The current loop is the key to realizing high-precision permanent magnet synchronous motor (PMSM) drives. However, the performance of the current loop is deteriorated by the dc and ac disturbances. To deal with the problem, this article proposes a discrete-time repetitive control-based active disturbance rejection control (ADRC) for the current loop. The discrete-time repetitive controller is embedded in the control law of the ADRC and operates in parallel with the discrete-time extended state observer (DESO), which is the core of the ADRC. The repetitive controller is designed to compensate for the ac disturbance, while the DESO is designed to suppress the dc disturbance. The stability and antidisturbance capability of the studied scheme are analyzed theoretically in the discrete-time domain, which can guide the tuning of control parameters. Compared with the existing ADRC-based PMSM control schemes, the studied scheme can suppress the dc and ac disturbances simultaneously, thereby improving dynamic- and steady-state performance of the current loop. Finally, the effectiveness of the studied scheme is validated on a 4.4-kW PMSM drive platform.

This paper proposed a quadrature phase-locked loop (PLL) with a second-order complex-coefficient filter (SCCF) to improve the position estimation accuracy of sensorless interior permanent magnet synchronous motor (IPMSM) drives. Due to the inverter nonlinearity and flux spatial harmonics, it exists 5th and 7th harmonics, and dc offsets in the back-electromotive force, reducing the accuracy of estimated position. For this problem, the studied SCCF-PLL can suppress the harmonics and dc offsets with the operation-frequency-adaptive character. The stability of the studied PLL are analyzed by establishing a small-signal model. Compared with the conventional PLL, the studied PLL can reduce the estimation error of the position observer effectively. The availability of the studied method is verified by the contrastive simulation results.

In the traditional overmodulation, the reference of voltage loop is selected as a constant. However, it keeps the motor working in the six-step mode regardless of torque demand, resulting in current and speed harmonics. To address this problem, an operation-area-selected (OAS) overmodulation strategy is proposed for flux-weakening control of surface-mounted permanent magnet synchronous motor. The compensation of the inscribed circle reference voltage is obtained by logical judgments of two designed reference voltages in an auxiliary closed-loop. The proposed scheme can adjust the reference of voltage loop dynamically according to different working conditions, leading to reduced current and speed ripples. The simulation results verify the effectiveness of the studied scheme.

Under low PWM-to-output frequencies ratio conditions, the performance of the discrete adaptive full-order observer (DAFO) deteriorates sharply with the forward Euler (FE) method. This seriously affect the stability of speed sensorless induction motor (IM) drives. To address the problem, this paper proposes an improved DAFO with the "prediction-correction" method. The studied method can improve the discretization accuracy of the observer. Compared with the conventional DAFO, the proposed DAFO ensure stable operation at low PWM-to-output frequencies ratio. The experimental results from a 3.7kW IM drive system verify the effectiveness of the studied observer.

Incremental encoders are popular for detecting angular position and speed. However, the conventional incremental encoder-based methods for estimating speed are prone to poor performance at low speed where the encoder output rate is correspondingly low, leading to a number of research efforts in improved low-speed detection algorithms. To improve speed control performance, a speed detection method based on reduced-order observer is presented in this paper. The observer can interpolate speed and position which is the integration of speed and compared with real data from the encoder when the encoder pulse is detected between encoder pulses. Furthermore, inertia identification based on recursive extended least square algorithm is presented to reduce sensitivity of the speed estimation. Experimental results show that low speed control performance using the reduced-order order is superior to that of conventional one

Target's spectral emissivity changes variously, and how to obtain target's continuous spectral emissivity is a difficult problem to be well solved nowadays. In this letter, an activation-function-tunable neural network is established, and a multistep searching method which can be used to train the model is proposed. The proposed method can effectively calculate the object's continuous spectral emissivity from the multispectral radiation information. It is a universal method, which can be used to realize on-line emissivity demarcation.