شبیه سازی غیر خطی زیردریایی

شبیه سازی غیر خطی زیردریایی

اسلاید 1: زمستان 1382Dr. H. BolandiExample: Consider the following equations of motion for a submarine in the dive plane

اسلاید 2: زمستان 1382Dr. H. BolandiWhere :

اسلاید 3: Now say we want to linearize these equations for a level flight path when the dive plane angle is zero, δ=0 . Then by setting all time derivatives to zero ( this corresponds to equilibrium ) we get : زمستان 1382Dr. H. Bolandi

اسلاید 4: زمستان 1382Dr. H. BolandiIf we choose to linearize around the equilibrium we have

اسلاید 5: زمستان 1382Dr. H. BolandiOne thing we have to emphasize is that in the submarine examples in these notes U is the forward speed (not control). The control is designated by δ; this is standard notation.

اسلاید 6: زمستان 1382Dr. H. Bolandi

اسلاید 7: زمستان 1382Dr. H. BolandiSTOP

اسلاید 8: زمستان 1382Dr. H. Bolandi3)

اسلاید 9: زمستان 1382Dr. H. Bolandiروشهای محاسبه ماتريس انتقال حالت 1) روش سريها 2) روش تبديل لاپلاس 3) روش قطري سازي 5) روش سيلوستر4) روش كيلي ـ هاميلتون

اسلاید 10: محاسبة ماتريس انتقال حالت با استفاده از روش قطري‌سازيزمستان 1382Dr. H. Bolandiحالت اول: ماتريس A داراي قطبهاي مجزا و غيرتكراري است. لذا قطري‌سازي آن امكان پذير است.D ماتريس قطري شده است كه به فرم زير مي‌باشد:

اسلاید 11: مثال : زمستان 1382Dr. H. Bolandi

اسلاید 12: زمستان 1382Dr. H. Bolandiحالت دوم : مقادير ويژه A تكراري باشند و بتوان A را به فرم جردن تبديل نمود.ابتدا ماتريس S را بنحوي محاسبه مي‌كنيم كه ماتريس A را بتوان بفرم جردن تبديل نمود: J ماتريسي به فرم كانونيكي جردن است.

اسلاید 13: زمستان 1382Dr. H. Bolandiبا توجه به تعريف تبديل همانندي فوق: مثال :

اسلاید 14: زمستان 1382Dr. H. Bolandi مثال :

اسلاید 15: زمستان 1382Dr. H. Bolandiقطری سازی با استفاده ازروش ماتريسهاي همبسته

اسلاید 16: زمستان 1382Dr. H. Bolandi

اسلاید 17: زمستان 1382Dr. H. Bolandiمثال : ماتريس A را قطري كنيد.

اسلاید 18: زمستان 1382Dr. H. Bolandiروش كيلي هاميلتون

اسلاید 19: زمستان 1382Dr. H. Bolandi nريشه دارد

اسلاید 20: زمستان 1382Dr. H. Bolandiاين معادله دلالت بر قضيه كيلي هاميلتون دارد كه هر ماتريس مربع مانند A معادله مشخصه خود را برآورده مي‌سازد لذا چند جمله‌اي متناظر با معادلات (4) و (5) بصورت زير هستند كه :طريقه عمل :

اسلاید 21: زمستان 1382Dr. H. Bolandiمثال : ماتريس انتقال حالت ماتريس A را محاسبه كنيد.

اسلاید 22: زمستان 1382Dr. H. Bolandiدر صورتيكه مقادير ويژه A تكراري باشند از روش بالا ديگر نمي‌توان استفاده نمود.

اسلاید 23: زمستان 1382Dr. H. Bolandiمثال :

اسلاید 24: زمستان 1382Dr. H. Bolandiمثال :روش كيلي ـ هاميلتون

اسلاید 25: زمستان 1382Dr. H. Bolandi مثال :حل :

اسلاید 26: زمستان 1382Dr. H. Bolandiمثال :

اسلاید 27: زمستان 1382Dr. H. Bolandi1. In this particular simulation we used a time step ∆t = 0.01 seconds. Is this small enough? The easiest way to check this is to reduce (or increase) ∆t, say by a factor of 10, and re–run the program. If the results do not change, the above choice for ∆t was good. A more rational way to do the same thing would be to look at the natural time constant of the dynamics of the system. The system poles were found in page 18. It seems that the fastest pole of the system has real part −0.5159, and the time constant that corresponds to this is about 1/0.5 or 2 seconds. This means that it takes a couple of seconds for the boat to “listen” to its dive planes, so ∆t = 0.01 should give very accurate results. In fact in this case we could go as far as ∆t = 0.5 and we would still be reasonably accurate.2. Look again at the system eigenvalues: one of them is certainly dominant, −0.0297, so the response should approximate that of a first order system with a time constant 1/0.0297, or about 33.5 seconds. Now look at the response of the figure: does it take approximately 33.5 seconds to go up to 60% of its final value?3. By now we are convinced that the transient response we see in the figure agrees with our engineering intuition. How about the final or steady state value of the response? This is something we can compute exactly. At steady state we should have, so that our equations become at steady state:Typical results of the simulation in terms of the pitch angle θ are shown in Figure 8. As with any numerical results, however, the real question is: are they correct? The answer to this borders between art and science, and in the context of system simulations here is a set of a few checks:

اسلاید 28: so that our equations become at steady state: زمستان 1382Dr. H. Bolandi

اسلاید 29: زمستان 1382Dr. H. BolandiSimulation of a nonlinear set of equations proceeds in a similar manner. Let’s assume that the only important nonlinearities in our example come from the trigonometric functions and not the hydrodynamic forces and moments; in other words the nonlinear equations of motion areThe numerical integration proceeds in exactly the same way as before; the only difference is that here the values for and are computed from the new equations. Typical results are shown in the previous figure where the difference between linear and nonlinear simulations is also shown. Naturally, whenever possible, simulations must be performed for the nonlinear systems since these model the underlying physics more accurately. The steady state value for θ can be computed from the nonlinear equations in the same way as before, the algebra is easy in the example case but keep in mind that for general nonlinear equations it may be very difficult. Here we can find and

اسلاید 30:

اسلاید 31: تبديل همانندي در معادلات حالت و خروجيزمستان 1382Dr. H. Bolandi