Research On The Sensorless Control Technology Of Direct-drive Permanent Magnet Synchronous Generator For Wind Power

As a clean and renewable energy, wind energy has obtained more and morewidespread application, and wind power technology has also been developed quickly. Thetrend of the wind power generator is that single machine capacity becomes larger and larger.Since larger machines require higher reliability, power density and efficiency, thepermanent magnet synchronous motor(PMSM) has remarkable advantages over other typesof motors in the wind power generation. And along with the decrease of permanent magnetmaterial prices, the permanent magnet direct-driven wind power system has got rapidlydevelopment in recent years.In order to ensure the security and efficient operation of permanent magnetdirect-driven wind power system, permanent magnet synchronous generator(PMSG) mustbe well controlled to meet the critical requirement. Generally, PMSGs for wind powergeneration are low rotation speed multipolar generators. There are two challenges in thecontrol of PMSG needing special attention: one is ensuring the static and dynamicperformance of current controller and the other is obtaining the accurate information ofrotor position and speed. This paper proposed an improved deadbeat predictive currentcontrol algorithm to overcome the weakness of the robustness to parameter variation. Inorder to get accurate real-time information of the rotor position and speed, and eliminate theproblem of the mechanical encoder installation and poor reliability of encoder, a sensorlesscontrol technology which is suitable for direct-drive wind power system based on thephase-locked loop theory was developed.The essence of the deadbeat control is a kind of control method based on model, whichhas the weak point of poor robustness to parameter variation, and the inherent one-step lagin the digital control will further reduce the robustness. This paper proposed an improveddeadbeat predictive current control algorithm based on predictive current control, i.e.predicting the motor current in one-step lead, thus making up for the influence of systemone-step lag and improving the system robustness to the parameter variation. Simulationsand experiments have validated the correctness and effectiveness of the theoretical analysis. An observer was designed based on PLL estimation of rotor position and speedaccording to PMSG electromotive force equation and phase-locked loop structure. In orderto solve the problem of the difficulty in obtaining precise motor parameters and parametervariation in generator operation, the influence of the parameter variation on the sensorlesscontrol technology was discussed. Study shows that q-axis inductance has the maininfluence on the precision of estimation, and current controller stability will be subject tocertain limitations when estimation error occurs.Finally, a10kW prototype system was designed and established, which has validatedthe effectiveness of the proposed improved deadbeat predictive current controller andsensorless control technology based on phase-locked loop.