Workshops, Tutorials, and Sessions
Pre-Conference Workshops (Sept 6 afternoon, 4 hours. Fee: $150 for full registration and $70 for students)
1. Practical Methods for Real World Control Systems (Daniel Y. Abramovitch; Agilent Technologies, Inc.)
Presenter: Daniel Abramovitch
A question one should ask of any advanced algorithm is, “How do we make that work in a real system?” A question one should ask of any industrial control system is, “How do we apply better algorithms to this problem?” The two questions are dual sides of the same “bridging the gap” problem that has hounded control for decades. This workshop will examine practical methods that address this problem from both sides: ways to implement advanced algorithms on real systems and ways to improve industrial control using advanced methods. We will examine which system identification methods work on which physical systems, as model-based control requires a model. We will discuss why so many industrial controllers are PIDs, present a universal framework for different PID implementations, describe how to tune the PID to the identified system model, and show how to augment these with higher order controller dynamics (a.k.a. filters). We will discuss how to make state-space models more useable in real-time systems. Speaking of which, we will explain how to program filters and PIDs in real-time control systems. We will discuss things to know about hardware implementation and tradeoffs with ADCs, DACs, and analog filters. We will present a unified view of feedforward control methods and how to incorporate them in a practical way with feedback systems. We won’t bridge the gap in an afternoon, but we can move the needle.
A web page that holds the information from the brochure can be found here.
A PDF version of the workshop flyer can be found here.
A 15-minute invitation video to an earlier version of this workshop is here.
Prerequisites skills for participants: Undergraduate level knowledge of feedback systems, sampled data systems, and programming. An honest interest in being able to translate control theory into physical control systems. The workshop is designed to be useful to industry practitioners wishing to apply more advanced control methods as well as academics wishing to make their algorithms more applicable to real world problems.
2. Smart Predictive Maintenance (SPM) of Mechatronic Systems based on Smart Big Data (SBD) and Digital Twins (YangQuan Chen*, Pawel D. Domanski, Furkan Guc, Jairo Viola, Jing Wang; The University of California, Merced)
Presenters: YangQuan Chen and Furkan Guc, Jairo Viola, Pawel Domanski, Jing Wang
In engineering designs, we employ a variety of sensors, actuators, processes and structures to achieve desired objectives within an acceptable performance window. Although the design process consists of lots of consideration regarding performance, maintaining successful operation may be a challenging task for many engineering applications. Therefore, the field of health monitoring, fault diagnosis/prognosis and predictive maintenance is expanding dramatically to cope with the performance degradation over time of the corresponding systems.
For both academia and industry, mechatronic control system design starts with asking two main questions 1) “What do we have/know?” and 2) “What do we want?” to comply with the systematic approach in the design process. After the engineering design, we need to give answers to two questions 1) “How optimal?” and 2) “How robust?”. With the aid of increasing application of “Digital Transformation” such as Industry 4.0, we emphasize the importance of asking the third question: “How smart?”.
Within the context of this workshop, the concept of “smartness” is re-evaluated in the sense of Smart Predictive Maintenance (SPM). A smart predictive maintenance program can learn from past actions and errors, uncover hidden patterns in complex signals, detect anomalies within regular system life time and classify possible fault scenarios with comparative examples.
This workshop introduces a new frontier for engineers: Smart Predictive Maintenance (SPM), using Smart Big Data (SBD) and Digital Twins (DT) as the enabler technology combined with IAI (industrial artificial intelligence) and groundbreaking technologies like Deep Learning, AI, Data Analytics, Big Data, and edge-computation. We will introduce not only the theoretical framework of SPM, but also employ example case studies of mechatronic applications from various systems. Matlab Predictive Maintenance Toolbox will be used in this tutorial.
