Keywords: Numerical Simulation and Co-Simulation, Mechanics, Mechatronics, incl. Robotics, Electrical, Electronic and Power Systems
Abstract: This contribution presents a model for electronic circuits built with dielectric elastomers (DEs). The core of this kind of circuits is the dielectric elastomer transistor (DET), a highly sensitive piezo-resistive component that can be used, like a conventional transistor, to realize circuits. DETs can be directly integrated into more complex DE structures, as signal processing units, since they are made with the same technology and thus work under the same conditions. In this contribution, a Simulink model that can reproduce the electro-mechanical behavior of DE electronic circuits and predict their output is presented. The model takes into account how the mechanical and electrical properties of DETs influence one another and connects them to simulate complete circuits. The required system parameters, such as actuation and resistance change, are obtained from measurements using mathematical equations. The model is then assembled and tested. All the logic gates have been set up using DETs, and their physical behavior was experimentally investigated and compared to the simulations. It is shown that the model can accurately reproduce the logical behavior of the gates, and correctly takes into account mechanical characteristics such as viscoelasticity.

Keywords: Mechanics, Mechatronics, incl. Robotics, Modelling Uncertainties and Stochastic Systems
Abstract: In recent years, Dielectric Elastomer Transducers (DETs) have been of top-notch interest as an alternative solution to conventional mechatronic transduction systems, thanks to their features as low-cost and affordable materials, silence operation, low-power consumption, and high level of energy density. Generally, in their most uncomplicated layout, these devices form an electrostatic system, composed of a Dielectric Elastomer (DE) membrane, embedded between two opposite compliant electrodes, constituting a highly deformable capacitor capable of transforming electrical energy into mechanical and vice versa. However, DETs applicability is strongly affected by several engineer constraints. One of their principal failure modes is related to the electrical breakdown of the DE membrane, which occurs when an applied input electrical load exceeds the dielectric strength of the DE. In order to address this problem, the materials and the input load conditions must be chosen appropriately to assure a desired lifetime of operation. For this purpose, this work proposes a preliminary static reliability assessment procedure to evaluate the failure probability of a DET for static events as the electrical breakdown, with specific electric input load conditions. The resulting reliability model comprises the stochastic comparison of the dielectric strength of the DE material and the extreme values distribution of electrical loads in a specific period, and the forecast reliability evaluation for a more extended period, multiple of the original one.

Keywords: Mechanics, Mechatronics, incl. Robotics, Model Reduction, Model Simplification and Optimization, Computing Systems, Discrete and Discrete-Event Systems, incl. Discretisation of Continuous Systems, Petrinets
Abstract: Shape Memory Alloys (SMAs) are a class of smart materials which exhibit different thermo-mechanical properties than conventional metals. When SMA material is heated, transformations in the crystal lattice structure are induced, which generate a macroscopic change in shape on the order of 4-8%. This effect can be exploited for the development of novel actuators that react to an external thermal input with a mechanical deformation. In most applications, SMA material is shaped as a thin wire. In this way, the thermal activation can be simply induced via an electric current, thus resulting in a mechatronic tendon-like actuator. Despite their remarkable benefits such as compactness, lightweight, and high energy density, SMA materials are characterized by a highly nonlinear response, which is mainly due to a load-, temperature-, and rate-dependent hysteresis. Accurate modeling and compensation of such hysteresis is fundamental for the development of high-performance SMA applications. The goal of this work is to provide an accurate and numerically efficient model, which can be used to perform accurate simulations, model-based design optimization, and control of complex structures driven by polycrystalline SMAs. Our approach is based on a reformulation of the physics-based Müller-Achenbach-Seelecke (MAS) model for polycrystalline SMA wires within a hybrid dynamical framework. In this way, we are able to significantly reduce the numerical complexity and computation time, without losing numerical accuracy and physical interpretability. In future research, the model will be used for hybrid control of SMA systems.

Keywords: Numerical Simulation and Co-Simulation, Mechanics, Mechatronics, incl. Robotics, Medicine, Physiology and Biology
Abstract: In this paper, we propose a path-planning problem for stereotactic neurosurgery using concentric tube robots. The main goal is to reach a given region of interest inside the brain, e.g. a tumor, starting from a feasible point on the skull with an ideally short path avoiding certain sensitive brain areas. To describe the shape of the entire cannula from an entry point to the point of interest we use an existing mechanical model for continuum robots. We show numerically that our approach enables the surgeon to reach areas within the brain that would be impossible with a straight cannula as it is currently state of the art.

