The commercial transport aircraft industry is currently developing new “more electric aircraft” (MEA) designs in which various conventional mechanical, hydraulic and pneumatic power systems are replaced with electrically-based power systems. Their objective is to improve the over
...
The commercial transport aircraft industry is currently developing new “more electric aircraft” (MEA) designs in which various conventional mechanical, hydraulic and pneumatic power systems are replaced with electrically-based power systems. Their objective is to improve the overall flight performance by reducing the aircraft weight and by a lower overall energy requirement for the systems. The vision for the future is to ultimately replace all systems with electrical systems and even to replace a part of the fuel used as primary source of energy for the propulsion system by an electrical power supply and thereby to achieve either a hybrid electric aircraft (HEA) or even all electric aircraft (AEA) if permitted by future developments in battery technology. In recent years, many small scale electric aircraft were developed to demonstrate the AEA concept. It has been determined that although the MEA, HEA and AEA concepts reduce the overall complexity of the aircraft, it significantly increases the complexity of electrical and electronic systems (E/E systems) and their integration into the aircraft, introducing a new challenge for the aircraft design industry. Two specific categories of aircraft, currently in operation, face the same challenge. These categories are; (1) unmanned aerial vehicles (UAVs), which by nature have more electrical and electronic systems (E/E systems) on-board and require an automatic flight control system due to the absence of a pilot and (2) aircraft which are inherently unstable and therefore require automatic flight control systems for stabilization. These two aircraft categories can be classified as typical mechatronic products. E/E systems have a significant impact on the overall flight performance, directly determine the flying qualities of aircraft, and are critical for safety. Thus, these systems should be developed synchronously with the other traditional engineering domains such as aerodynamics, structures and propulsion. However, several challenges need to be overcome before this can be achieved effectively. Three specific challenges are identified and addressed in the current research study: The development of high fidelity multiphysics simulation models for analysis and development of the E/E systems is a complex, time consuming and multidisciplinary task that requires a large amount of manual work from simulation experts; The design of consistent automatic flight control systems for use throughout the entire flight envelope and for all aircraft weight and c.g. combinations is labor intensive and requires the availability of high fidelity multiphysics simulation models in the early design phases; The development of control software components is prone to errors due to inconsistencies between the description of the top level physical configuration, the control architecture and the associated software components. Traditional aircraft design methods which are largely dominated by the mechanical engineering domains are not suitable to synchronously design complex integrated E/E systems. Moreover, the conventional design process, which is sequential to a large extent, cannot support concurrent engineering requirements. Therefore, novel methods and tools to support the development of the E/E systems on-board aircraft are needed. The overall objective of this research study is to reduce the development time of aircraft with a high level of integrated E/E systems by automating the design process of the flight control systems, by creating more consistent control software through the entire design envelope. Besides a reduction in development time this will also improve the quality of the final (mechatronic) product. The three challenges described above will be tackled in particular. The novel methods and tools are based on the knowledge based engineering (KBE) approach. The KBE approach is highly suitable because it cannot only automate non-creative, repetitive design tasks done for example by simulation experts but also support for multidisciplinary design, analysis and optimization (MDAO). Compared to other existing KBE systems, the proposed system integrates the flight control system design with the physical design in three specific areas. First, in order to ensure a consistent design representation, the concept of a multiphysics information model (MIM) is proposed in order to integrate the design knowledge present from multiple engineering domains. The proposed MIM (a KBE system) defines objects with attributes to represent various aspects of physical entities (e.g., mass, inertia, geometry, material properties). Moreover, it uses functions to capture non-physical information, such as the control architecture, relevant test maneuvers, simulation procedures, etc. The problem of system coupling and interactions between disciplines involved are taken into account by the proposed KBE system in a knowledge acquisition process. Next, depending on the requirements, the proposed KBE system extracts necessary knowledge from the MIM which is needed for the development of a multiphysics simulation model, which is composed of a physical plant, flight control systems including the embedded control software and simulation configurations. By capturing the expertise of simulation experts, the proposed KBE system is able to automatically instantiate the multiphysics simulation models. This multiphysics simulation model can be used to evaluate the flight control systems in operation practice throughout the flight envelope, for example when performing maneuvers. Altogether, the MIM enables rapid development of high fidelity multiphysics simulation models for analysis and development of the flight control systems. Second, in order to evaluate the inherent flying qualities of unstable aircraft in a simulation environment, an automatic flight control system is required. For this purpose, model based inversion control is applied. This method has the advantage that tuning is not required. The techniques, processes and knowledge required to develop a model based control system based on the (nonlinear) multiphysics simulation model are captured by the KBE system. Model based inversion control has its disadvantages when implemented on real-life aircraft. For the final design solution developed by the framework, which will enter the detailed design phase and which will ultimately be produced, other control methods and architectures can be developed, more appropriate for a real-life situation. Such a control system will only have to be tuned and developed once in contrast to the thousands of designs evaluated in an MDO framework. This application of model based inversion control is considered new. Third, in order to avoid errors in the embedded control software as a result of manual programming activities, the dependencies of parameters in the software on physical parameters of an evolving design and the high complexity (thousands of lines of code), control software components of flight control systems should ideally be developed in an automated fashion. The proposed KBE system has the ability to generate consistent control software components. The system extracts the variable definitions and values from physical configurations and control architecture from the information model to specify the variables in the software components. In addition, the system divides software components into basic elements and writes them into strings, which can, in principle, be any computer language. When the top level configuration and control architecture changes, the proposed KBE system can operate the basic elements in specific order and automatically create new software components by capturing the expertise from software engineers. Summarizing, because both the geometry model and multiphysics simulation model including flight control system are obtained from one source, the MIM links the physical modeling and control system design with the development of software components with respect to data and topology structure. A multirotor UAV configuration is used as test case to demonstrate the novel methods and tools described above. This is an inherently unstable configuration with a wide range of applications. A computational framework is developed which enables the conceptual/preliminary design and optimization of this typical mechatronic product. The proposed KBE system automatically creates thirty thousand designs of multirotor UAVs with different topologies and then evaluates each solution by automatically simulating five test maneuvers and by checking twenty-two constraints. Results show that the proposed KBE system can automatically generate multiphysics simulation models to support the multidisciplinary analyses not restricted to the mechanical domain but also applicable for evaluation of flight control systems and other domains. Even though different design solutions can have a highly different topology, automatic flight control systems based on the model inversion control method are created automatically for each design solution, enabling the evaluation of the inherent flying qualities of the unstable aircraft configuration. Furthermore, within the framework, design processes are automatically completed from the initial definition of top-level aircraft requirements, to the design and optimization, and finally down to selecting feasible solutions. The approach demonstrated leads to: a reduction in manual work, improved quality of the final solution, and consistent control system and software components. Key to the MIM concept is that it focuses on capturing the intrinsic properties of physical systems by the KBE approach and a specific format of representation is avoided. Although the current research study focuses on the software of the flight control systems in particular, the concept of the MIM can in principle be applied to design the complete E/E system, including hardware components, as well as other multiphysics systems.