Recently, hairpin windings (HWs) have been among the most studied and implemented winding topologies to enhance torque and power density of electrical machines (EMs), especially in the automotive sector. The aim of this work is threefold. First, it proposes an asymmetrical configuration for HWs, which exploits the possibility of reducing AC losses through the implementation of conductors featuring variable cross-section within the slot. Second, this asymmetrical winding is envisioned using additive manufacturing (AM), given the constraints imposed by the conventional manufacturing process of HWs. Third, a suitable aluminium (Al) alloy is selected as winding material, thus replacing copper (Cu), with the aim of achieving a more sustainable design. These three objectives are achieved through extensive analytical and numerical analyses and comparisons between conventional and asymmetrical windings, as well as among different materials. A practical demonstrator is then built to validate analyses, methodologies and concepts proposed in this paper. Finally, an experimental campaign is carried out on the asymmetrical winding demonstrator, highlighting the loss reduction and sustainability potentials of the proposed layout.
13 February 2020 in IEEE Open Journal of the Industrial Electronics Society
Electric drives applications have been worldwide adopted for the transportation electrification. The electric drive system is constituted by the power electronics and by the electrical machine. The common design workflow implies the separate realization of the power electronics and of the machine, with many assumption regarding the operating conditions. This paper presents a unified analysis of the two subsystems considering the actual operating conditions, including the power loss of the machine due to the switching operations. As a case study, this paper presents a comparative study among three different converter topologies designed to drive a 8.5kW-120krpm surface PMSM. The study aims at comparing the considered systems in terms of converter complexity, control complexity and overall efficiency. At first converters are simulated in Matlab-Simulink to estimate the losses and the current waveforms, that are then used in the Finite Elements model of the electrical machine so to estimate the loss components in a real scenario. Experimental validation is used for the verification of the developed models. The study gives an understanding of the interactions between the sub-systems and allows concluding which converter topologies gives the best performance.
Incremental encoders and their use in electric drives are nowadays considered standard and not given much thought. Nevertheless, speed measurement through the signals coming from an incremental encoder can be subject to errors in some operating conditions if special care is not taken. Techniques exist in literature to enhance the precision of speed measurement, but they often require special hardware (i.e. FPGA) or a non-negligible portion of computing power. This paper proposes a novel technique for precise speed measurement, with minimal error, that can be implemented on a low-cost microcontroller with standard quadrature decoder peripheral.
16 August 2021 in IEEE Open Journal of the Industrial Electronics Society
When targeting mission critical applications, the design of the electronic actuation systems needs to consider many requirements and constraints not typical in standard industrial applications. One of these is tolerance to faults, as the unplanned shutdown of a critical subsystem, if not handled correctly, could lead to financial harm, environmental disaster, or even loss of life. One way this can be avoided is through the design of an electric drive systems based on multi-phase machines that can keep operating, albeit with degraded performance, in a partial configuration under fault conditions. Distributed architectures are uniquely suited to meet these challenges, by providing a large degree of isolation between the various components. This paper presents a system architecture suitable for scalable and high-performance fault tolerant machine drive systems. the effectiveness of this system is demonstrated through theoretical analysis and experimental verification on a six-phase machine.
In modular distributed architectures, the adoption of a communication method that is at the same time robust and has a low and predictable latency is of utmost importance in order to support the required system dynamics. The aim of this paper is to evaluate the consequences of the random jitter on machine drives distributed control, caused by the messages’ re-transmission in case of an error in the received data. To achieve this goal, two different Forward Error Correction (FEC) techniques are introduced in the chosen protocol, so that the recipient of the message can correct random errors without the need of any additional round trip delays needed to request and obtain a re-transmission. Experimentally validated simulations are used to evaluate the impact of random network derived jitter on a real world closed loop control system for distributed power electronic converters.
14 September 2021 in IEEE Open Journal of the Industrial Electronics Society
In applications that require a high availability and high performance (for example aerospace),modular power electronics and multi-phase machines represent an advantageous choice. In this framework, a control system able to handle a high number of PWM signals and communication interfaces as well as featuring a high computational power is required. This paper proposes a novel HDL plus soft-core approach to be implemented on System-on-Chip hardware which allows for the efficient and modular implementation of the modern control techniques with strong guarantees in terms of determinism. The proposal lies in the adoption of a very simplified and optimized floating-point soft-core, the femtocore (fCore) and its tool-chain, which allows C-like implementation of complex algorithms in a HDL-design power electronics control framework. Several fCore units can be arranged for parallel processing to handle the time requirements of a complex modular system even with low sampling time (100 kHz or more). The proposed architecture is experimentally validated in a proof-of-concept, six-phase electric machine including a comparison against a traditional method.