited to those mentioned in the internal structural representation) are specified, using different process sequences for metalbased parts and composite parts. It should be noted that this approach is not conducive to estimate the cost of radically new designs and cost comparison between different manufacturing approaches. From the explicit product definition (geometry, material type, power plant, and systems data), aircraft performance is derived using physics-based models. This is again given a sanity check to make sure the aircraft is not physically implausible. The discreteevent simulation model estimates the operation and maintenance costs for a fleet of aircraft using mission details, aircraft performance, and product definition as inputs. The air base is assumed to have two types of aircraft: combat aircraft with low aspect ratio for the closer to ground military missions and high altitude long range aircraft for the reconnaissance missions. Also, It is assumed that there is a limited range of mission and threat descriptions that can be performed. These are not unrealistic assumptions, given that each air base has a limited set of aircraft and a limited number of generic mission activities. Aircraft are launched on the missions according to sortie demand and the systems that fail due to battle damage and/or unreliability are determined by performing survivability JAERO574

Life cycle cost modelling as an aircraft design support tool

and reliability analysis, respectively. The simulation model estimates the fuel, repair, and maintenance cost for each aircraft after every mission using the logistics data and these costs are then added to estimate the operation costs for the fleet of aircraft. The maintenance and operation costs combined with the acquisition costs provide the life cycle cost of the aircraft.

9

CASE STUDY

A life cycle cost comparison between a fleet of metalbased UAVs and a same-sized fleet of non-metal-based UAVs is performed. The UAV configuration chosen for this study is a surveillance/reconnaissance aircraft and it has a wing span of 15 m, length of 9 m, and height of 2 m. A fleet size of 10 is chosen, but the fleet size can also be varied to examine whether metal-based or non-metal-based achieved better LCC per UAV. Both the UAVs have identical geometry and the propulsion system is also assumed to be same; the study here is to identify the better material choice over the whole life cycle between metal-based and composite UAVs with the same explicit product definition (except for the UAV structural material). The acquisition costs are estimated first for both the vehicles. This is performed by calculating the dimensions of the individual parts of the UAV from the high level dimensions such as wing span, sweep, chord length, fuselage length, width, and height etc. The structure of the UAV considered here is fairly simple and generic. It has a conventional configuration with fuselage, wing, and tail. The wing consists of stringers, spars, ribs, and an outer skin. The number of stringers, spars, or ribs is variable and can be modified and the cost model structure in itself is independent

Fig. 7 JAERO574

485

from this variation. The fuselage is assumed to have a semi-monocoque structure with frames, stringers, and skin. Again, the number of frames and stringers can be varied. In this case study, the wing has three spars, 15 stringers, and 10 ribs while the fuselage has eight frames and 15 stringers. The dimensions of these individual parts are then calculated from the high level geometry. The raw materials and the manufacturing costs for each part are estimated from their dimensions. The costs of all the structural sets are then added to achieve the overall acquisition cost of the UAV. It was observed that the structural cost of the non-metalbased UAV is higher than that of a metal-based UAV. This can be attributed to the high raw material and manufacturing costs of composite materials. It should be noted that the tooling and assembly costs are not included in this study. Implicit product definition calculated from the explicit product definition is different for the UAVs and this in turn leads to different operational and maintenance costs as estimated using the simulation model. The UAVs are assigned missions according to a sortie file which is input into the simulation model. Aircraft may experience system failures and combat damage during the course of the mission. The battle damage rate is different for metal-based and non-metal-based UAVs. This difference is attributed to the fact that composite UAV performs better than the metal-based UAV. Both UAVs have the same aerodynamic properties and the same propulsion system, but the composite UAV is lighter; hence, it is more agile and faster. Also, composite UAV are stealthier compared to metal-based UAV. Thus, the battle damage probability is lower for composite aircraft as estimated using survivability analysis. The fuel burn rate is also lower for the composite aircraft due to its lower weight and this difference is estimated using Breguet’s equation. All the

LCC and the LCC difference plotted against time Proc. IMechE Vol. 224 Part G: J. Aerospace Engineering

