双井抽油机用频繁换向大功率开关磁阻电机电磁力矩冲击机理及抑制

Energy efficiency could be improved by 40% when a novel dual oil well pumping unit driven by the Switched Reluctance Motor (SRM) substitutes for the conventional beam pumping unit. However the new-type pumping unit suffers a lot from violent shocks caused by the motor's frequent reversal operations and its heavy time-varying working load, which compromises safety and stability of the oil pumping equipment. The severe working conditions paralyzes conventional SRM torque ripple minimization techniques, and previous relevant research results hardly achieve high performance for suppressing the vibration, shock, and noise of the high power SRM systems. ..In this study, characteristics of the torque ripple and chaotic instability are comprehensively analyzed to discover the origins of the SRM's shocks and explore the underlying principles of the dynamic system oscillation. ..Then feature abstraction and pattern reorganization techniques are introduced to process the Extended Status Time-series of the SRM dynamics. The time-domain evolutionary track of the SRM chaotic dynamics is reconstructed within the transformed sparse feature space, and corresponding mapping algorithms are proposed, hereby a novel online real-time chaos identification model is built to forecast the motor driving system dynamics...With the aid of proposed chaos forecasting model, a composite scheme is promoted to prevent the motor torque shocks via combining the torque vector controller with a robust complementary impulse controller for chaos suppression. To implement the proposed models and algorithms an embedded real-time control systems is to be developed on the multiple computational cores of DSP+FPGA+ARM hardware... Consequently the presented control systems are validated and tested on a dual oil well pumping unit platform. Research results will yield fundamentals for SRM's stable control and pave way for long-term safe operation of the dual oil well pumping unit.

基于大功率开关磁阻电机驱动的双井抽油机替代传统游梁式抽油机可节能40%，但时变重载条件下电机频繁正反转引起的电磁力矩冲击，严重影响设备的安全性和稳定性。现有转矩脉动抑制理论无法从根本上解决频繁换向的大功率开关磁阻电机振动、冲击、噪声抑制难题。.本项目通过研究开关磁阻电机转矩脉动与混沌失稳的耦合关系，揭示电机电磁力矩冲击成因及系统振荡失稳机理。.研究开关磁阻电机动力学状态变量的时间序列信号获取、特征提取与模式识别，提出稀疏特征空间内混沌演化轨迹重构与数值计算方法，建立开关磁阻电机驱动系统混沌特性在线实时判定与预报模型。 .提出转矩矢量控制与混沌系统脉冲控制相结合的复合控制算法，开发基于DSP+FPGA+ARM多计算核心的大功率开关磁阻电机嵌入式实时控制软硬件实验系统，并在双井抽油机实验场验证理论模型与控制算法有效性。研究成果将为实现油田双井抽油机长期稳定、安全、可靠运行提供理论和方法。

【研究背景】本项目通过研究开关磁阻电机转矩脉动与混沌失稳的耦合关系，揭示电机电磁力矩冲击成因及系统振荡失稳机理。.【研究内容】研究开关磁阻电机动力学状态变量的时间序列信号获取、特征提取与模式识别，提出电磁转矩扩张状态空间内实时混沌动力学演化轨迹重构与数值计算方法，建立开关磁阻电机驱动系统混沌特性在线实时判定与预报模型。.【重要结果】提出滞环调速控制、转矩实时分配控制与混沌系统脉冲控制相结合的复合控制算法，开发基于DSP+FPGA+ARM多计算核心的大功率开关磁阻电机嵌入式实时控制软硬件实验系统，并在多功率等级实验场验证理论模型与控制算法有效性。研究成果为实现油田双井抽油机长期稳定、安全、可靠运行提供理论和方法。.【关键数据】电机动力学实时观测方面，采用CMAC的嵌入式开关磁阻电机电磁转矩矩动力学建模比传统神经网络方法计算精度和实时性同时提高2个数量级（模型计算实时性由185us提升至2.7us，模型计算平均误差由2.2%提升至0.20%），实现了电机基于模型的FPGA高速硬件嵌入式转矩在线实时解算；电机动力学特性方面，采用基于动力学扩张状态观测和多核心同步计算方法，将混沌定性预报实时性又传统5000数据点提升至小于1500采样点，对应于3000rpm转速为4个换向周期，为混沌动力学的早期预报提供了工程性借鉴方法。.【科学意义】高精度非线性实时多核协同设计与解算模型为复杂电机高性能驱动奠定了工程基础；电机混沌动力学的扩张状态观测、转矩实时分配控制为非线性系统强实时性控制方法探索出了新方法。