Keywords: Model Calibration, Model Validation and Verification, Design of Experiments, Search Based Testing, Medicine, Physiology and Biology, Fitting Models to Real Processes, incl. Identification and Calibration
Abstract: Musculoskeletal simulations are widely used to increase our insight in healthy and pathological. Typically, a generic musculoskeletal model is scaled to a participant and afterwards employed to calculate joint angles and estimate musculoskeletal loadings. This approach, however, neglects subject-specific musculoskeletal geometry. At the femur the neck-shaft angle (NSA) and femoral anteversion angle (AVA) are the most important anatomical features. Modifying the NSA and AVA affect hip joint contact forces. Furthermore, personalizing the AVA has been shown to increase the accuracy of hip joint contact force calculations. The impact of personalized femoral geometry on muscle activations and forces has not been assessed yet and therefore was the aim of the current study. We hypothesized that modifying the femoral geometry will alter muscle activations and forces. Furthermore, we assumed that a personalized femoral geometry would improve the agreement between the muscle activations obtained from the simulations and the experimentally measured electromyography (EMG) signals. Three-dimensional motion capture data and EMG data of lower limb muscles were collected. Magnetic resonance images (MRI) of each femur were segmented and used to calculate the subject-specific NSA and AVA. We created a reference model based on the NSA and AVA obtained from the MRI images and created additional models with altered NSA and AVA. Joint kinematics, joint kinetics, muscle activations, muscle forces and joint contact forces were calculated for each model using OpenSim. We compared muscle activations and forces between the different models. Furthermore, we compared the EMG data with the activations from the simulations and quantified how much hip, knee and ankle joint contact forces differ between models. We showed that the femoral geometry affects muscle and joint contact forces at all joints. Personalizing the geometry did not improve the agreement between EMG and muscle activation from our simulations. More comprehensive studies are needed to evaluate if the personalized femoral geometry can improve the accuracy of muscle force calculations in musculoskeletal simulations.

Keywords: Automation of Modelling and Software tools, incl. Computer Modelling, Machine Learning, Deep Learning, Data Analytics, Big Data, AI for Modelling
Abstract: A widely accepted notion of football matches in performance analysis (PA) is to consider them as dynamic interaction processes with emerging behaviours (Gréhaigne et al., 1997; McGarry et al., 2014). The description and analysis of these processes requires specific methods. Recurrence analysis is a technique to examine the recurrence behaviour of a system, as in when, how often and how close its trajectory in a phase space returns to a previous state. The aim of the study is to apply recurrence analysis to football matches in order to design an according analysis tool. Positional data from 21 football matches of a German Bundesliga team were examined. The phase space was made up of the field players’ xy-positions at each second of the match. For each pair of seconds, the average distance of all players was calculated. Recurrence plots (RPs) were obtained by colour-coding these distances. With a recurrence threshold of 9 meters and a minimum line length of 3 seconds, general recurrence parameters were calculated to characterize the individual recurrence behaviours of each match. Three football-specific recurrence parameters were defined to represent recurrence properties of open play. RPs showed commonalities (typical features indicating set plays and continuous gameplay) as well as unique structures during each match (number, distribution, and sequence of typical features). By implementing an interactive RP analysis tool, there is great potential for recurrence analysis to improve performance analysis in football practice.

Keywords: Medicine, Physiology and Biology, Machine Learning, Deep Learning, Data Analytics, Big Data, AI for Modelling
Abstract: Markov chains are a well-known tool in mathematical modelling. In performance analysis, finite Markov chains and the connected transition probabilities can be used to evaluate game behavior. Herein finite Markov chains are especially suited in the modelling of net games like tennis, due to their structure as an alternating sequence of discrete strokes. Finite Markov chain modelling thereby can be used as a descriptive tool, as well as to gain insight in the relationship between sports behavior and outcomes. The transition matrix can be used to display the frequency of game actions in relation to the respective game structure. Likewise, the model calculations from the theory of Markov chains enable determining the relevance of actions therein, concerning overall performance. To permit the usage of finite Markov chain modelling, adherence to the Markov property, often described as the property of “memorylessness”, must be assured. Present study implemented finite Markov chain modelling for elite level tennis at the Australian and French Open. The goal of the present study was to verify the usage of Markov chain modelling in tennis, using a newly designed transition matrix. Furthermore, the aim was to gain insight in the game structure of tennis. Results showed that the new model adhered to the Markov property in an extent, which allowed the usage of selected model predictions for the analysis. Further the analysis revealed insights in the current game structure of tennis, as well as connected influence of the factors court surface and gender of players.