486

P Thokala, J Scanlan, and A Chipperfield

Fig. 8

Cost estimation website

other parameters in the simulation model are assumed to be same for both vehicles. The logistics data are also the same for both the vehicles (i.e. the repair times, repair costs, spares, and personnel required are assumed to be same). These assumptions are not strictly true, but this can be refined when data become available. It was observed that the operational costs for the fleet of metal-based UAVs are higher than that of the fleet with non-metal-based UAVs. This is because of the high repair costs and higher fuel consumption of the metal-based UAVs. The simulation model is then run for different time periods, starting from 1 year to 30 years. The LCC for the both fleets is estimated and plotted as a cost versus time graph as shown in Fig. 7. The LCC difference between metal-based and nonmetal-based fleets of UAVs is also plotted against time. It can be observed that in the first few years, the fleet with non-metal-based UAVs cost more than the fleet with metal-based UAVs, but in the long run the overall life cycle cost for the fleet of non-metal-based UAVs is lower than that of the metal-based fleet. This is due to the lower operational costs of non-metal-based UAVs. Proc. IMechE Vol. 224 Part G: J. Aerospace Engineering

10

CONCLUSIONS

A life cycle cost model that can be integrated into a multi-disciplinary design framework has been developed. The acquisition cost model developed using DecisionPro™ has the capability to estimate the product acquisition costs of an UAV from its design specifications. A discrete-event simulation model for a fleet of aircraft is developed using Extend™ to estimate the repair and maintenance costs. The models have been published on the local internet network for remote access; Fig. 8 shows the website which allows the user to verify the UAV geometry before estimating its acquisition costs. In future, it is planned to deploy these cost models on a secure web server for public access.

11

FUTURE WORK

The LCC framework developed needs to be integrated into the design process to facilitate the comparison between different configurations. This can be achieved by automating the framework, which will allow JAERO574

Life cycle cost modelling as an aircraft design support tool

trade-off studies and optimization to be performed without human intervention. Also, prior to cost estimation, structural analysis needs to be performed making use of the data from aerodynamic analysis to avoid structurally unsound aircraft. Value-driven design is an emerging topic in the engineering community which makes use of a mathematical value model in a formal optimization framework to balance performance, cost, schedule, and other measures to identify the best possible outcome [27, 28]. It is planned to use this value-driven methodology in contrast to cost-centric methodology to improve the conceptual aircraft design process. ACKNOWLEDGEMENTS The authors would like to thank BAE Systems for supporting this research work and allowing this information to be released for publication. © Authors 2010

13

14

15

16

17

18

REFERENCES 19 1 Curran, R., Ragunathan, S., and Price, M. Review of aerospace engineering cost modelling: the genetic causal approach. Prog. Aerosp. Sci., 2004, 40, 487–534. 2 Asiedu,Y. and Gu, P. Product life cycle cost analysis: state of the art review. Int. J. Prod. Res., 1998, 36(4), 883–908. 3 Shishko, R. and Smith, J. L. Lessons learned from the space station program: the role of life-cycle cost. Beyond the ISS: The Future of Human Spaceflight, NASA technical report, 2001. 4 Scanlan, J., Hill, T., Marsh, R., Bru, C., Dunkley, M., and Cleevely, P. Cost modelling for aircraft design optimization. J. Eng. Des., 2002, 13(3), 261–269. 5 Layer, A., Brinke, E. T., Houten, F. H., Kals, H., and Haasis, S. Recent and future trends in cost estimation. Int. J. Comput. Integr. Manuf., 2002, 15(6), 499–510. 6 Ben-Arieh, D. Cost estimation system for machined parts. Int. J. Prod. Res., 2000, 38(17), 4481–4494. 7 Stockton, D. J., Forster, R., and Messner, B. Developing time estimating models for advanced composite manufacturing processes. Aircraft Eng. Aerosp. Technol., 1998, 70(6), 445–450. 8 Rehman, S. and Guenov, M. A methodology for modelling manufacturing costs at conceptual design. Comput. Indl Eng., 1998, 35(3–4), 623–626. 9 Wei, Y. and Egbelu, J. P. A framework for estimating manufacturing cost from geometric design data. Int. J. Comput. Integr. Manuf., 2000, 13(1), 50–63. 10 Kulkarni, U. A. and Bao, H. P. Close-loop cost equation for objects manufactured by milling. J. Manuf. Sci. Eng., 2003, 125, 436–448. 11 Dean, E. B. and Unal, R. Design for cost and quality: the robust approach. J. Parametrics, 1991, 11(1), 73–93. 12 Bao, H. P. and Samareh, J. A. Affordable design: a methodology to implement process-based manufacturing cost models into the traditional performancefocused multidisciplinary design. In Proceedings of the JAERO574

