Transfer learning has been proposed as an effective approach to address the problems of insufficient data. It has been utilized for condition monitoring of wind turbines (WTs). Most works focused on how to transfer the knowledge of fault diagnosis from a single source WT to target WT, which overlooked the potential useful knowledge from other source WTs. In order to utilize more knowledge of source WTs and address the challenge of data distribution alignment between source WTs and the target WT, a dual-source domain adaptation-based approach is proposed for condition monitoring. A two-stage prediction framework is employed to align the domain-specific distribution and domain-specific regression prediction. The proposed approach is validated by a case study on condition monitoring of WT. Results show that the performance of condition monitoring of the target WT using dual-source domain WTs adaptation is better than that of using single-source domain WT adaptation.

This paper proposes and compares transient currents limiting techniques for a variable speed drive short term power interruption ride-through improvement module. The module comprises of an additional capacitor that is switched onto the dc link during a short-term power interruption. The switching on of the additional capacitor to the dc-link results in high inrush currents which if not mitigated could damage the variable speed drive components and cause nuisance tripping of incoming supply breakers. Proposed inrush current limiting techniques are investigated which include the use of a fixed resistor, inductor, variable resistor and the equipotential voltage switching. Experimental results indicate that the variable resistor technique results in the least torque pulsations of 12% and 9% at the start and end of dc-link compensation. The inductor technique produced the greatest disturbance at the start of compensation with a torque pulsation of 15%. The equipotential voltage switching technique produced torque pulsations of 6.8% and 9.8% at start and end of compensation respectively. It also produced the least drops in torque and speed of 9.9% and 3.2% respectively.

The specific energy (kWh/kg) in energy storage systems (ESS) [capacitors and batteries] presents a major stumbling block into the wild breakthrough of electric vehicles(EVs). Consequently, this results into less mileage coverage by EVs which is further exacerbated by the less accessibility of EV charging infrastructures. Even if EV chargers are available, the low C-rate in batteries impedes the charging process. As a result, EV manufacturers increase on board battery sizes in EVs. Unfortunately, this make EVs to be heavy and therefore more expensive as compared to internal combustion engine (ICE) vehicles. Therefore it becomes extremely imperative that the onboard energy must be used optimally during a drive schedule. This paper develops a Matlab/Simulink electric bus model based on the insert permanent magnet synchronous machine drive train and then uses it to come up with an energy-efficient driving strategy. Optimal velocities were then generated for each road gradient of the drive route. When compared to a conventional velocity profile, consumption rate of approximately 8.9% can be saved on the (8360−8910m) range that have a road gradient of 0.076.

This paper utilizes the energetic macroscopic representation formalism to model a port crane load hoisting system. A rule-based energy management strategy is developed to manage the power split between the battery, supercapacitor, and grid energy sources. The crane model developed includes the mathematical model, the crane’s local control system, and a MATLAB/Simulink model for simulation. This study investigates and confirms the suitability of a battery-supercapacitor hybrid energy storage system for port crane applications. This is done by evaluating two objectives which are using the battery-supercapacitor hybrid energy storage system to reduce both the crane’s energy consumption from the grid and the peak power it consumes from the grid. The supercapacitor achieves 100% peak shaving for energy consumed from the utility grid and the battery achieves 65% energy savings during peak time based on the energy price arbitrage technique.

The paper presents for the first time the use of a low cost Hall effect position sensor for a synchronous reluctance machine (SynRM). The use of vector control for a SynRM requires information of the rotor position. Already the SynRM is independent of the use of permanent magnets (PMs) which results in reduced cost of the SynRM. Hence the use of a low cost position sensor contributes to an additional cost reduction of the electric drive for traction applications. The paper also presents a comparison between the low cost position sensor with the digital hall effect position sensor used in the electric power steering (EPS) machines. The different dynamics of the SynRM were demonstrated through the use of the low cost position sensor. The low cost position sensor had a relatively simpler algorithm for position information and better position resolution. Through the use of the low cost position sensor the machine produced low torque ripples of less than 9% with a total current harmonic distortion (THD) of less than 2%. The proposed position sensor registered an error less than 1 mechanical degree when compared with a 12-bit industrial encoder.

This paper investigates the performance of a two level and a three level inverter drives for a permanent magnet synchronous machine (PMSM) under different switching frequencies. The experiments were conducted via a real time based system called Opal-RT. The inverters were operated at 1 kHz and 5 kHz switching frequencies. The torque was varied between 2 N.m and 4 N.m at a constant speed of 800 rpm. However the torque ripples and the copper losses between the two inverter drives were similar and the copper losses were independant of the swicthing frequency. At 5 kHz inverter switching frequency the two level inverter performed better at 4 N.m with a total current harmonic distortion of 2.55% and a torque ripple of 10%. However overall the three level inverter had low voltage and current harmonic distortion which translates to lower iron losses. Swicthing at 5 kHz frequency showed to have lower overall harmonic distortion than at 1 kHz for both inverters. The greater the capcitor voltage error the more the torque ripples even though the machine registered lower current total harmonic distortion for the three level inverter.