Prof. Dr. Sandro Wartzack

Chair of Engineering Design, Department of Mechanical Engineering

Research and validation of processes, methods and tools to support the design of medical technologies and interventions.

Research projects

  • Biomechanical Engineering
  • Musculoskeletal Modeling and Simulation
  • Development of Exoskeletons and Ortheses
  • Preoperative Planning of Endosurgeries
  • Analysis and Optimization of User-Product-Interaction

Current projects

  • Product development ontology supporting the provision of decision-relevant knowledge for multicriteria evaluation of decision alternatives

    (Third Party Funds Single)

    Term: 1. October 2022 - 30. April 2024
    Funding source: Deutsche Forschungsgemeinschaft (DFG)

    Due to the growing complexity of modern products, new challenges arisefor product developers, whose central task is to fulfill the requirementsdemanded by the market. These requirements concern specific properties, which haveto be fulfilled by the product and ensured through related productcharacteristics. Due to the increase of product complexity, various relationsbetween the characteristics and properties occur, which makes theirconsideration more difficult. Changes in product characteristics lead todifferent properties of the whole system, which in the case of complex relationnetworks result in unmanageable and often undesirable property changes. Theeffects of this include cost-intensive iterations for the adaptation to the demandedproperties or even the failure of products in the market. For this reason,these influences have to be identified already during development and then takeninto account during multi-criteria evaluation of decision alternatives. Forthis purpose, consistent decision models are required that contain theserelations. In interdisciplinary product development, these relations as well asthe relevant information and data originate from domain-specific models and areavailable as heterogeneous databases. This results in challenges regarding theprovision of decision-relevant knowledge for multi-criteria decision processesand models.

    Motivated by these challenges, the objective of the proposed researchproject is to develop fundamental findings about the formalization and provisionof context-specific knowledge in the fields of requirements-based, multi-criteriadecision making. Ontologies offer a suitable method for the formalization andprovision of context-specific knowledge, therefore, the main objective of theproject has to be achieved with a decision-supporting product developmentontology (PEO). Within the ontology development the own preliminary work fromthe first phase of the project (WA2913/33-1) has be considered, which providesthe necessary fundamentals regarding computer aided multi-criteria decisionmaking based on all product requirements. The extension of the PEO through dataand knowledge integration from heterogeneous sources enables product developersto provide relevant decision knowledge consistently and reduces the effort formanual provision. Verification of the developed PEO requires new approaches fordefining suitable queries. In summary, the research project focuses on scientificfindings regarding the applicability and usability of ontologies for an efficientsupport of multi-criteria decision making.

  • Methodik zur Auslegung passiver, strukturoptimierter Orthesen für die Behandlung bzw. Kompensation pathophysiologischer Bewegungsmuster anhand muskuloskelettaler Menschmodelle

    (Third Party Funds Single)

    Term: 1. October 2021 - 30. September 2024
    Funding source: DFG-Einzelförderung / Sachbeihilfe (EIN-SBH)

    Ankle joint orthoses are the most common rehabilitation device used to treat pathophysiological gait patterns such as foot drop syndrome. For this purpose, passive orthoses are preferably used which enable a supporting or stabilizing function or even the support of a rotational direction of the joint with spring-damper systems. In the case of extensive paresis or paralysis of the lower leg musculature, active orthoses can be used to restore a healthy gait. With the appropriate actuators, these can support both directions of rotation of the joint. The use of active orthoses, however, is accompanied by many problems, such as their high weight, the less integral design and the complex control and power supply of the actuators. For this reason, a methodology is to be developed within the framework of the research project with which the design of passive ankle joint orthoses can be realized to support both directions of joint rotation. For this purpose, the orthosis is to be provided with suitable fibre-reinforced plastic lightweight structures which, in the event of elastic deformation, exhibit the structural response corresponding to the required support of the pathophysiological gait. The core objective of the research project is the development of a method which enables the structures to be designed in such a way that their structural response is in symbiosis with the patient's gait behaviour and thus enables the best possible treatment or compensation of the patient's disease. To this end, a coupling between an FE model for structural optimization and a musculoskeletal human model for mapping the physiological and pathophysiological gait is to be introduced. The orthosis design optimized with the help of the developed methodology is finally manufactured as a prototype and evaluated on the basis of a test set-up before the orthosis can be tested in practical trials in conjunction with patients.

