Andreas Johansson

About me

Research

Almost all my research efforts has been inspired by applications, mainly in the local process industry. Here follows a compilation of the areas where I have made contributions.

• Fault detection: Developing the concept of Dynamic threshold generator which is a dynamic system that generates a time-varying alarm threshold taking into account the uncertainties and the inputs of the process. Applications in pulverized coal injection, jet engines, flotation (supervising Michael Bask), and lambda processes (supervising Magnus Berndtsson)
• Fault detection: As co-supervisor of Soheil Salehpour, I have assisted in developing algorithms for change detection based on mixed integer linear programming and applied the result to the detection of faults in a blending process in the production of iron ore pellets.
• Fault detection: As co-supervisor of Saleh Sayyaddelshad, I have participated in developing design algorithms for observers for unknown input reconstruction in nonlinear systems
• Model validation: Developing a validation algorithm for systems with parameter uncertainty based on a Fréchet derivative approximation of the mapping from parameters to output.
• Nonlinear observers: An observer for systems with nonlinear output map that deals with systems that become nearly unobservable in parts of the state-space. Application to the LD converter
• Nonlinear observers: As co-supervisor of Saleh Sayyaddelshad, I have participated in finding LMI based methods for robust observer design for a class of nonlinear systems with time-delay
• Optimal control: Introducing the window norm to be used as the basis for optimal control
• Distributed parameter systems: Modeling, model reduction, and simulation of two industrial processes, the paper pulp digester (with Limei Ding) and the cone crusher
• Complex systems: Developing tools for using the signal-flow graph for visualizing complex industry plants. The signal-flow graph technique for visualization has been implemented by Wolfgang Birk and co-workers as a software application and applied to a stock preparation process
• Fail-bounded control systems: In cooperation with Jonny Vinter developing an integrator implementation that is robust to bit-flips
• Extremum seeking control: As supervisor of Khalid Atta, aiding in the development of an extremum seeking controller based on phasor estimation suitable for systems with large and time-varying phase lag. This algorithm has been applied to cone crushers and hydro-power turbines

Education

I have been teaching on many levels in the subject of Automatic Control, ranging from a basic course for students in a vocational college program to advanced courses on multivariable and nonlinear systems. I am currently examiner and teacher in two basic courses in the field.

R0001E Basic Automatic Control External link.. For students at several bachelor level programs, e.g. Automotive Engineering, Mechanical Engineering and Electrical Power Engineering.

R0002E Modelling and Control External link.. For students in the master level programs Sustainable Energy Engineering, Mechanical Engineering, and Industrial and Management Engineering

Both of these courses are given according to a flipped classroom concept with recorded lectures and learning activities where interaction between students as well as individual help and quick feedback from the teacher is in focus.

Some selected qualifications

• PhD in Automatic Control from LTU 2002
• Docent degree 2007
• Head of the division of Signals and Systems since 2011
• Main supervisor of Khalid Atta, who received his PhD in 2015. Co-supervisor for 3 PhD students who received PhD degrees and currently main supervisor of Riccardo Lucchese and co-supervisor of Sina Sharif Mansouri
• Supervisor for 10 degree projects in master level programs
• Erdös number External link. 4