Aus der Praxis für die Praxis 1 Introduction A Condition Monitoring System (CMS) can be implemented as an indicator of the wind turbine drive train condition. CMS have demonstrated to be effective to reduce the Operation and Maintenance (O&M) costs in wind turbine applications. Currently, most of the CMS applied in wind turbines are based on vibration measurements. However, these systems have a limited performance concerning the early detection of wear-related damages on gearboxes and bearings. In the recent years, the interest in oil-based condition monitoring systems (OCMs) for wind turbines as a complementary approach has increased. Several publications have shown, how the implementation of particle-counter sensors could give an indication of wear-related damage in the gearbox and of filtration problems [1] [2] [3] [4] [5]. One of the most important point related to OCMs is to assess the capability of the oil sensors to deliver reliable values in order to avoid costly repairs, unnecessary oil sampling or to extend the oil change intervals. Testing sensors before their installation on wind turbines can give the gearbox and wind turbine manufacturer the necessary confidence to implement OCMs, by providing a proven sensor functionality and interpretation of the output signals. Testing several types of oil sensors on a single test bench involves numerous challenges due to the different sensor measurement principles, installation Tribologie + Schmierungstechnik 63. Jahrgang 3/ 2016 45 Oil-Sensors Test Bench: An Approach to Validate Oil Condition Monitoring Systems for Wind Turbine Applications D. Coronado, A. Bustamante, C. Kupferschmidt* Ziel dieser Studie ist es, die wichtigsten Aspekte für die Prüfung von Ölsensoren zur Zustandsüberwachung von Windenergieanlagengetrieben unter realistischen Bedingungen zu identifizieren. Die in den Getrieben von Windenergieanlagen herrschenden, variablen Betriebs- und Umgebungsbedingungen beeinflussen die Sensormessungen. Mittels geeigneter Testszenarien lässt sich in Funktionstests die grundsätzliche Eignung von Ölsensoren für die Zustandsüberwachung prüfen. Ein Prüfstand, der die wesentlichen Betriebsparameter in einem Getriebe nachbildet, ermöglicht die Untersuchung der Detektionsfähigkeit der Sensoren als Teil eines Öl-basierten Zustandsüberwachungssystems. Dieser Artikel widmet sich den Ergebnissen von Vorversuchen mit einem Öleigenschaftssensor unter dem Einfluss von Temperaturänderungen und Verunreinigung des Öls mit Wasser. Praktische Überlegungen zur Homogenisierung des Öls bei Wasser- oder Partikelzugabe werden beschrieben und die wichtigsten Herausforderungen und Verbesserungsansätze werden vorgestellt. Schlüsselwörter Ölsensorik, Püfstand, Windenergieanlagen, Getriebe, Schmierstoffe, Verunreinigungen, Wasser, Abriebpartikel The main objective of this paper is to identify the most relevant aspects for testing oil sensors for condition monitoring of wind turbine gearboxes under realistic conditions. In a wind turbine gearbox, several operating parameters and environmental conditions affect the sensor measurements. By means of suitable testing scenarios, functionality tests can be carried out to validate oil sensors for condition monitoring. A test bench that recreates the most important operating parameters in a gearbox will allow assessing the capability of the sensors to be used as a part of an oilbased condition monitoring system. This paper shows the results of preliminary tests carried out on an oilproperties sensor under influence of temperature variations and water contamination. Practical considerations for achieving homogenization of the oil in case of water or particle insertion are described and the most important challenges and improvement approaches are presented. Keywords Oil sensors, Test bench, wind turbine, gearbox, lubricants, contamination, water, particles Kurzfassung Abstract * MSc. Diego Coronado MSc. Augusto Bustamante Dr.-Ing. Claus Kupferschmidt Fraunhofer Institute for Wind Energy and Energy System Technology IWES Northwest, 27572, Bremerhaven, Germany T+S_3_16 05.04.16 09: 01 Seite 45 Aus der Praxis für die Praxis methods and measured parameters. Therefore, a previous suitability analysis of the testing methodology will improve the assessment of oil sensors for OCMs. This paper shows a methodology to test oil sensors including pre-tests to identify the main challenges to perform testing and assessment of OCMs. The paper is structured as follows: Section 2 describes the methodology implemented to test oil sensors under realistic conditions on a purpose-designed test bench. Based on the requirements from this purpose-designed test bench, initial tests carried out on an experimental setup with an oilproperties sensor are presented. These initial tests serve to identify possible challenges in the testing and assessment methodology. Section 3 shows the initial test results including the influence of temperature, different oil types and water contamination. The results obtained from oil sampling and online measurements for several testing scenarios are presented and discussed. Section 4 includes a discussion of the obtained results. Section 5 presents the conclusions from the experiments, the related challenges as well as suggestions for the improvement of testing oil sensors for wind turbines. 