Research On Control Method Of Electric Inertia System For Brake Tester

The brake tester is a special mechanical equipment to synthetically mensurateand analyze the brake performance of rotational mechanical device. Mechanicalinertia flywheel is usually used to imitate rotational machinery inertia at present.This method can provide a sufficient and accurate reappearance of the brakemachine. But because there is a inertial mass flywheel group in the tester, itinfluences the tester with the following several shortcoming, such as the dynamicequilibrium of the flywheel is highly required; the weight and volume of theflywheel group is large; there is simulation differential, so we can't regulate inertiasteplessly; we can't imitate the vertical operating mode etc.Due to the shortcomings of traditional friction brake tester, considering theDC motor in the original system, we propose to reduce or cancel mechanical inertiapart, utilize a certain technology to compensate DC drive system appropriately, andthe dynamic characteristic this new system will be consistent with the characteristicof system with big quality inertia flywheel. This control technology is rotationalinertia electric simulation technology abbreviated as electric inertia technology.The machinery system with this technology will not only have simple structure andhigh automatic degree, but also can improve experimental precision, and theelectric simulation system of machinery rotational inertia is abbreviated as electricinertia system.Electric inertia system realizes through rotational speed control, and the outputrotational speed of the motor must imitate the rotational speed of the traditionalfriction brake tester. Because two closed-loop DC drive system is becomingaccepted universally, and its project design method is simple, convenient andpractical, electric inertia system will adopt the separately excited DC motor, supplyadjustable voltage with silicon controlled rectifier, and the study on the electricinertia system control method is based on two closed-loop DC drive system andinvariable moment test project.The electric inertia system is a speed regulation system that uses DC motor asthe core, and this system should use varying-voltage method to regulate motorspeed optimistically. This paper deduces the transfer function and dynamic modelof two closed-loop DC drive system, and it is known that two closed-loop systemhave good stable-state characteristic and can meet dynamic performancerequirement. According to the project design method, rotational speed loop andelectric current loop adopt PI regulator, and should be corrected as Model I andModel II system respectively, so as to ensure ASR good anti-disturbingperformance and good following performance. During invariable moment test project, through comparing dynamic equationsof mechanical inertia system and two closed-loop system we can know thatchanging voltage of rotational speed loop or electric current loop can realizeelectric inertia system theoretically. Considering the amplitude limitation ofrotational speed loop and the perturbation of brake load etc., electric simulationsystem should be realized by regulating the voltage of electric current loop. Whilediscussing the working state of the motor, we conclude that part of motor energycompensate the rotational inertia losing when big flywheel changes into smallflywheel, and other part overcome the loss of the motor itself. For the influence ofbrake moment perturbation, this electric inertia system applies feedforwardcompensate control of brake perturbation, and becomes a complex DC drivesystem. This paper also analyses the control method of electric inertia system undervariable moment test project, and set up the systematic structure principle chart. The two closed-loop DC drive system above adopts traditional PID regulatorthat often can't meet parameter change and non-linear characteristic of the target,and the step response of the system will overshoot. But the neural networks canlearn about the system's uncertainty and non-linear, find out systematic structureand parameter, and provide the control law that control system needs. Thecontroller algorithm based on neural networks is simple, ready to operate, and stillcontains good control performance even when the target's parameter is changing.The single neuron is elementary cell of the neural networks, and it also is the basiccontrol part in the neural networks control. To the single neuron controller, its adaptive function is realized throughchanging weight value, and its core is learning algorithms that reflect its learningability. The learning algorithm of electric inertia system has the ability ofself-teaching and self-organizing, and can make adjustment online to the coefficientof integral, proportion, differential according to the error reflected by learningsignal, which make the single neuron controller equivalent to a adaptive PIDcontroller with variable coefficient. At this moment, the dynamic performance ofsystem only depends on its error signal, can immune to the change of target's modelparameter, and has high working performance and strong robustness. Thiscontroller has also utilized the peculiar non-linear characteristic of neuron, breaksthrough the limitation of the linear regulator, and realizes the steady saturationcontrol of rotational speed loop. The electric current loop adopts PI regulator, andis verified as a Model I system, while the rotational speed loop adopts singleneuron adaptive PID controller. Based on the electric inertia systemic model, we did the control method...