Keywords: Machine Learning, Deep Learning, Data Analytics, Big Data, AI for Modelling
Abstract: Artificial neural networks (ANN) are able to detect patterns in sports games. The aim of the present study is to predict the setting speed and its target area within the complex I in volleyball using ANN. For the analysis, 289 rallies from one setter of the 2nd Austrian Volleyball League Women were considered. Player positions and ball trajectory data prior to the setter’s pass were recorded. Subsequently, ANN software NeuroDimension was trained with 60% of the datasets to predict the target area and setting speed (supervised learning). The rest was used for cross-validation (15%) and predictions (25%). The accordance of the predicted and the real values was assessed by the percentage of correct predictions. To rate the prediction quality, the results of the ANN were compared with predictions from high-level volleyball coaches. The ANN correctly predicted the target area in 68.1% and the setting speed in 79.2%, which was significantly higher than by chance (p<0.01). The accuracy of the ANN-predicted target areas was 2.8% higher than from coaches (not significant). The ANN’s prediction rate for the setting speed was significantly higher by 14.6%. Information from ANN predictions could be helpful to support coaches to train the athlete’s skills to anticipate opponents’ playing actions and to train the setter of the own team.

Keywords: Medicine, Physiology and Biology, Machine Learning, Deep Learning, Data Analytics, Big Data, AI for Modelling
Abstract: The present study applied social networks analysis to objectively discriminate and describe interpersonal interaction dynamics of players coached by the top-level professionals Jürgen Klopp, Pep Guardiola, and Mauricio Pochettino, across the UEFA Champions League seasons from 2017 to 2020, considering different quality of opposition. In total, passing data on 92 games of the UEFA Champions League were analysed, according to 2 different qualities of opposition: strong and weak opponents. Statistical analysis revealed that two of Jürgen Klopp’s team passing network metrics differ according to the quality of opposition. Density (U(34) = 202; p = 0.02) and largest eigenvalue (t(32) = -3.24; p = 0.03) were lower when Liverpool played against strong opponents. Pochettino also showed higher values for density (t(26) = -2.77; p = 0.01) and largest eigenvalue when playing against strong opponents (U(28) = 148; p = 0.01), compared to when playing against weak opponents. Additionally, the average shortest-path length was significantly lower when playing against strong opponents (U(28) = 148; p = 0.01). Guardiola’s network metrics were not statistically different according to the quality of opposition in the analysed matches. Klopp and Pochettino, which stand out for more the flexibility of interpersonal linkages synergies in their passing dynamics, presented lower density and largest eigenvalue when playing against strong opponents. Thus, they are impacted negatively in the structural cohesion inside their teams, as well as the network strength and susceptibility to errors. Guardiola is able to maintain his footprint which is related to integrated and coordinated connection between groups of players, thus keeping the relevant players connecting the attacking plays, regardless of the quality of opposition. This is novel evidence on sports teams’ coordination and cooperation relationships through passing in football association, along seasons of a high-level European competition.

Keywords: Fluidics and Thermodynamics, Numerical Simulation and Co-Simulation, Bio-Technical, Bio-Chemical and Chemical Engineering Processes
Abstract: Biomass gasification and subsequently fuel and chemicals synthesis allow for a seamless transition from fossil raw materials to a renewable economy. A computational particle fluid dynamics (CPFD) simulation model is developed for biomass gasification in a bubbling fluidized bed reactor with steam as the fluidizing gas. CPFD methodology incorporates the multiphase particle-in-cell (MP-PIC) method and the particle parceling algorithm to solve the gas phase as a continuous fluid in an Eulerian grid of cells where particles are modeled as discrete Lagrangian points. The model has been validated by experimental results and shows good agreement. Reactor temperature effects as a key process operating parameters have been investigated and the results of the product gas composition on the temperature of 950, 1000, 1050, 1100, 1150, and 1200 K are shown. The results illustrate that at a higher temperature, the product of CO and H2 is promoted whereas the generation of CO2 and CH4 is diminished.

Keywords: Bio-Technical, Bio-Chemical and Chemical Engineering Processes, ODE, DAE, SODE, SDAE Systems, Numerical Simulation and Co-Simulation
Abstract: In this paper, a model of ex-situ biological biogas upgrading process is first developed. Then, parameter sensitivity analysis is performed, in order to determine the most relevant parameters for subsequent parameter identification, based on a linear fractional transformation (LFT) reformulation of the model. Biogas composition and volumetric production have been well predicted by the calibrated model, allowing its adoption as a designing tool for start-up operation of experimental pilot-scale activity.

