كنترل بهينه ساختمان‌هاي مجهز به ميراگر MR با استفاده از شبكه عصبي فازي ANFIS

Optimal Control of Structures Equipped

كاربرد ميراگرهاي MR به دليل ظرفيت بالا و مصرف انرژي كم، امروزه براي كنترل پاسخ‌هاي لرزه اي سازه ها گسترش فراواني پيدا كرده است. پس ارايه روش‌هاي كنترلي هرچه دقيق‌تر و نزديك‌تر به (عوامل غيرخطي، عدم قطعيت‌ها و ...) مهندسين را در به كارگيري هرچه مفيدتر اين نوع ميراگر ياري مي‌كند. در اين مقاله براي دست يافتن به كنترل گري كه بتواند در حين زلزله ولتاژهاي ورودي به ميراگرهاي MR را هرچه سريع‌تر و دقيق‌تر تعيين كند، رهيافت كنترلي بهينه اي براي ساختمان‌هاي مجهز به ميراگر MR با استفاده از شبكه عصبي فازي ANFIS ارايه شده ؛ نشان داده شده است كه با به كارگيري اين روش مي توان به سرعت و با دقت ولتاژهاي بهينه ميراگرها را درحين وقوع زلزله تعيين نمود. براي نشان دادن تاثير پارامترهاي موجود در تابعك هدف و مقايسه كارايي اين كنترل گر نسبت به روش‌هاي معمول پيشين، دو مثال عددي مورد بررسي قرار گرفته است. نتايج حاصل از كنترل، كاهش مناسبي را در جابه‌جايي‌هاي جانبي نسبي، شتاب‌هاي مطلق و برش پايه نشان مي‌دهند و همچنين طراح كنترل گر را در تعيين ضرايب تابعك هدف راهنمايي مي‌كند تا بتواند كاهش مناسبي را در تمام معيارهاي استاندارد دلخواه داشته باشد.

Nowadays, due to high capacity and with MR Dampers Using ANFIS low energy consumption of MR dampers, they are vastly utilized to control seismic responses of structures. Thus presenting control methods which are more precise and include more actual physical behaviors (e.g. nonlinearities, uncertainties and …) will help engineers employ this kind of damper more efficiently. In order to achieve a controller that quickly and accurately commands the input voltages to the dampers, in this paper, a strategy for optimal control of structures equipped with Magneto-Rheological (MR) dampers using Adaptive Network based Fuzzy Inference System, ANFIS, is presented. To obtain optimal time histories of voltages, firstly, a new objective functional (J) is suggested and some control criteria for reduction of relative drifts, absolute accelerations and absorbed energy by elements of stories and also input voltages of dampers are included in it. Next, considering the equation of motion of the structure and a nonlinear model of MR damper, the objective functional is minimized by the optimal method (described by Moharrami and Fayezi [3]) so that the optimal time histories of input voltages of dampers are numerically computed for a 15-storey building subjected to two deterministic earthquakes. In this method the voltages are calculated without any limitation that assumes the earthquake signal is a white noise or a Gaussian white noise. Next, velocities of stories and optimal voltages are used as input-desired output data to learn the ANFIS as a quick and accurate controller. Finally, to establish a context for assessment of the effectiveness of the proposed strategy in comparison with other conventional methods and to analyze the effects of coefficients in the objective functional, two numerical cases are presented. In the first case the mentioned 15-storey building, by giving three pairs of values to the coefficients in the objective functional, is controlled against some earthquakes which are not applied for learning process of ANFIS. Results of control show quite much reduction in relative drift, absolute acceleration and base shear and also will help designer of the controller apply appropriate values of the coefficients so as to reach a proper reduction in all of desired performance criteria. In the second case the three proposed controllers show more flexibility in reduction of the standard criteria in comparison with the conventional methods