CAREER: Parameter Estimation for Ill-Conditioned Systems with Application to Electric Drives and Power Systems

该研究计划的重点是病态动力系统的参数估计及其在电力驱动和电力系统中的应用。电力驱动和电力系统中的物理限制限制了参数估计可用数据的质量。因此，使关联参数估计成为一个病态估计。111-条件调节防止获得所有系统参数的一致的良好估计。本研究提出发展合理的方法来处理电力传动和电力系统中的病态参数估计问题。首先，我们建议发展量化疾病严重程度的方法。其次，我们建议开发可应用于产生物理上有意义的参数估计的病态问题的参数估计算法。从系统仿真，稳定性分析，优化，状态估计和控制需要准确的电力系统参数来看，开发可处理低质量数据的鲁棒参数估计算法的重要性显而易见。随着监管的放松，以及电力系统在接近满负荷、安全/储备边际较低的情况下运行的需求，这一点变得更加重要。在电力驱动领域，无传感器控制的强劲增长促使需要开发鲁棒参数估计算法，因为大多数确定机械位置或速度的方法在很大程度上依赖于准确的机器参数。由于电力系统的大规模和电力驱动应用中实时实现的限制，这项工作的一个重要焦点是用于状态评估和参数估计的算法的计算效率。大规模问题的高效算法开发一直是波多黎各NSF-EPSCOR计划下的高性能计算（HPC）计划的重点，HPC仍然是该岛科学发展的主要信任领域。将这一研究项目加入到高性能计算工作中，将加强这一领域的应用组成部分，有助于实现PR-EPSCOR项目的目标。教育计划中提出的活动将影响到本科生、研究生和高中生。作为教育计划的一部分，将开展四项主要活动：修订和开设研究生和高级水平的电力驱动和电力系统建模和控制领域的课程，修订本科电力系统实验室并将其纳入电力系统本科课程。指导研究生和本科生，并让他们参与研究生和本科生水平的研究项目。加强PI为参加upm暑期工程预科课程的高中生提供的电力系统介绍讲习班。教育计划的课程开发部分建立在以技能为基础和以实践为基础的教育活动的基础上，包括合作学习和教学方法的使用以及研究人员参与的作为nsf资助的制造工程教育合作伙伴关系的一部分的新颖评估技术。将对课程进行修订，使研究工作融入课堂。工程前讲习班将修订形式和内容，编写的材料将分发给其他机构和高中教师。研讨会和高中教师活动将与普遍定期审议科学与工程资源中心合作进行，该中心部分由国家科学基金会资助。

ECS-9702860 Velez-Reyes The research plan is focused in the area of parameter estimation for ill-conditioned dynamical systems and its application to electric drives and power systems. Physical restrictions in electric drives and power systems limit the quality of the data available for parameter estimation. Therefore, making the associate parameter estimation an ill-conditioned one. 111-conditioning prevents uniformly good estimates of all system parameters from being obtained. This research proposes to develop sound methodologies to deal with ill-conditioned parameter estimation problems in electric drives and power systems. We first, propose to develop methodologies to quantify the severity of the ill-conditioning. Second, we propose the development of parameter estimation algorithms that can be applied to ill-conditioned problems producing physically meaningful parameter estimates, The importance of developing robust parameter estimation algorithms that can deal with low quality data is evident from the need of accurate power system parameters for system simulation, stability analysis, optimization, state estimation, and control. This is becoming even more important with deregulation and the need of operating power systems near their full capacity with lower safety/reserve margins. In the area of electric drives, the strong growth of sensorless control have prompted the need to develop robust parameter estimation algorithms since most approaches to determine mechanical position or speed depend heavily on accurate machine parameters. An important focus of this work is the computational efficiency of algorithms to be developed for condition assessment and for parameter estimation because of the large scale of power systems and the constraints in real-time implementation in electric drive applications. Efficient algorithm development for large scale problems has been a focus of the High Performance Computing (HPC) initiative under the Puerto Rico NSF-EPSCOR program in the past and HPC is still a main trust area of scientific development in the island. The addition of this research program to the HPC effort will strengthen the applications component of this area contributing to the attainment of PR-EPSCOR program goals. The activities proposed in the educational plan will impact undergraduate, graduate, and high school students. Four major activities will be undertaken as part of the educational plan: Revision and creation of courses at the graduate and advanced senior level in the areas of modeling and control of electric drives and power systems, Revision of the Undergraduate Power Systems Laboratory and its integration into undergraduate courses in power systems. Mentoring of graduate and undergraduate students and their engagement in graduate and undergraduate level research projects. Enhancement to the Introduction to Power Systems Workshop offered by the PI to high school students participating at the pre-engineering summer program at UPRM. The curriculum development component of the education plan builds on activities in skill-based and practice-based education including the use of the cooperative learning and teaching methodologies and novel assessment techniques in which the researcher has been involved as part of the NSF-funded Manufacturing Engineering Education Partnership. The courses will be revised to integrate the research work into the classroom. The pre-engineering workshop will be revised in format and content and the materials developed will be disseminated to other institutions and high school teachers. The workshop and high-school teacher activities will be done in collaboration with the UPR Resource Center for Science and Engineering partly funded by NSF.