Keywords: Modelling in Manufacturing and Process Engineering, Fitting Models to Real Processes, incl. Identification and Calibration, Model Reduction, Model Simplification and Optimization
Abstract: In the process industry, where several products are produced in different batches, the optimal plant operation mode and control to meet the quality requirements have a high priority. In order to be able to perform the corresponding analysis for verification, a reliable model is required to represent the process accurately. In practical applications, linear models can usually be determined using step tests at operating conditions of the process, which are not always known. Furthermore, performing step tests can be challenging due to operational constraints. Thus, nonlinear system identification approaches offer possibilities to obtain rigorous models on a data-driven basis assuming that the functional structures for the process model are known. In order to be able to perform nonlinear system identification without a predefined functional structure, sparse regression with L_1 regularisation offers the advantage of simultaneously performing functional structure identification and parameter estimation in open-loop cases. This paper seeks to provide a nonlinear system identification approach for closed-loop systems containing a proportional-integral (PI) controller using sparse regression. For this purpose, a nonlinear dictionary function is constructed containing all possible nonlinear bijective function candidates representing both the process and controller dynamics. The regularisation property performs a feature selection and provides the most suitable functions with the related weighting parameters for the final model. The approach is tested on a continuous stirred-tank reactor model (CSTR) and it is demonstrated that the controller and process dynamics are identified accurately with R^2 = 1. Furthermore, it can be shown that normalising the dictionary function promotes sparsity.

Keywords: Bio-Technical, Bio-Chemical and Chemical Engineering Processes, Medicine, Physiology and Biology, Modelling Uncertainties and Stochastic Systems
Abstract: In this work, we present an optimization-based design strategy for gene regulatory networks (GRNs) in the stochastic regime (i.e., in the presence of molecular noise). The approach exploits a recently developed framework for the efficient simulation of stochastic GRNs based on a Partial Integro Differential Equations (PIDE) model formulation, which is here further accelerated with a parallel implementation in GPUs to maximize the performance. The simulator is combined with a global Mixed Integer Nonlinear Programming algorithm to efficiently address the optimization of the design through topology and parameter spaces simultaneously. We illustrate the performance of the methododology proposed through two different case studies: a biocircuit with a pre-defined target dynamics, and a biocircuit with a stationary bi-modal distribution fulfilling a number of requirements (in terms of distance and ratios of probabilities between modes).

Keywords: Bio-Technical, Bio-Chemical and Chemical Engineering Processes, Model Reduction, Model Simplification and Optimization, ODE, DAE, SODE, SDAE Systems
Abstract: The necessity to focus and work with renewable energy for value-added product generation has gained interest in recent years, which has led to the development of mathematical models that allow a better understanding and optimization of these processes. In this work an extension of the anaerobic digestion model (ADM No. 1) with H2 and CO external gas injection was proposed. Therefore, the modification of the volumetric mass transfer coefficient in terms of gas injection and the use of CO as a substrate of the process were proposed. Then, a model reduction was performed applying the principal process analysis (PPA) methodology with two threshold values δ=0.05 and δ=0.1. The R^2, AICc criterium, and Global Relative Error (%Error) were used to compare the model and reductions performance. The threshold value δ=0.05 presented the best results with an R^2 > 0.99 and AICc criterium of -114 compared to the experimental process. For the %Error, values of 2.32%, 1.38%, and 2.18% were achieved for H2, CH4, and CO outlet gas flowrates when the reduction δ=0.05 is compared with the complete model. This reduction also allowed to decrease the simulation time from 1.94s to 0.82s. Thus, concluding that a first reduced model approximation is possible for the biomethanation process.

Keywords: Medicine, Physiology and Biology, Bio-Technical, Bio-Chemical and Chemical Engineering Processes, Machine Learning, Deep Learning, Data Analytics, Big Data, AI for Modelling
Abstract: Mathematical modeling of biological systems are useful in (i) gaining better understanding about the physiological processes in an organism, (ii) simulating alternative scenarios, (iii) finding targets for improved performance within metabolic engineering context (iv) performing several functional analyses, e.g. identify drug targets (v) process scheduling within the context of industrial biotechnology etc. Increasing the predictive capability of these models is of common interest within systems biology studies which allows identification of more effective and personalized treatment strategies for complex metabolic diseases such as cancer by investigation of disease metabolism and providing correct subtyping and staging. By transforming gene-level information to flux/metabolite level information, current disease state can be analyzed and diagnosis of cancer subtype can be performed using a less invasive methods.

In this study, subtyping and staging of lung cancer, that is one of the main causes of cancer related deaths, was performed by integrating publicly available RNAseq data of normal, lung adenoma and adenocarcinomas and lung squamous cell neoplasms to human genome scale metabolic model and classification of obtained flux distributions using linear support vector machine (SVM) classifications. Differential flux analysis and pathway enrichment methods showed that model adequately represented tumour metabolism. SVM classification accuracies were calculated as more than 99% for normal and cancer cells and 94% for adenomas and adenocarcinomas and squamous cell neoplasms, indicating high predictive capability of flux distributions.

