He recently defended his doctoral thesis at Linköping University, titled “Efficient Fluid Structure Interaction Analysis in the Conceptual Design Phase: Industrially Proven and Validated Unique Methodology based on Aerospace Concepts.” Early in his career, he worked in the aerospace industry optimizing wing surfaces and flexible surfaces that interact with their surroundings and airflow. In that field, unexplained vibrations have caused, and still cause, major problems.
– Aircraft break apart in the air, fans explode, and pipes and hoses with internal and/or external axial flow begin to vibrate, resulting in cracks and failures. These phenomena are all examples of problems caused by aeroelastic, or more generally fluid-elastic, instability and response.
Seemingly random vibrations
In the years after working in the aerospace industry, Jari Hyvärinen worked with similar phenomena in several different fields, including the oil and truck industries. In the oil industry, they had problems with vibrations that seemed to arise randomly in gas pipelines.
– Phenomena involving “random” vibrations are all examples of problems caused by aeroelastic, or more generally fluid-elastic, instability and response. The purpose of this thesis is to demonstrate the opportunities that open up when a frequency-domain approach is used in studies of systems in which fluid-structure interaction, FSI, phenomena have a significant influence on behavior. In 2015, Jari Hyvärinen was recruited to Epiroc, where similar issues were relevant for their rock drilling equipment. In 2018, he was enrolled as a doctoral student at Linköping University to continue his earlier research, which had resulted in a licentiate degree, and to investigate vibration problems in rock drilling machines as part of his doctoral work.
– These were vibrations that suddenly appeared without anyone understanding why. The result was problems ranging from wear to hoses being completely torn loose. They claimed that the vibrations occurred without any visible influence. From my perspective, nothing happens without a reason; it is just a matter of understanding all dimensions of the system in order to trace the problems, he states.
Studied the behavior of hydraulic hoses
Fatigue failure in hydraulic hose systems, caused by severe vibrations, has become a critical factor creating operating and maintenance costs for end users of rock drilling equipment. Similar behavior also occurs in, for example, forestry machines.
– Hoses are used as parts of the energy supply system in machines used for mining and construction work. One purpose of my research was to create an understanding of the dynamic behavior of a selected hydraulic hose. The chosen numerical modeling method included a boundary element method in the fluid-elastic analysis of the dynamics of a pressurized hose with transport fluid. Experimental modal analysis was used to validate the numerical model. The analysis and experiments show that a complex coupling of purely structural bending modes arises when the hose is exposed to internal flow. In the research, we saw how these coupled modes become increasingly sensitive to external or internal excitation as flow velocity increases.
– These are complex relationships involving multiphysical phenomena that are linked together. What makes it so difficult to calculate is that, in order to understand the relationships, you need to be able to calculate everything from acoustics and fluids to mechanics and structural dynamics. What people believe are random vibrations occur because, in certain situations, several different parameters, such as flow velocity and temperature, interact and cause vibrations. This may happen at a specific point or within a certain range.
Developed a methodology
During the first half of the thesis work, Jari Hyvärinen developed a methodology that was then tested more experimentally during the second part.
– The first part of the thesis was completed a number of years ago, and I then tried to sell the method to industry. But since it is so complex and requires at least one year of training to use, it was difficult for companies to adopt it. Since then, I have used it in various failure investigations in which I have participated as an expert. It also formed the basis of the doctoral work I carried out at Linköping University, which resulted in a doctorate this spring.
Tested the methodology
The application studies included in the second part of the thesis involve failure investigations of a ventilation fan, a vehicle cooling fan and a hydraulic diesel engine torque converter. In the presented cases, the calculation results have either been validated against experiments, or the results from the calculations have been used as a decision basis for successfully updating a design.
– The presented examples clearly illustrate that the entire system envelope must be studied, not just individual conditions, in order to avoid unexpected failures. The reason is that instability problems can arise anywhere within the envelope. To efficiently search through the envelope, other methods and tools are needed than the ALE-CFD/FEM programs that have been available on the market for the past decade or so.
Investigated the influence of different parameters
To illustrate the strength of the method from the first part of the thesis, the second part of the work presents a parametric study of hose dynamics for hoses with typical dimensions used in industrial applications.
– Among other things, I investigated how different parameters affect the dynamic properties of hydraulic hoses and showed, for example, that the stiffness of the hose end support has a major impact on the stability and dynamic behavior of the hose. A soft support tends to create a static instability-like behavior, where the lowest frequency decreases to levels close to zero as flow velocity increases. Pre-tensioning the hose has a stabilizing effect on hose dynamics. In the case where the internal pressure of the hydraulic hose does not generate tension in the hose, increasing or decreasing the internal pressure has limited impact on hose dynamics; this is at least a conclusion that is valid within the investigated pressure range of 100–210 bar.
– In addition, a hose with a smaller diameter is more sensitive than one with a larger diameter, and this applies as long as the pre-tension is high enough to maintain static stability throughout the entire flow velocity range.
Machine learning can strengthen the methodology
In the second part of the thesis, Jari Hyvärinen has shown that the methodology he developed during the first part can be used very generally.
– The next step we have discussed at Linköping University is to connect the method with machine learning algorithms. One possible result could be that, in the future, you can ask a machine learning tool whether a system with certain conditions is likely to suffer from problematic vibration phenomena or not. That would be very helpful when designing, for example, a hydraulic system. Discussions are ongoing with Linköping University about whether he may help them build a research group in this area.
– But we are not there yet. Until then, I will continue working part-time at Epiroc and running my own consulting business, where I use my calculation method to help companies both understand failure processes and try to avoid them.