20

21

22

23

24 25 26

27

28

487

8th AIAA/NASA/USAF/ISSMO Multidisciplinary Analysis and Optimisation Symposium, Long Beach, September 2000, AIAA paper 2000-4839. Gantois, K. and Morris, A. J. The multi-disciplinary design of a large-scale civil aircraft wing taking into account of manufacturing costs. Struct. Multidiscip. Optim., 2004, 28, 31–46. Marx, W. J., Mavris, D. N., and Schrage, D. P. Effects of alternative wing structural concepts on high speed civil transport airport life cycle costs. In Proceedings of the 37th AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics, and Materials Conference, Salt Lake City, Utah, 15-17 April 1996. Hackney, J. and Neufville, R. Life cycle model of alternative fuel vehicles: emissions, energy and cost trade-offs. Transp. Res. A, 2001, 35, 243–266. Bruening, G. B. The potential impact of utilizing advanced engine technology for a combat capable unmanned air vehicle (UAV). J. Eng. Gas Turbines Power, 2001, 123, 508–512. Andijani, A. and Duffuaa, S. Critical evaluation of simulation studies in maintenance systems. Prod. Plan. Control, 2002, 13(4), 336–341. Upadhya, K. S. and Srinivasan, N. K. System simulation for availability of weapon systems under various missions. Syst. Eng., 2005, 8(4), 309–322. Cook, T. N. and DiNicola, R. C. Modeling combat maintenance operations. In Proceedings of the Annual Reliability and Maintainability Symposium, San Francisco, California, January 1984. Cobb, R. Modeling aircraft repair turntime – simulation supports maintenance marketing efforts. J. Air Transp. Manage., 1995, 2(1), 25–32. Adamides, E. D., Stamboulis, Y. A., and Varelis, A. G. Model-based assessment of military aircraft engine maintenance systems. J. Oper. Res. Soc., 2004, 55(9), 957–967. Orsagh, R. F., Allen, E. C., Schoeller, M., and Roemer, M. J. A mission-based logistics and maintenance support cost simulation environment. In Proceedings of the IEEE Aerospace Conference Big Sky, Montana, 8-15 March 2003. Sandberg, M., Boart, P., and Larsson,T. Functional product life-cycle simulation model for cost estimation in conceptual design of jet engine components. Concurr. Eng., 2005, 13(4), 331–342. DecisionPro developer. Vanguard Software Corp, 1100 Crescent Green, Cary, NC, USA. Extend. Imagine That Inc., San Jose, California, USA. Patel, N., Chipperfield, A. J., and Keane, A. J. Missile endgame analysis via multiobjective optimization. In Proceedings of the 16th International Federation of Automatic Control (IFAC) World Congress, Prague, 2005. Collopy, P. Value-driven design and the global positioning system. AIAA paper 2006-7213, 2006 (American Institute of Aeronautics and Astronautics, Reston, Virginia). Castagne, S., Curran, R., and Collopy, P. Implementation of value-driven optimization for the design of aircraft fuselage panels. Int. J. Prod. Econ., 2009, 117, 381–388. Proc. IMechE Vol. 224 Part G: J. Aerospace Engineering

488

P Thokala, J Scanlan, and A Chipperfield

APPENDIX

Pk|hi

Notation

Pk|Hi

LCC Ph|Hi PH

life cycle cost probability of hit on ith system given a hit on the aircraft probability of hit on the aircraft

Proc. IMechE Vol. 224 Part G: J. Aerospace Engineering

PK PK/H UAV

probability of ith system kill given a hit on that system probability of ith system kill given a hit on the aircraft probability of aircraft kill probability of aircraft kill given a hit on the aircraft unmanned air vehicle

JAERO574