Feature ArticleAutomated Conceptual Design Utility For Unmanned Underwater Vehicles
By Candace Brown
Richard P. Clark
Lockheed Martin Mission Systems and Sensors
Riviera Beach, Florida
Compared to the naval architecture and aerospace industries, the historical development of unmanned underwater vehicles (UUVs) has been more focused on low-cost implementation. As a result, designers often employ commercial-off-the-shelf products and more readily accept nonoptimal capability in final design. However, as UUV mission requirements grow in sophistication and complexity, the demand for near-optimal engineering design of UUVs is expanding, driving the use of subsystems and components with increasingly specialized function and exotic, expensive materials. With this expansion, the historical disparity in the level of design optimization between industries has become more apparent in the methods for conceptual, parametric and preliminary design.
Engineers at Lockheed Martin Mission Systems and Sensors have recently created a rapid, low-cost parametric UUV design tool, the Automated Vehicle Conceptual Design Utility (AVCDU).
The utility, which Lockheed uses for customers during UUV conceptual design, is ideally suited for exploration of multidimensional design spaces and performing conceptual design studies. The utility conducts studies at a level comparable to the tools that support aerospace engineering and naval architecture.
AVCDU runs within MathWorks Inc.'s MATLAB software. Users of the system provide input parameters through a single script, which allows the program to perform vehicle analysis. Output is then provided as data in a spreadsheet and in graphical form. A future implementation of the utility, which Lockheed plans to develop in 2011, will provide computer-aided design (CAD) output.
Background on UUV Design
A closed UUV design indicates that the vehicle satisfies all top level requirements (TLRs) using subsystems that individually satisfy their allocated system level requirements. The complexity of identifying a closed design results from the interdependencies of subsystem sizes. The size of almost every subsystem is influenced by the aggregated elements. For example, adding a higher-capacity battery might require a larger vehicle with increased resistance, which would require more energy to travel the same distance, thus negating the additional capacity.
In the past, closed UUV designs have been identified through manual analysis with a spiral design process. Vehicle elements are iterated in a 'spiral' by rotating through each subsystem, performing design modifications to satisfy requirements. Automation of the spiral design process represents an opportunity for both a dramatic reduction in the cost of engineering products and the identification of a superior-performing design variation.
Lockheed Martin's AVCDU requires two sets of input parameters to converge on a conceptual vehicle design: a mission description and a selection of vehicle configuration options. A vehicle mission description is divided into a user-prescribed number of discrete segments. Each segment is defined by its unique parameters, which include vehicle conditions (ground speed, range, depth, hotel power), payload conditions (payload power, weight, volume) and environmental conditions (current, sea state, seawater density). The vehicle mission parameters are typically obtained from vehicle TLRs and a concept of operations description.
The vehicle configuration options dictate the set of vehicle design options to be evaluated by the AVCDU. The most important of these options are the fineness ratio and block coefficient, which dictate the basic packaging and sizing of all other vehicle subsystems. Another basic option is the selection of vehicle appendages such as fins, specialized sensors or communication devices and payload appendages—these impact the vehicle drag estimation.
Basic AVCDU Operation
Leveraging the legacy spiral design process, the AVCDU begins with a presumed vehicle scale and evaluates the vehicle's total resistance using a Reynolds number defined from the mission segment information. The resistance defines the required propulsor and motor size. The energy storage system is sized to power the propulsor, hotel devices and payload over the specified vehicle range. Weight and volume is allocated to miscellaneous elements defined by the configuration selection options. The components are housed in pressure vessels sized to address the maximum depth in the mission segment information. Finally, sufficient flotation foam or ballast is allocated to set the total wet weight of the system to the user-desired value (typically zero). The overall vehicle size is computed using a packing factor and this computed size is compared against the initial size estimate. An error in the size estimate will dictate the next iteration.
AVCDU Code Validation
The code was validated by comparing AVCDU conclusions against several vehicle case studies. After the design utility generated estimates of the subsystem sizes and the aggregate vehicle size, the components were assembled in a CAD environment. Independent engineering estimates confirm that the resulting solution was indeed a closed vehicle design that minimally satisfied the performance requirements. A second form of validation used existing vehicle designs.
The AVCDU was run with the same mission information and configuration options as an existing vehicle and the utility sizing conclusions were compared with the true vehicle design. In four validation cases examined by Lockheed Martin, the average AVCDU error was five percent for aggregate metrics like length, diameter, volume, wet weight, dry weight and energy capacity.
Design Study of a Long-Range UUV
The ability to change any input values and rapidly reach sizing convergence makes the AVCDU ideal for studying vehicle sizing trends over a range of input parameters. In the case of studies where requirements are incomplete or not clearly defined (typical in a concept study or preproposal setting), the AVCDU can be used to study the impact of those requirements and identify design drivers. Such is the case for a long-range UUV (LRUUV) design study conducted at Lockheed Martin.
The LRUUV must be capable of transiting between 20 and 300 nautical miles while carrying a large-volume, variable payload of unspecified weight. The LRUUV is designed to pick up, carry and drop off a large volume, large wet-weight payload. The scale of the payload is a significant driver of the vehicle size, making the integration of the payload with the vehicle an important design element. A variable ballast system is necessary to maintain neutral buoyancy when the payload is picked up and dropped off.