For more information: https://mechatronics.ucmerced.edu/spm
Tutorials (part of the conference, open to all conference attendance)
1. Control co-design of Wave Energy Converters (Dr. Giorgio Bacelli, Sandia National Laboratories)
The concept of impedance matching for maximization of energy absorption by a wave energy converter (WEC) has been well understood for perhaps 40 years. In the intervening decades, research in this area has focused primarily on the practical means with which the impedance matching condition can be achieved (or approximated). More specifically, the problem of WEC control design is often posed in terms of maximizing mechanical power. The assumption in this approach of designing for mechanical power, which is often made implicitly, is that the hydromechanical system can be decoupled from the drive-train and electrical systems. As has been recognized previously, and will be shown in this tutorial, such an assumption can be quite erroneous, and can lead to exceptionally poor performance. In fact, since the power take-off (PTO) system can, and often does, contain dynamics, a controller designed based on the hydro-mechanical system can produce negative electrical power. Further, while much attention has been given to improving control design, the marginal improvement available for a higher performing controller is actually quite small compared with potential benefits obtainable by improving the PTO and system design to align with the controller function (so called “control co-design”).
Clearly, the WEC control problem must be reformulated to more fully represent the WEC system and enable a control co-design approach. In this tutorial, we show how to extend the linear theory for maximization of power absorption by a WEC to include the dynamics of the PTO. By representing the subsystems of the WEC as cascading multi-port elements, we maintain the ability to apply impedance matching principles, but may now do so with a complete “wave-to-wire” view, that is, with the objective to maximize the useful power at the output of the device.
In many previous studies, the efficiency of the PTO is captured through a single term that represents all losses. While using a single efficiency point can work well in static systems, the inherent oscillator nature of WECs makes a static efficiency model impractical. Alternatively, a nonlinear efficiency term (e.g., where efficiency is a function of torque and rotational velocity) can be used, but this formulation can obscure system dynamics and interactions. A nonlinear efficiency model can often appear as a “black-box” within a design study, forcing a “guess-and-check” approach and limiting co-design. The cascading impedance model formulation shown herein extends the hydro-mechanical impedance modeling to encompass the entire WEC system. By using this approach, it is possible to use network synthesis methods to design WEC PTO and control systems, not just based on trial and error, but with direct knowledge of the complete dynamical system.
The co-design problem will be effectively illustrated in this tutorial by considering a generic oscillating body WEC, the motion of which is constrained to a single degree of freedom. This first part covers the formulation of a generic WEC model, including both hydrodynamic interactions and PTO, as a multiport system. By means of this approach, it is shown that many of the results currently available for the design of controllers for maximization of mechanical power can be still be used, provided the impedance model is adequately derived from the multiport formulation. The second part covers detailed numerical example which illustrate the enormous benefit of using a control co-design approach compared to the classical approach of separately designing the PTO and control system from the hydrodynamics. The Matlab code will be made available on GitHub.
2. A Tutorial on Computational and State Space Models for Real-Time Control of Mechatronic Systems (Daniel Y. Abramovitch, Agilent Technologies, Inc.)
The day may come when mechatronic control engineers do not need to worry about the speed of their real-time computation – but it is not this day. A day when we move from high order Simulink models to functional hardware, sampling at 200 kHz with the mere touch of a button – but it is not this day. A day when even microsecond sample periods provide enough time to do Model Predictive Control optimizations between each sample – but it is not this day. A day when we can drop 1000 parameter models on a system without worrying about their physical meaning or which represent clusters of lightly damped dynamics – but it is not this day.
Today we fight … for every scrap of computation time and latency minimization… to minimize noise without dramatic increases in latency and the severe negative phase and bandwidth limiting implications… to properly identify highly flexible dynamics and build filter and state-space structures that preserve both numerical fidelity and physical intuition in our real-time models.
Actually, not today, but on Wednesday, September 8, at the 2022 IFAC MS/MoViC, I will give a tutorial explaining a lot of this stuff in very practically applicable terms. I will discuss the three-layer computation model I’ve been teaching for a few years as a way of understanding how to program different parts of a control system. I will discuss issues with converter chips and how they affect the choice of sample rate and the effective delay. Finally, I will teach the Biquad State-Space (BSS) structure, a state space form that not only has outstanding numerical fidelity, but also excellent physical intuition. The BSS has the unique property that the discrete-time states and the continuous-time states – taken two at a time, correspond to each other.
So, for all you hold dear in making advanced control methods work in high-speed, lightly-damped, mechatronics systems on this green Earth – I bid you to come to the tutorial in the West (actually at UCLA).
3. A Tutorial on the Forgotten Signals in Control (Daniel Y. Abramovitch; Agilent Technologies, Inc.)
An old control engineer once said, “When you’ve done everything else right, what limits you is time delay. Oh, and the noise. Noise really screws things up, too.”
