Oct 2, 2014 - creased in renewable energy in recent years for a variety of ... imizing Control of Wave-Energy Con- verters,â by ... âCSS Newsâ announces the ...
»
about this issue
Renewable Energy and Optimization-Based Control
R
enewable energy is the energy that comes from resources that are replenished on a human time scale. Examples of renewable resources are sunlight, wind, rain, tides, waves,
Digital Object Identifier 10.1109/MCS.2014.2333211 Date of publication: 16 September 2014
and geothermal heat. Interest has increased in renewable energy in recent years for a variety of reasons, such as increasing oil prices, the accumulation of radioactive waste products from nuclear plants, and concerns over increasing carbon dioxide concentrations in
the atmosphere. The most established forms of renewable energy are from hydroelectricity and biomass, the latter which has been used for centuries in wood-burning stoves and more recently to generate liquid biofuels for transportation. The production of energy is
Contributors
John Ringwood competing at a World Cup Wildwater Race on the River Shannon as a member of the Irish National Team.
Francesco Fusco hiking in the Haleakala crater on the Hawaiian island of Maui.
Iakovos Michailidis.
Giorgio Bacelli.
Simone Baldi.
8 IEEE CONTROL SYSTEMS MAGAZINE » OCTOBER 2014
Elias B. Kosmatopoulos.
growing rapidly for many renewables, including wind, solar, and geothermal. One enticing source of energy is the ocean, where energy can be produced by waves, tides, salinity differences, or temperature differences. Wave energy is generated by the movement of water. The first feature article, “Energy-Maximizing Control of Wave-Energy Converters,” by John V. Ringwood, Giorgio Bacelli, and Francesco Fusco provides an overview of the motivation, background to, and state of the art in energy-maximizing control of waveenergy devices. The underpinning mathematical modeling is described and the control fundamentals established. Two example control schemes
are presented, which include an optimization-based control algorithm. The article also presents some algorithms for wave forecasting, which can be a necessary requirement, due to the acausal nature of some optimal control strategies. One of the control schemes is extended to show how cooperative control of devices in a wave farm can be beneficial. The article also includes perspectives on the interaction between control and the broader objectives of optimal wave-energy device geometry and full technoeco-
nomic optimization of wave-energy converters. The second feature article, “A ‘Plug and Play’ Computationally Efficient Approach for Control Design of Large-Scale Nonlinear Systems Using Cosimulation,” by Simone Baldi, Iakovos Michailidis, Elias B. Kosmatopoulos, and Petros A. Ioannou considers simulation-based control design, aka cosimulation. The controller optimizes a cost function related to system performance that involves an iterative process of system simulation and controller
Petros A. Ioannou.
Xiaoxiang Zhu.
Jared A. Frank working in the Mechatronics Laboratory at NYU Polytechnic School of Engineering.
(From left) Lily Gossage, Bei Lu, and Panadda Marayong with student Jasmine Barreras and counselor Debra Goodlow from the International Elementary School. (Photograph courtesy of Christina Yoon.)
Vikram Kapila at the 2014 NYU Polytechnic School of Engineering Research Expo.
OCTOBER 2014 « IEEE CONTROL SYSTEMS MAGAZINE 9
“Member Activities” describes a program organized by Bei Lu, Lily Gossage, and Panadda Marayong to introduce young women and their parents to robotics and control technology. redesign. Such control systems do not have any knowledge of the state equations for the model of the system, and a commonly applied algorithm for the design of such control systems is the sequential approach, in which the control parameters are determined by coupling any standard optimizer directly to the simulation code. The term “sequential” refers to the notion that the optimizer sends candidate control parameters to the simulation codes, which then send the control performance objectives back to the optimizer for choosing improved control parameters, with the iterations repeated until convergence. The calculations are typically sped up by employing automatic differentiation algorithms so that the first- and possibly second-order derivatives of the cost function with respect to the control parameters are supplied to the optimizer. This sequential approach is commonly used both in offline optimal control calculations and in real-time implementations of nonlinear model predictive control. The article proposes an alternative approach to cosimulation that does not require any derivative information by combining cognitive-based adaptive optimization with a convex approximation to the Hamilton-Jacobi-Bellman (HJB) equation. The cognitive-based adaptive optimization can compute control updates that are close to the gradient descent direction without requiring the derivatives. Unlike other alternatives, the convex approximation to the HJB equation means that the cosimulation approach does not require the online so-
lution of a nonconvex high-dimensional mathematical program. In his “President’s Message,” Jay Farrell discusses advanced sensor technologies, control ambassadors, and the “Impact on Control Technology” report, among other things. “CSS News” announces the members of the IEEE Control Systems Society (CSS)Board of Governors who will start three-year terms in 2015, provides an update of the progress of IEEE Transactions on Control of Network Systems, and reminds CSS members to nominate worthy colleagues for awards. “Letters” discusses control presentation. “Member Activities” describes a program organized by Bei Lu, Lily Gossage, and Panadda Marayong that introduces young women and their parents to robotics and control technology. “Technical Activities” discusses the CSS technical committees and control-related Wikipedia pages. “Publication Activities” discusses the CSS Publication Digest, the publication quality of CSS transactions, and ethics in publishing. “People in Control” has interviews with IEEE Fellows Edwin K.P. Chong and Dawn Tilbury. In “Focus on Education,” Xiaoxiang Zhu and Richard D. Braatz provide an introduction to data-based fault detection and identification, with the latter being the determination of system variables that are most closely associated with a detected fault. The contribution plot is the most commonly used tool in the process industries for quickly identifying the most affected variables.
10 IEEE CONTROL SYSTEMS MAGAZINE » OCTOBER 2014
Contribution calculations are revisited in the context of principal component analysis and multivariate statistics, and a two-dimensional contribution map is demonstrated for the examination of time series data under faulty conditions, to enable greater understanding of the fault and how its effects are propagated through the dynamical system. Another contribution to “Focus on Education” is provided by Jared Alan Frank and Vikram Kapila, who describe the development of mobile interfaces for interacting with automatic control experiments in the laboratory and remotely. Preliminary results are presented on the usability and user experience associated with the applications, as assessed by students. Implementation considerations such as application development, wireless communication, and hardware interfacing and control are discussed. A microcontroller-based interface to the laboratory hardware is shown to provide a low-cost solution that can monitor, command, and control experiments via mobile devices such as iPads, iPhones, or iPods. “Historical Perspectives” is the fourth in a series that commemorates the 60th anniversary of the CSS with reminiscences from past CSS presidents. This issue includes recollections from N. Harris McClamroch, who was the CSS president in 1998. Among the regular columns, “25 Years Ago” revisits a 1989 article that assessed air-to-air missile guidance and control technology. “Conference Calendar” lists upcoming conferences sponsored or cosponsored by CSS. “Book Announcements” provides summaries of books recently published in the control field. “Random Inputs” considers control-related techniques that are essentially equivalent but have different names. Richard D. Braatz
