1,721,175 research outputs found
Model reduction of switched affine systems
Full text linkThis paper addresses model reduction and extends balanced truncation to the class of switched affine systems with endogenous switching. The switched affine system is rewritten as a switched linear one with state resets that account for the affine terms.
Balanced truncation can then be applied to each mode dynamics, independently. As a result, different reduced state vectors are associated with the different modes, and reset maps are here appropriately redefined so as to account and compensate for this mismatch, possibly preserving the continuity of the output. The overall behavior of the reduced switched system is determined by both the selected reduction per mode and the adopted reset maps. In this paper, we consider a stochastic setting and propose a randomized method for the selection of the reduced order. The performance of the proposed approach is illustrated through a multi-room temperature control example
A model partitioning method based on dynamic decoupling for the efficient simulation of multibody systems
No full textThe presence of different time scales in a dynamic model significantly hampers the efficiency of its simulation. In multibody systems the fact is particularly relevant, as the mentioned time scales may be very different, due for example to the coexistence of mechanical components controlled by electronic drive units, and may also appear in conjunction with significant nonlinearities. This paper proposes a systematic technique, based on the principles of dynamic decoupling, to partition a model based on the time scales that are relevant for the particular simulation studies to be performed, and as transparently as possible for the user. In accordance with said purpose, peculiar to the technique is its neat separation in two parts: a structural analysis of the model, that is general with respect to any possible simulation scenario, and a subsequent decoupled integration, which can conversely be (easily) tailored to the study at hand. Also, since the technique does not aim at reducing but rather at partitioning the model, the state space and the physical interpretation of the dynamic variables are inherently preserved. Moreover, the proposed analysis allows to define some novel indices relative to the separability of the system, thereby extending the idea of “stiffness” in a way that is particularly keen to its use for the improvement of simulation efficiency, be the envisaged integration scheme monolithic, parallel, or even based on co-simulation. Finally, thanks to the way the analysis phase is conceived, the technique is naturally applicable to both linear and nonlinear models. The paper contains a methodological presentation of the proposed technique, that is related to alternatives available in the literature so as to evidence the peculiarities just sketched, and some application examples, illustrating the achieved advantages and motivating the major design choice from an operational viewpoint
Control and Design of Computing Systems: What to Model and How
No full textThe application of feedback control to computing systems is a promising research area, but has to date been hindered by the almost unanimously perceived complexity in creating control-oriented system models. Computing systems are in fact considered very hard to describe with dynamic models allowing for simple and powerful control design tools, so that complex ones need bringing in to the detriment of efficiency and result assessment. In this work a novel approach to the modelling of computing systems is proposed, in a view to explain and partially avoid such complexity, by capturing only their relevant dynamics with the simplest possible models. The approach is shown to work at least on two relevant case studies, so that a significant generality can be inferred from it, being able to reproduce the relevant parts of the system’s behaviour and paving the way to control design and synthesis
Antiwindup-aware PI Autotuning
No full textThis manuscript addresses the problem of “tuning the antiwindup mechanism” of PI(D) controllers, an issue seldom addressed in the literature but yielding improvement and reducing undesired large-transient behaviours if tackled correctly. A previously proposed framework is extended and related to both alternative antiwindup implementations and tuning methods, resulting in more methodological insight on the matter, and guidelines to set up the extension of a given method to include antiwindup. Simulation examples are reported to illustrate the achievable advantages
Automating efficiency-targeted approximations in modelling and simulation tools: Dynamic decoupling and mixed-mode integration
No full textModelling and simulation nowadays permeate virtually any engineering activity, requiring tools capable of managing complex models efficiently. Nonetheless, whereas modern modelling languages and tools allow to construct such models even on lightweight platforms (e.g., a laptop), the same is not true when it comes to numerically integrate those models. For the latter purpose, modellers usually pursue efficiency by resorting to approximation and reduction techniques. However, such techniques are unnatural to include in modelling and simulation tools. This is particularly true with object-oriented ones, which on the other hand are the most interesting for dealing with complexity from the model construction viewpoint. This paper presents a novel approximation technique that can be easily included in modelling and simulation tools, and relates the proposal to literature alternatives so as to evidence its peculiarities. An extended manipulation toolchain is also proposed, allowing for the introduction of other (classical) efficiency-targeted approximation techniques, within a unified framework. Some application examples illustrate the achieved advantages and motivate the major design choices from an operational viewpoint
Model Separability Indices for Efficient Dynamic Simulation
No full textThis paper proposes a means to quantify the aptitude of a dynamic model to be partitioned into weakly coupled submodels. The proposal applies to both linear and nonlinear systems, is largely independent of their scale, and requires information that is easy to provide on the part of the analyst. The presented indices can be exploited in different manners, from easing numerical integration on monolithic solution platforms to tailoring distributed or co-simulation ones based on the particular characteristics of the model at hand. Examples are reported to show the usefulness and the practical viability of the proposal
Modelling and Control of Big Data Frameworks
No full textWe present a model library conceived to design and assess critical components of big data frameworks, with a control-centric approach. The library adopts the object-oriented paradigm, using the Modelica language. Continuous-time and algorithmic models can be mixed, allowing to represent control code with high fidelity, and to reduce the simulation effort to the minimum required. We discuss the used modelling principles, describe the library, and show some design examples
- …
