ДОСЛІДЖЕННЯ РЕГУЛЯТОРА ПОТУЖНОСТІ ДУГ ДУГОВОЇ ПЕЧІ 3 НЕЙРОРЕГУЛЯТОРОМ MODEL REFERENCE CONTROLLER
Електронний архів E-archive DonNTU – (Electronic archive Donetsk National Technical University)
Переглянути архів Інформація| Поле | Співвідношення | |
| Title |
ДОСЛІДЖЕННЯ РЕГУЛЯТОРА ПОТУЖНОСТІ ДУГ ДУГОВОЇ ПЕЧІ 3 НЕЙРОРЕГУЛЯТОРОМ MODEL REFERENCE CONTROLLER
Исследование регулятора мощности дуг дуговой печи с нейрорегулятором Model Reference Controller Study of Arc Furnace Arcs Power Regulator with MODEL REFERENCE CONTROLLER Neural Controller |
|
| Creator |
ЛОЗИНСЬКИЙ, О.Ю.
ЛОЗИНСКИЙ, О.Ю. LOZYNSKYY, O.Y. ПАРАНЧУК, Я.С. PARANCHUK, Y.S. МАЦИГІН, А.Б. МАЦИГИН, А.Б. MATSYHIN, A.B. |
|
| Subject |
дугова сталеплавильна піч
регулятор потужності дуг нейрорегулятор комп 'ютерна модель стабілізація дисперсія перерегулювання дуговая сталеплавильная печь регулятор мощности дуг нейрорегулятор компьютерная модель стабилизация дисперсия перерегулирование arc steelmaking furnace arcs power regulator neural controller computer model stabilization dispersion overshoot |
|
| Description |
The issue of complex electrotechnological efficiency indices improvement during steel melting in electric arc fumaces is urgent and important for electric metallurgy industry of Ukraine. Great energy consumption of arc furnaces metal production is the result of outdated and wom out electric mode control systems. Existing arc power regulators are characterized by a significant lag, nonlinearity, limited stiffness and allow significant dispersion of electric mode coordinates during control process. Such characteristics do not allow to obtain accurate mathematical models of the furnace power circuit and control system. We propose to increase performance and decrease power consumption through the improvement of electric mode coordinates control systems. To improve dynamic accuracy of arcs lengths and electric mode coordinates control it is proposed to use neural network control principles. The structure of arcs power regulator based on Model Reference Controller neural controller was designed. Neural controller is included in the input of electrodes movement electric drive thyristor converter. A digital Simulink-model of proposed arcs power control system with Model Reference Controller neural controller was developed. Identification of control object model based on neural network using a digital model was performed. Traimng of reference model, which defines the desired trajectory of the electrode movement, was performed. The appropriateness of aperiodic law of electrodes motion is substantiated. The neural network controller is composed of two layers and has 5 inputs, 13 neurons in the hidden layer and 1 output. It contains delay elements, included in inputs and outputs, which allows to give such direct neural network dynamic properties. To design neural controller and study the effectiveness of its work a digital Simulink model of control object, namely the ARDM-T arcs power regulator regarding its parameters for super-power DSP-200 electric arc furnace, was created. Neural controller synthesis is performed using Model Reference Controller block from the MatLAB Simulink library. The efficiency of the developed electric mode control system with Model Reference Controller neural controller is estimated by the results of comparisons of the dynamics when working out determimstic and random perturbations of arc length by typical ARDM-T-12 arcs power regulator with and without neural controller. For this purpose, using the created Simulink-model of ARDM-T-12 controller and model of the proposed system with Model Reference Controller neural controller the dynamics of working out identical implementations of determimstic and random perturbations of arc length have been studied. The results of research on the digital model confirmed the feasibility of the practical use of the proposed ARDM-T regulator structure with Model Reference Controller neural controller. Розроблено структуру системи регулювання потужності дуг дуговоі сталеплавильноі печі на основі неирорегулятора MODEL REFERENCE CONTROLLER. Опрацьовано структурну схему та цифрову Simulinc-модель неиромережевоі системи регулювання потужності дуг, запропоновано методику параметричного синтезу неирорегулятора та проведено моделювання динаміки регулювання координат електричного режиму. Отримані результати підтвердили поліпшення показників якості динаміки за використання розробленоі неиромережевоі системи у порівнянні з типовим регулятором потужності дуг типу АРДМ-Т |
|
| Date |
2013-10-28T06:18:59Z
2013-10-28T06:18:59Z 2013 |
|
| Type |
Article
|
|
| Identifier |
Наукові праці Донецького національного технічного університету. Серія "Електротехніка і енергетика". № 2 (15), 2013 р., 338 с.
http://ea.donntu.edu.ua/handle/123456789/23402 |
|
| Relation |
Электротехника и энергетика;
|
|
| Publisher |
ДонНТУ
|
|