The input to the code requires a mission description and a vehicle configuration: The LRUUV mission description is divided into five segments—launch, transit, work, return transit and recovery—each defined by key mission parameters like range and speed and environmental parameters such as current and salinity. The vehicle configuration description is parameterized with fineness ratio values of three, five and seven, indicating the ratio of length to diameter. Other configuration options are varied to account for the differences between the three vehicle payload configurations.
The large variations in mission distance and ground speed result in some nonpragmatic scenarios. For instance, executing a 10-nautical-mile mission at 12 knots completes the mission in less than two hours. Alternatively, completion of a 150-nautical-mile mission at two knots requires more than six days. Therefore, the study was conducted with a fixed total transit time of 24 hours, creating a one-to-one correspondence between ground speed and range.
A nominal hotel power load was selected for the vehicle based upon estimation of its required components, such as a Doppler velocimeter, an inertial navigation unit, a forward-looking sonar (to support collision avoidance) and a navigation computer. However, due to the large payload and long-range missions, the hotel power load is considered much less significant to the vehicle scale than propulsion power.
Three methods of payload integration were considered in the design study, each with varying impact on overall vehicle scale:
Internal Payload Vehicle. The UUV is designed with an open space within its envelope that is the size and shape of the payload.
Expandable Payload Vehicle. The UUV is designed to provide no payload space while stowed, but once deployed, expands in length to allow space for the payload within its envelope.
External Payload Vehicle. The UUV carries the payload beneath the hull rather than within the vehicle. Because carrying the payload externally during transit increases vehicle drag, the variable ballast system is comprised of air ballast bags that, when inflated, provide fairings for the payload.
LRUUV Conceptual Design Results
Results illustrate the vehicle length, which for most applications is the largest driver of the handling, transportation and platform stowage requirements. The vehicle diameter can be derived from the stated aspect ratio of each design point.
The length output of the AVCDU for each of the three vehicle configurations at three aspect ratios was investigated. Note that the reported lengths for the expandable payload vehicle are for its stowed configuration (not containing the payload).
As expected, the higher-aspect-ratio vehicles for each configuration are longest. The expandable payload vehicle is shorter than the internal payload vehicle, and the most dramatic length differences occur at short transit distances (and low speeds).
At long transit distances, the vehicle size is dominated by energy storage, and the vehicle length reduction incurred by collapsing the payload volume decreases, eventually to the point that there is no difference between the two internal payload vehicle sizes. For the external payload configuration for long transit distances (at high speed), the additional drag incurred on the vehicle from the payload dominates the vehicle size.
To validate the results of the LRUUV analysis, three design points were selected for further definition in a CAD model. A design point was selected for each vehicle configuration at a different aspect ratio; specifically, the internal payload vehicle had an aspect ratio of seven, the expandable internal payload vehicle had an aspect ratio of five and the external payload vehicle had an aspect ratio of three.
Subsystems were modeled based on the sizes defined by the AVCDU output and then arranged within a hull shape sized according to the overall length and diameter predicted by the AVCDU. The validation exercise showed that the AVCDU is capable of producing a converged design with accurate system sizes and an appropriate margin that can be arranged into a realistic vehicle configuration.
The LRUUV parametric analysis provides valuable insight into the key drivers for a long-range vehicle design. For long-range vehicles at high speed, energy stowage dictates the vehicle size. For these missions, the design must focus on drag reduction and increased energy density to minimize vehicle size. For short missions at low speed, depth and variable ballast requirements have more impact on vehicle size. The AVCDU identified a transit range of 70 to 90 nautical miles as a critical transition between these drivers for the given payload scale and a 24-hour total transit time.
The ability of the AVCDU to reach a fully converged vehicle size in a matter of seconds allowed numerous design parameters to be explored in the LRUUV study. More than 1,000 closed vehicle design points were defined with the AVCDU; using traditional UUV concept analysis methods, only a few closed design points would have been defined in that same amount of time. The speed, flexibility and automation of the AVCDU enables engineers to more easily and consistently explore the impact of new technologies, along with changing requirements and trades in subsystem design early in the process of an overall vehicle design. The LRUUV analysis shows that Lockheed Martin is now able to address UUV conceptual design studies with the same fidelity and sophistication that is applied in naval and aeronautic conceptual design studies.
Candace Brown has more than six years of experience as a systems analyst at Lockheed Martin Mission Systems and Sensors in Riviera Beach, Florida. Since completing undergraduate studies in naval architecture at the Webb Institute, she has executed systems engineering and analysis for diverse undersea products. Brown earned a master's degree in ocean engineering from Florida Atlantic University in 2009.
Richard P. Clark is a systems analyst at Lockheed Martin Mission Systems and Sensors. His recent work focuses on conceptual unmanned underwater vehicle design, with particular emphasis on hydrodynamics, propulsion and thermal management. He earned master's and undergraduate degrees in mechanical and aerospace engineering from Princeton University.