Jan Bartolj, dipl. inž., dr. Ana Trajkovski, univ. dipl. inž., oba Univerza v Ljubljani, Fakulteta za strojništvo; mag. Anže Čelik, univ. dipl. inž., Poclain Hydraulics, d. o. o., Žiri; doc. dr. Franc Majdič, univ. dipl. inž., Univerza v Ljubljani, Fakulteta za strojništvo

Abstract:
When designing new products for production using additive manufacturing processes, developers have fewer constraints than when using traditional manufacturing processes such as CNC machining or die casting. SLM, or selective laser melting, is a method of metal 3D printing in which the laser beam is guided by mirrors to melt the metal powder and create a solid body of the final product. By using the laser beam as main energy source for manufacturing, we are talking about significant savings in energy consumption during the manufacturing process. However, one of the reasons why 3D printing has not been serious competitor to traditional manufacturing methods is the time it takes to produce the final products. But if metal 3D printing were to be used for larger batches of 1000 parts or more, the cost would decrease due to the mass savings, and with many positive aspects that this process offers, these parts would be more comparable to other products on the market.
We were faced with the challenge of developing a new 3D-printed hydraulic valve that would have better characteristics than conventionally manufactured valves, yet be cost effective to produce and sell on the market with ever increasing demands. The decision was made to use topology optimization based on a known internal geometry that was numerically tested.
First, we created the inner geometry of new valve as solid body. This geometry was to be inverted to create a hollow structure at the end. Then, we numerically tested the strength of a model using FEA analyses and flow characteristic with CFD analyses. We mainly wanted to test how the structure would handle the pressure and forces applied, since much less material is used compared to conventionally manufactured valves. We focused on the displacement of the leading edge (fig. 4) of the valve, which proved to be a critical point of numerical results. These displacements had negative values at some locations which should be avoided at all costs, as this could mean blocking of the spool in the housing, making the valve unusable. We reduced these negative values by increasing the wall thickness.
Numerically testing fluid flow through the housing (fig. 5), we simulated different scenarios based on varying inlet flow, spool displacement and temperature characteristics. We also compared numerical results of 3D-printed housing with conventionally manufactured one to ensure that pressure drop and other fluid-based properties were improved (fig. 6).
Using internal geometry and numerically applied loads as starting point for topology optimization, we created different types of meshes and decided on the best one based on the results. The optimal result had to give us the best stiffness-to-weight ratio. The result of topology-based optimization is a body consisting of small volumes derived from finite elements (fig. 8). Resulting geometry had to be parametrically shaped, giving us a model ready for printing (fig. 9). Printing was done on the EOS M290 metal 3D printer and was completed in 5 hours (fig. 10). The material chosen was stainless steel, which offers the best corrosion-resistant properties for use in water hydraulics. After printing, the housing was micro-pinned as primary way of surface. It was then CNC machined and fine grinded on leading edges to ensure tight fit between housing and spool.
Finally, we tested the valve using water as hydraulic fluid. We tested pressure drop (fig. 11) at different flow rates and spool displacements. We also measured Q-s characteristic (fig. 13) at different pressure drops with the load applied to the valve outlet port. When comparing the pressure characteristics of numerical and physical tests, we found a deviating difference (fig. 12). This is because in the numerical tests of the housing, we measured the parameters at the inlet and outlet of the valve, while in the real tests, we added the mounting plate, knee, and T-hydraulic connectors between valve and pressure transducers all adding up as resistance components. In future, we will need to either develop a mounting plate that has pressure sensors integrated close to the valve or to create a numerical model of all the interference factors.
All in all, the entire design process of this 3D-printed valve was a success, as we were able to reduce the mass of the final product by 63% compared to conventionally manufactured valves. In addition, all flowbased characteristics were also improved, such as pressure drop, which is 25 bar lower at a flow rate of 50 l/min, a reduction of 20%.

Keywords:
metal 3D printing, hydraulics, additive technologies, optimization, numerical analysis, topology optimization