Okay, it was me, but to be fair, I am old, and I’ve made the above case many times (sometimes in the same workshop) with nobody standing up to prove me wrong.
So, what do we do about it? Well, a proactive application of Bode’s Integral Theorem called PES Pareto allows us to isolate the sources of noise around the loop, back-filtering the closed-loop measured quantities into independent reference inputs. From there, their contributions can be forward filtered to any point in the loop providing a measure of which noise sources are most critical at different measurement points e.g., error signal or output signal. It has recently been used to estimate the measurement and process noise covariances for Kalman filter design, but it has historically been used to identify which noise source are the most critical to eliminate. The first half of this tutorial will teach how to use PES Pareto to dramatically improve the quantification of noise in a very practical way.
The second half of this tutorial is all about how we limit the effects of these input noises in a latency sensitive way. We will show why the choice of an anti-alias filter should never simply be left to the circuit engineer and how that should be tied into the choice of converter circuits and sample rates. Furthermore, a key subset of the signals with excess noise around the loop are sensor signals that have been modulated. Modulation is used in nature and engineering as a way to immunize systems from baseline noise.
The key to using a modulated signal is how we demodulate it, and it turns out that while our knowledge of control methods is very 21st century, our demodulation methods are stuck in the 1950s. There are plenty of brilliantly modern demodulation methods used in communication and storage systems, but these are for problems that are largely insensitive to latency, time-delay, and the resulting degradation of phase margin. We will show the benefits of adapting modern demodulation methods to low latency needs. Taken together, these two “forgotten signals” methods (PES Pareto for noise analysis and low latency, coherent digital demodulation) push back substantially on the noise and latency in sensor signals around the loop.
Networking and Career Development for Young Professionals
Special Sessions (part of the conference, open to all conference attendance)
1. The Water Power Technologies Office at the Department of Energy (DOE): interdisciplinary prize competition (Carrie Schmaus; US Department of Energy)
Presenter’s short bio:
Dr. Daniel Abramovitch
Danny Abramovitch earned degrees in Electrical Engineering from Clemson (BS) and Stanford (MS and Ph.D.), doing his doctoral work under the direction of Gene Franklin. Upon graduation, and after a brief stay at Ford Aerospace, he accepted a job at Hewlett-Packard Labs, working on control issues for optical and magnetic disk drives for 11 1/2 years. He moved to Agilent Laboratories shortly after the spinoff from Hewlett-Packard, where he has spent 19 years working on test and measurement systems. He is currently in Agilent’s Mass Spectrometry Division working on improved real-time computational architectures for Agilent’s mass spectrometers, including the new Ultivo Tandem Quad product.
Danny is a Fellow of the IEEE and was Vice Chair for Industry and Applications for the 2004 American Control Conference (ACC) in Boston. He was Vice Chair for Workshops at the 2006 ACC in Minneapolis and the 2022 ACC in Atlanta, for Special Sessions at the 2007 ACC in New York, and for Industry and Applications for the 2009 ACC in St. Louis. He was Program Chair for the 2013 ACC and was General Chair of the recent 2016 ACC in Boston. He has helped organize conference tutorial sessions on topics as varied as disk drives, atomic force microscopes, phase-locked loops, laser interferometry, and how business models and mechanics affect control design. He served as the Chair of the IEEE CSS History Committee from 2001 to 2010. Danny is credited with the original idea for the clocking mechanism behind the DVD+RW optical disk format and is co-inventor on the fundamental patent. He was on the team that prototyped Agilent’s first 40Gbps Bit Error Rate Tester (BERT) and was able to cite a Douglas Adams book in one of his patents relating to that device. Along with his co-author, Gene Franklin, he was awarded the 2003 IEEE Control Systems Magazine Outstanding Paper Award. His favorite paper remains the one prompted by a question from his then 3-year-old son, which showed that the outrigger was a feedback mechanism that predated the water clock by at least a 1000 years. He was a Keynote Lecturer at the 2015 Multi-Conference on Decision and Control in Sydney, Australia and a plenary speaker at the 2020 Conference on Control Technology and Applications (this conference). His recent work for Agilent was on future atomic force microscopes and high precision interferometers. His current work involves improving the real-time data collection and signal processing on Agilent’s Mass Spectrometers and is part of the team that created Agilent’s multi-award winning Ultivo Tandem Quad LC Mass Spectrometer. He is the holder of over 20 patents and has published nearly 50 reviewed technical papers.
