Wind energy systems (WES) like wind turbines are designed to last at least 20 years. But, most of these wind giants fail before that term. The most common causes are failures of the electrical system. They can be repaired in just a few hours to some days. In contrast to this, gearbox defects can cause long and expensive downtime.. Generally, the gearbox has to be replaced, because, with hub heights of more than 100m, repair is not possible. In addition, lead times for new parts take up a considerable amount of the downtime.
Work on reducing the number of gear defects, as well as downtime reduction, is being undertaken by research scientists at the University of Bremen in Germany. The main objectives of the research are:
- Understanding the influences of the manufacturing process leading to geometric deviations at the final gear
- Analysis of the correlation between deviations of the gear geometry and defects within its lifecycle
- Analyzing deviations of the gear geometry at nearly every point of the work piece by an extended measurement and areal analysis
- Realizing a flexible and fast way to manufacture spare parts for broken gears
The first three points of the list prevent failures of gear-driven wind turbines by taking a closer look at the gears themselves. The last point tends to reduce both the costs of repairing a WES gearbox and the delivery time of spare parts by increasing the number of suppliers. In general, only the original manufacturer of the gear knows its design parameters. Alternatives to obtaining a spare part of a defective gear requires reverse engineering by the following steps:
1. Measurement of a damaged gear and calculation of the unknown design parameters
2. Analysis of the material and ordering blanks
3. Analysis of the heat treatment
4. Flexible rough manufacturing of the gear
5. Heat treatment process
6. Finishing of the hardened gear
This list includes two challenging tasks: the unknown gear measurement and the flexible rough manufacturing. In contrast to this, the analysis of the material and the heat treatment process is possible but not subject of this article.
Measurement of unknown gears
The Bremen Institute for Metrology, Automation and Quality Science (BIMAQ) at the University of Bremen commissioned a new 3-D coordinate measuring machine (CMM) in 2011. It is a portal-type CMM Leitz PMM-F 30.20.7 (see Figure 1) and is the heart of a new centre for measurement of large gears.
The Leitz CMM is located in a certified laboratory of air-condition category 2. All dimensional measurements are performed with the software Quindos 7.5. The maximum permissible errors of the CMM are specified according to ISO 10360 by MPEE = 1.3 µm + (L in mm/400 mm) µm, MPEP = 1.3 µm and MPETHP = 1.8 µm in 68 s.
The largest gears to be measured can have diameters up to 3 m (9.8 ft). The loading is performed by a high precision automatic delivery system. In contrast to the standard measurement of four teeth of a gear, the aim is an areal measurement of all gear flanks. This provides flank topographies with high resolution. Based on a 3-dimensional evaluation approach, these measurements enable statements about areal deviations. This is proved on gears with a diameter of 120 mm. They were measured and analyzed by a holistic approach developed by the BIMAQ (Figure 2). These evaluations extend the 2-dimensional values of a standard gear measurement. Additionally, the current development of e.g. optical sensors enabling fast areal measurements will raise the demand for areal evaluations and areal quality parameters in the near future.
A reverse engineering of unknown gears is possible with the software-module “gearxy” included in Quindos, for example. But, this demands a highly skilled user to be able to iteratively correct the geometric parameters. The aim is to extend the mathematical capabilities of the existing 3-dimensional evaluation to an automatic calculation of these parameters which will minimise operator influence.
Flexible manufacturing of single gears or small lots
Ordering a gear spare part can be a time-consuming task. If the design parameters of an involute gear are either known or reverse engineered, one or more gears can be manufactured. Single gears or small lots of gears demand a highly flexible manufacturing process. In mass production, the most common process is hobbing. But hobbing is cost-intensive and not flexible, as each hob is specialized to a certain gear design. This leads to the idea of rough milling involute gears with standard machine tools. For very large gears, some solutions have been already developed.
The flank surface of involute gears is mathematically describable. In general, the tool paths require a 5-axes milling machine. To develop this this research idea, the director of the BIMAQ, Prof. G. Goch, was given a Machine Tool Technologies Research Foundation Loan Award. The loaned machine tool is a Mori Seiki NMV 5000DCG. The machine supports 5-axes milling of work pieces with a diameter of up to 500 mm. Using this machine, gears were milled with the tool rolling on the flanks (Figure 3). The achieved quality class of this rough machining was between 5 and 8, but further optimizations seem to be possible.
In conclusion, the research has found that the main parts of the reverse engineering process are feasible. A software toolbox for areal gear measurements was developed and the first results of 5-axes milling small gears are promising. Both the machining time and the workpiece quality can be optimized by the tool type and the milling strategy. Some works with slot cutters were published in the last two years, e.g. by Sandvik Tooling, reducing the machining time to several minutes. This is possible by a rolling motion of the slot cutter on an involute path on the flank. The aim is to manufacture single gears or small lots with slot cutters in reasonable time and high quality classes.
Future research will concentrate on including tooth modifications in both the measurement algorithms and the tool path calculation. The combination of a sophisticated measurement strategy with a flexible machining process might be also capable of performing hard milling and, thus, to replace finishing by grinding.l
The large CMM PMMF 30.20.7 is an essential part of the research and was funded by the Federal Ministry of Environment, Nature Conservation and Nuclear Safety (BMU).
A part of the research described in this paper was funded by the German Research Foundation (DFG – Deutsche Forschungsgemeinschaft) within the project B9 of the collaborative research centre 570 “Distortion Engineering”. The authors also gratefully acknowledge the support of the Machine Tool Technologies Research Foundation (MTTRF) and the DP Technology Corp. providing a high-precision 5-axes milling centre (Mori Seiki NMV 5000 DCG) and an efficient CAM software system (ESPRIT). We also thank the company Sandvik Tooling Deutschland GmbH / Sandvik Coromant for loaning high precision slot cutters specialized for gear milling.
BIMAQ: Martina Fuhrmann