Contents
- 1 Modeling Methods
This section describes the procedures of developing the aircraft dynamics model on a one-third scale Yak-54. The aircraft model is later implemented into a simulation platform to perform either software-in-the-loop (SIL) simulations or hardware-in-the-loop (HIL) simulations supported by the autopilot systems. This process is especially crucial for the Piccolo II autopilot as users are required to perform gain tuning in its HIL simulations environment to acquire its preliminary stability and performance before the first autonomous flight. These gain values are to be tested and re-tuned, if needed, in an iterative process to refine its performance during flight tests (see above figure for Piccolo gain tuning procedures). Acquiring a good set of flight control gain values requires an accurate aircraft dynamic model. This can not only minimize the number of flights required to re-tune the system but also significantly reduces the risks, costs, and lead time for the flight test program.
This process, however, does not apply to the wePilot 2000 system as it does not require users to perform any gain tuning activities. The wePilot utilizes the robust nonlinear control technique which employs a complex mathematical process to calculate its optimal gains. Moreover, access is not opened for users to adjust any gain values on the wePilot system. The set of gain values for wePilot is tuned by the supplier using either system identification technique through flight tests or parametric method with required derivatives data given from the dynamics model.
The modeling techniques discussed here are primarily developed and used for the Piccolo II autopilot system in the effort for later simulations and gain tuning purpose.
Modeling Methods
Three different modeling techniques are used and compared in this study. An introduction for each method is briefly given below:The Cloud Cap Technology (CCT) initially provided a modeling tool that allows users to rapidly develop a preliminary aircraft model making use of basic geometric data and look-up tables to define the aerodynamics parameters, such as CL, CD, CM, for each lifting surface. The SCCS implements those user inputted data directly to its already integrated HIL simulation platform which supports pilot-in-the-loop command using Futaba console. That is the biggest advantage of this modeling method. Its drawback, however, is the undocumented process used to implement the data which posts questions about its precision. In addition, no preliminary modeling data are available to perform analysis, and thus the only way to verify its accuracy is to perform its HIL simulation, which makes it difficult to troubleshoot.
The CCT later released another modeling tool for its Piccolo system. This modeling tool makes use of the AVL code written by MIT. This method provides full access for users to well define the geometry of the vehicle, which makes it a useful tool for modeling an unconventional aircraft configuration. It utilizes the vortex lattice method to generate the stability and control derivatives. An interface written by CCT is used to import the AVL data into the Piccolo simulator that makes the HIL simulations possible using the AVL method.
AAA is a professional aircraft design software program designed by DARcorporation. The configuration of the vehicle is first defined by users. The AAA software will then perform interpolation and extrapolation from its comprehensive built-in database, making use of DATCOM to estimate the derivatives for various trim conditions. The list of stability and control derivatives outputted from AAA allows users to perform a complete dynamics mode analysis of a vehicle using conventional state space model technique [1]. This cannot be done by the SCCS method and can only be done on AVL method through approximation technique due to its incomplete derivatives data. The major disadvantage of the AAA method is its incapability to perform HIL simulations, as the existing Piccolo HIL simulation platform can only import data generated from SCCS or AVL models.
Stability and Control Derivatives
Two sets of derivatives are now available from AAA and AVL for a one-third Yak-54 RC model. Note that the AVL method does not support engine models into its modeling process. Therefore, not all thrust related derivatives are available from AVL. The AVL model also does not provide the derivatives related to drag and the rate of change of the angle of attack. The table below shows the comparison of the derivatives data given from these two methods.
The only way to truly evaluate the accuracy of each modeling technique is in flight tests. System identification flight tests are performed to excite the true dynamics of each mode through a series of singlet and doublet control inputs [2]. The data reduction method [2] is then applied to analyze the dynamics of each mode from the time domain flight test data. The analyzed results are then compared with those from the modeling results.
The above analysis technique cannot be used on the SCCS method as derivatives are not available from the SCCS model and thus cannot proceed any state space model analysis. Since the HIL simulation is available for SCCS, the identical system identification technique as used on the flight tests is also applied on SCCS to obtain its mode dynamics through simulation activities. This exercise is also repeated on the AVL model. For the AAA model, only the analytical method is applied as it currently does not support HIL simulations. The complete comparison process is illustrated in a flow chart as shown in the figure above.
Comparison Results
Four sets of results from HIL simulations and analytical methods are now available and are compared with the flight test data. The summaries of comparisons are listed in the table below. Note that the time constant of the roll mode dynamics is very small (fast) for Yak-54, making it highly affected by the elapsed time of the aileron input during flight tests and HIL simulations. The best way to truly compare the roll dynamics responses is to make a side-by-side comparison using identical aileron inputs. This can only be done on MATLAB/Simulink, and only the AAA and AVL state space models can be implemented into MATLAB/Simulink for simulation. The side-by-side comparison results for the roll dynamics is also shown in the figure below. More comparison results and discussions are available in Ref [3].
Conclusions
From the comparison study on the three different modeling methods, the following summary can be made:
- Among the three modeling methods, the AAA model provides the closest predictions to the flight test results on the Dutch roll, short period and Phugoid mode dynamics.
- The lack of an engine model on the AVL method makes its approximation results less accurate for short period and Phugoid mode estimates.
- The missing longitudinal derivatives from AVL make it less useful for further longitudinal simulation activities.
- The side-by-side comparison technique is introduced to truly compare the roll dynamics. This technique provides great benefits for validation processes, but this can only be done on the MATLAB/Simulink platform.
- According to the side-by-side comparison, the roll dynamic given by the AVL model is more precise than that given by the AAA model.
- Based on the final comparison results, the longitudinal derivatives given from the AAA model and the lateral derivatives provided by the AVL model are chosen to use for futher 6DOF simulation model development as discussed in the section of Simulations.
[1]: Roskam, Jan. 'Airplane Flight Dynamics And Automatic Flight Controls Part I,' DAR Corporation, Lawrence, Kansas, 2003.
[2]: Donald T. Ward and Thomas W. Strganac, 'Introduction to Flight Test Engineering,' 2nd edition, Kendall/Hunt, Dubuque, Iowa, 2001.
[3]: Leong, H.I. (Edmond), Jager, R., Keshmiri, S., and Colgren, R., 'Development of a Pilot Training Platform for UAVs Using a 6DOF Model with Flight Test Validation,' AIAA Modeling and Simulation Technical Conference and Exhibit, Honolulu, Hawaii, August 18-21, 2008. (PDF)
May 30, 2019 Despite the fact that Apple keeps the Mac Library folder hidden from users, it’s good to learn how to show the Library folder on Mac. You may need to access the Library folder when you need to clear cache, remove the app’s leftovers, edit preference files manually, etc. Access Library Folder On Mac. The Library Folder on your Mac contains Preference Files, Caches, and Application Support Data. While regular Mac users may not find the need to access the Hidden Library Folder, advanced Mac users will at times come across the need to access files inside the Library Folder for troubleshooting and other purposes. Ever since Lion (OS X 10.7), Apple has hidden your personal Library folder (/Library) by default. In Lion and Mountain Lion (OS X 10.8), you could make the folder visible, but it required some work. Jul 22, 2011 Fortunately, this is easy to reverse and if you wish you can switch the behavior and instead show Library all the time with a quick terminal command. This tutorial will detail four different ways to access the User Library folder in Mac OS X from Mt Lion and beyond.