Dr. YangQuan Chen
University of California, Merced
YangQuan Chen earned his Ph.D. from Nanyang Technological University, Singapore, in 1998. He had been a faculty of Electrical Engineering at Utah State University from 2000-12. He joined the School of Engineering, University of California, Merced in summer 2012 teaching “Mechatronics”, “Engineering Service Learning” and “Unmanned Aerial Systems” for undergraduates; “Fractional Order Mechanics”, “Nonlinear Controls” and “Advanced Controls: Optimality and Robustness” for graduates. His research interests include mechatronics for sustainability, cognitive process control, small multi-UAV based cooperative multi-spectral “personal remote sensing”, applied fractional calculus in controls, modeling and complex signal processing; distributed measurement and control of distributed parameter systems with mobile actuator and sensor networks. His most recent book is “Outliers in Control Engineering — Fractional Calculus Perspective”(280 pages with P. Domanski and M. Lawrynczuk, 2021 De Gruyter series in Fractional Calculus in Applied Sciences and Engineering). He is listed in Highly Cited Researchers by Clarivate Analytics in 2018, 2019, 2020 and 2021. He received Research of the Year awards from USU (12) and UCM (20).
University of California, Merced
Furkan Guc is a PhD student in Mechanical Engineering at the University of California, Merced, and working at the Mechatronics, Embedded Systems, and Automation Lab (MESA Lab). He received his master’s degree in Mechanical Engineering from Bilkent University, Turkey in 2020. His research interest includes Applied Fractional Calculus, Smart Control Engineering, Industrial AI, Control Performance Assessment, Smart Predictive Maintenance, Fault Diagnosis and Health Monitoring in Process Control.
Dr Paweł D. Domański
Warsaw University of Technology, Poland
Dr Paweł D. Domański (PhD, DSc) was born in Warsaw, Poland in 1967. He received his M.Sc. in 1991, Ph.D. in 1996 and D.Sc. in 2018 all in automatic control from Warsaw University of Technology, Faculty of Electronics and Information Technology. He works in the Institute of Control and Computational Engineering, Warsaw University of Technology from 1991. Apart from scientific research he participated in dozens of industrial implementations of APC and optimization in power and chemical industry. He is the author of more than 100 publications in books, journals and conferences. His most recent book is “Outliers in Control Engineering — Fractional Calculus Perspective”(280 pages with Y. Chen and M. Lawrynczuk, 2021 De Gruyter series in Fractional Calculus in Applied Sciences and Engineering). His main research interest is with industrial APC applications, control performance quality assessment and optimization.
Dr. Jairo Viola
University of California, Merced
Dr. Jairo Viola was an Electronic Engineer with a master’s degree in Electronics from Universidad Pontificia Bolivariana, Colombia. He earned his PhD on Mechanical Engineering, from the Mechatronics, Embedded Systems, and Automation Lab (MESA Lab) at the University of California, Merced in 2022. For five years, he worked as a professor in the faculties of Electronic Engineering and Computer Sciences at the Universidad Pontificia Bolivariana, Colombia, teaching Programming, Computer Architecture, and Operating Systems courses. His research topics are Process Control, Robotics, Artificial Intelligence, Machine Learning and Big Data, Edge Computing and Failure Detection for Industrial Processes, and Applied Fractional Calculus.
Dr. Jing Wang
North China University of Technology (NCUT), Beijing
Dr. Jing Wang, received the B.S. degree in industry automation and the Ph.D. degree in control theory and control engineering from Northeastern University, Shenyang, China, in 1994 and 1998, respectively. She has been a Professor in the Department of Automation, School of Electrical and Control Engineering, North China University of Technology (NCUT), Beijing, China since 2020. She was a Professor with the College of Information Science and Technology, Beijing University of Chemical Technology, Beijing, China. She was a Visiting Professor with the University of Delaware, Newark, DE, USA, in 2014. Her current research interests include application of advanced control schemes to nonlinear, multivariable, constrained industrial processes; modeling, optimization, and control for complex industrial process; nonlinear model-based control of polymer microscopic quality in chemical reactor; and process monitoring and fault diagnosis for complex industrial process.