Keywords: Medicine, Physiology and Biology, Fitting Models to Real Processes, incl. Identification and Calibration, ODE, DAE, SODE, SDAE Systems
Abstract: In this work, a novel insulin-glucagon-glucose model is proposed, where the glucagon effect on the endogenous glucose production (EGP) is described by the dynamics of the glucagon receptors. In order to assess the quality of the model, its parameters are fitted in such a way that the influence of glucagon on EGP is isolated. Experimental data is used to validate the model structure and show that the receptor dynamics allow to explain some of the glucagon-related phenomena observed in the clinical data. This physiology-focused model will be useful in the development of artificial pancreas algorithms both for more realistic in silico validations and in the development of model-based control strategies.

Keywords: Numerical Simulation and Co-Simulation, Medicine, Physiology and Biology, Automation of Modelling and Software tools, incl. Computer Modelling
Abstract: Utilizing machine learning techniques for classification of tumorous tissue has been shown to be extremely effective for external imaging and processing. However, when training an algorithm to work intraoperatively for identifying the extent of cancerous tissue within the body using intraoperative sensors, an adequate data set is not readily available or in many cases not feasible to obtain. For this reason, accurate and realistic simulations of cancer growth can be a powerful tool in supplementing datasets where it is not realistic to directly obtain sensor measurements in the body. This work introduces a simulation environment to generate cell-based cancerous tumors grown within a controlled oxygen environment. The growth and oxygen simulations are outlined along with how they are implemented within an open source simulation environment. Simulation results are provided showing the influence of the oxygen field on tumor growth shape and rate. The results follow intuitive growth characteristics and exhibit invasive behavior from lack of oxygen. The simulation environment lends itself to be easily and flexibly used for further detailed investigations of tumor growth.

Keywords: ODE, DAE, SODE, SDAE Systems, Bio-Technical, Bio-Chemical and Chemical Engineering Processes
Abstract: The removal of planktivorous fish results in ecosystem changes that increase water transparency and decrease the risk of harmful algal blooms. In many situations, complete eradication of fish is necessary to improve water quality, which can also protect natural populations. In this work, we construct a generic fish-zooplankton (FZ) model and we determine conditions to eradicate the fish population through repeated fish removal. The model accounts for the weight and age of fish, and is described by a system of ordinary differential equations with impulsive effects representing the reproduction process and fish removal. We demonstrate that the survival and extinction of fish can be determined from the roots of a polynomial, whose coefficients depend on maximal fecundity, mortality rates, and catching effort. If all roots of the polynomial lie within the unit circle, then eradication is ensured. To illustrate our results, we show the importance of removing young fish, and not only adults, to achieve eradication.

Keywords: Modelling for Control and Real-Time Applications, ODE, DAE, SODE, SDAE Systems, Medicine, Physiology and Biology
Abstract: In this article, we intend to explore the role of using an "infinite time horizon" framework to address the issues of sustainability and long-term strategies in the control of biological processes. We use two case study models to explain why considering a fixed or moving endpoint does not lead to the desired long-term effects. The first biological model considered concerns the spread of an infectious disease and its treatment as an infinite horizon optimal control problem. The second one deals with the metronomic chemotherapy cancer treatment over the remaining lifetime horizon of the patient. The latter is consistent with the conception of cancer as a chronic disease. Both models show structural differences in the choice of the objective functional, the first one uses a stabilization functional containing a weight function, the second one contains a damage functional which involves a density function.

Keywords: Financial, Economic and Macroeconomic Systems
Abstract: It is nowadays well-known that the classical deterministic mathematical tools, based on causality principles, generally fail when dealing with the complexity features of behavioral systems. In this talk, I want to address some of these complexity features, i.e. rationality vs. bounded rationality, homogeneity vs. heterogeneity, equilibrium vs. out-of-equilibrium and linearity vs. non-linearity. I will present a toy model to explore the impact of human altruistic vs. selfish aptitudes on the asymptotic wealth distribution of a simulated simple society, as an example to address the aforementioned features using the tools of the kinetic theory of active particles. Finally, I will change the prospective towards the empirical one, presenting an example on detecting lottery-type preferences of investors using market data from the New York Stock Exchange, by means of capital asset pricing models. In this case the emphasis is mostly placed on the role of computation, discussing some aspects related to classical regression vs. machine learning techniques.