| With the increasing demand on the capacity and density of the “more/all electric” aircraft power system,the integrated starter/generator(ISG)system has been drawing more and more attentions in recent decades.In the ISG system,the ISG is operated as a motor to start the aero engine in the start-up process,and as a generator supplying power for airborne equipments after the start-up.Compared with the traditional power system,the ISG system can decrease the weight and volume of the aircraft power system,which is of great significance for aircraft.Several machines have been considered for the ISG system,but due to advantages of mature power generation technology and high safety,multi-stage electrically excited synchronous machine(MEESM)becomes an attractive candidate for the aircraft ISG system.Precise rotor position knowledge is required to operate the multi-stage electrically excited synchronous stater/generator(MEESSG)in the motoring mode.Rotor position is usually obtained by a resolver or encoder fixed on the same shaft of the machine.However,the mechanical position sensor faces strict restrictions such as severe electromagnetic and space environment in aircraft applications,which make the mechanical sensor a vulnerable point and decrease the system reliability.Rotor position estimation technology can provide benefits on the above issues for the ISG system during the start-up process.For the MEESSG investigated in this paper,the armature current of the main machine(MM)varies in a wide range due to the complex load characteristics,which results in the magnetic saliency of the MM changed greatly during the start-up process.Therefore,the existing rotor position estimation methods based on the machine saliency are not effective on the MEESSG.Considering the special multi-stagestructure of the MEESSG,an alternative scheme is using the main exciter(ME)to estimate the rotor position of MEESSG(or MM).At the same time,to solve the problem of initial starting torque shock when the MEESSG is started with large inertia load by the existing starting control method,a smooth start-up control method of MEESSG is investigated in this paper.Firstly,the equivalent circuit of excitation system consisting of ME,rotating rectifier and field winding of MM was established,and the expression of ME rotor currents are deduced based on the equivalent circuit in this paper.The distribution of harmonic components in ME rotor current are obtained by Fourier transform.According to the operation principle of asynchronous machine,the expressions of harmonic currents in ME stator windings induced by rotor magnetomotive force of ME are derived,and the complete expression of ME stator current is obtained.Secondly,the rotor position of MSEESG is estimated by the way of “MM injection-ME detection” in zero and low speed region.The signal transmission mechanism in MSEESG is analyzed,and the rotor position estimation methods at zero/low-speed region based on mutual inductance between stator and rotor windings of MM are proposed.The optimum frequencies of excitation voltage and injection signal are determined by the Fourier analysis results of ME stator currents.The high/low frequency voltage signals is injected into the eauivalent α-axis andβ-axis of MM,and the sinusoidal and cosine signals of the rotor position are extracted from the stator current of the exciter.The initial rotor position is estimated by the arc tangent operation using the sinusoidal and cosine signals.On this basis,the rotating high frequency voltage signal is injected into the stator windings of the MM,and the response high frequency currents in ME stator windings are extracted by the rising and falling sampling function instead of the traditional filter.After the phase-locked loop processing,the online estimation of the rotor position in the low speed region is realized.The experimental results show that the proposed initial position estimation method has high estimation accuracy,and the proposed rotor position estimation method in low speed region is robust to the change of saliency rate of the MM.Thirdly,the harmonic currents in stator windings produced by ME rotor harmonic magnetomotive force are analyzed deeply,and a rotor position estimation method based on harmonic current in low speed region is proposed.The relationship between the specific harmonics and the rotor position of ME is analyzed,and the detailed expression of the harmonics are derived.The optimum excitation frequency and sampling frequency are determined based on the analysis results.A constant frequency sampling function is used to extract the position information from the stator harmonic currents of ME,and the rotor position is estimated on line using phase-locked loop processing.The estimated rotor position of the MM can be obtained by using the number of poles relationship between the MM and the ME.No high frequency signal is required during the rotor position estimation process,so the high frequency torque ripple and high frequency loss caused by the injected high frequency signal can be avoided in this method.The experiments are carried out and the experimental results showed that the estimated errors are very small in the low-speed region during the start-up process and the proposed method has highly robust to parameters change of the MM.Finally,in order to solve the problem that the starting of the MSEESG is not smooth and the rotor position estimation error is relatively large in the start-up process,a starting control method based on sub-time closed-loop of current/speed is proposed.The current closed-loop and speed open-loop control mode is adopted to make the MSEESG start from static state.After the machine accelerated to a lower speed,the current/speed double closed-loop is switched to complete the subsequent starting process.The experimental results show that there is no torque shock and the rotor position estimation error is small in the start-up process.And then,the sensorless smooth starting control of MSEESG under the specific aeroengine load is completed on the experimental platform. |
PDF Full Text Request