  • Development of a methodology for plausibility checks for linear structural mechanic finite element simulations using Deep Learning

    (Third Party Funds Single)

    Term: 1. August 2021 - 31. July 2024
    Funding source: Deutsche Forschungsgemeinschaft (DFG)

    Inthe current industrial environment, design accompanying linear finite elementsimulations are often carried out by product developers and not exclusively bycalculation engineers with several years of professional experience. This leadsto frequent iterations in the product development process and can lead toincorrect decisions based on insufficiently validated results. An automaticplausibility check for linear structural-mechanical FE simulations is animportant method to support product developers. The use of Convolutional NeuralNetworks (CNN) and Machine Learning methods represents an enormous potential toidentify correlations in data and to build a model with high prediction quality.In the applicant's preparatory work it could be shown that a plausibility checkfor FE calculations using Deep Learning and CNNs is possible. However, it isnecessary to increase the prediction quality of the artificial neural networkby adjusting the parameters and to demonstrate the application of the method tonew simulations. Furthermore, local areas of the FE-simulation shall beinvestigated, especially to detect numerical errors like singularities.


    Theaim of the project is to create a method for plausibility checks of similarlinear structural-mechanical FE simulations based on the preliminary work onthe projection method and singularity recognition. Furthermore, the networkparameters of different CNNs and machine learning methods are to be optimizedin order to implement a plausibility check with high prediction quality.FE-simulation results cannot be directly transferred to a neural network ormachine learning algorithm, but have to be converted to a uniformcomputer-processable form. Within the framework of the research project, theprojection method developed in preliminary work is applied, which usesspherical detector surfaces to transform arbitrary simulations into matriceswith uniform size. The generated matrices contain all relevant information toclassify a simulation as plausible or implausible. A Deep Learning CNN or SVMdoes the classification. In addition to the classification of the entiresimulation, local areas should also be examined. Especially singularities inFE-simulations shall be detected and accordingly give feedback to the user.


    Withan automatic plausibility check, errors in the simulation setup can be detectedautomatically at an early stage. It therefore represents an enormous potentialfor increasing the simulation quality in virtual product development.Especially if FE simulations are performed by product developers who have lesssimulation knowledge than experienced calculation engineers. A method will bedeveloped which allows to consider similar geometries and simulation boundaryconditions of linear FE-simulations.

  • Bereinigung multimodaler Bewegungsmessdaten mittels individualisierter muskuloskelettaler Menschmodelle

    (Third Party Funds Group – Sub project)

    Overall project: Empathokinästhetische Sensorik - Sensortechniken und Datenanalyseverfahren zur empathokinästhetischen Modellbildung und Zustandsbestimmung
    Term: 1. July 2021 - 30. June 2025
    Funding source: DFG / Sonderforschungsbereich (SFB)

    Es wird eine neuartige Methode zur Bereinigung und Filterung multimodaler Bewegungsmessdaten in Verbindung mit einer entsprechend notwendigen Multidomänenmodellierung muskuloskelettaler Ganzkörpermenschmodelle erforscht. Aufbauend auf den individualisierten Menschmodellen werden kinematische bzw. dynamische Trackingalgorithmen untersucht, die auf unterschiedliche Kombinationen multimodaler Messdaten angewendet werden können. Außerdem werden die neuen Methoden erprobt und evaluiert, sowie Benchmarks zu konventionellen Methoden durchgeführt. Neue EmpkinS-Messtechniken werden sobald verfügbar auf die neu etablierten Methoden transferiert.

  • Erforschung der posturalen Kontrolle basierend auf sensomotorisch erweiterten muskuloskelettalen Menschmodellen

    (Third Party Funds Group – Sub project)

    Overall project: Empathokinästhetische Sensorik
    Term: 1. July 2021 - 30. June 2025
    Funding source: DFG / Sonderforschungsbereich (SFB)

    A novel postural control model of walking is explored to characterise the components of dynamic balance control. For this purpose, clinically annotated gait movements are used as input data and muscle actuated multi-body models are extended by a sensorimotor level. Neuromotor and control model parameters of (patho-)physiological movement are identified with the help of machine learning methods. Technical and clinical validation of the models will be performed. New EmpkinS measurement techniques are to be transferred to the developed models as soon as possible.

Recent publications






Related Research Fields