2 Test methodology In order to test oil sensors it is necessary to identify the main parameters that play an important role in the operating conditions of the wind turbine lubrication system. According to IEC 61400-4 [6], wind turbines with a rated capacity of more than 500 kW should be equipped with a pressure-fed lubrication system. Figure 1 illustrates the basic configuration of a wind turbine lubrication system. With the objective to recreate the operating conditions in the gearbox lubrication system, an oil-sensors test bench has been designed at Fraunhofer IWES. The layout is depicted in Figure 2. This oil-sensors test bench offers the possibility to install several types of sensor and to test them under various operating conditions. Typically not all parameters affecting the oil condition can be directly measured by single OCMs. However, some of these parameters can be correlated to oil degradation; therefore their identification is essential. Table 1 summarizes the most relevant sources of disturbances affecting the sensor output signals. The conditions that can be recreated in the oil-sensors test bench are also shown. The selection of the parameter to be emulated in the test bench has been chosen because there is a lack of procedures available for testing. Vibration and ambient temperature have been evaluated with another procedure in [8]. Concerning pressure, a special configuration is required to recreate the real pressure in the pipes, which is dependent on the size of the wind turbine. In the testbench, gearbox oils with different aging states will be used. It should be noted that oil aging will not take place 46 Tribologie + Schmierungstechnik 63. Jahrgang 3/ 2016 Figure 2: Oil-sensors test bench layout Figure 1: Gearbox lubrication system [7] Table 1: Sensor disturbance parameters and related testing possibilities at the IWES oil-sensors test bench Parameters affecting sensor Emulation in the oiloutput signal sensors test bench Temperature YES Flow regime YES Pressure NO Particle content YES Water content YES Air content YES Oil aging NO Oil type YES Sensor positioning YES Ambient temperature NO Vibrations NO T+S_3_16 05.04.16 09: 01 Seite 46 Aus der Praxis für die Praxis in the test-bench. Instead, the oil is aged on a separate gear test rig. For further information concerning the IWES test bench the reader is referred to [9]. The monitoring of water and particles contained in the oil is a very important aspect due to the negative effect of these parameters on the oil condition. Additionally, water content sensors and particle counters should be tested in the IWES test bench. This means that a contamination unit performing an artificial water and particle oil mixture is essential to carry out these tests and to perform an appropriate assessment. In order to gain experience on the influence of water and particle contamination, some pre-tests for an oil-properties sensor have been carried out in a simplified laboratory set-up. The tests consist of: • Homogenization scenario: The oil mixing process with water and particles is evaluated. The main objective is to assess the effectiveness of the homogenization process for water, particles and bubbles production. • Temperature scenario: The oilproperties sensor is tested in the temperature range 40 - 90 °C to observe the influence of temperature variation in the sensor output signal. • Water contamination scenario: The oil-properties sensor is tested considering the influence of water in the oil. Two tests are performed: First, the water content remains constant and the temperature is increased from 40 - 90 °C. In the second test, the temperature is kept constant while the water content is increased. The purpose is to verify, if the sensor can detect the content of water by changes in the measured properties of the oil. The tests are carried out on a double-wall reservoir with a cover adapted to install the oil-properties sensor, a stirrer and a thermometer. The experimental setup is shown in Figure 3. The oil-properties sensor under test is designed to measure simultaneously the viscosity, density, dielectric constant and temperature. This multi-parameter sensor uses a tuning fork sensing element. The manufacturer of the oil-properties sensor states a wide range of applications (e. g. engine oil, hydraulic, transmission, fuel). For gearbox applications for example, the manufacturer states a typical dynamic viscosity of 35 mPa · s at an operational temperature of 50 °C. Moreover, the sensor manufacturer gives a measurement range for the dynamic viscosity between 0,5 mPa · s and 50 mPa · s with an accuracy of ± 5 %. The previous testing scenarios are carried out for three oils with an ISO viscosity grade (VG) of 320: • Oil-1: Mineral oil • Oil-2: Poly-alpha-olefin (PAO) • Oil-3: Polyalkylene glycol (PAG) Oil samples are taken before and after the test to analyze the difference between the quantity of water inserted into the oil and the quantity of water indicated by the laboratory report. 3 Results 3.1 Homogenization The homogenization is achieved by means of a stirrer. To verify the functionality of the mixing process, a sample of 0.5 l of oil and 1 ml of water were used. The behavior of the mixture oil - water at 40 °C is observed when the stirrer is rotating at different rotational speeds. The water is colored in order to have a better visualization of the homogenization process. The results are shown in Figure 4. Tribologie + Schmierungstechnik 63. Jahrgang 3/ 2016 47 Figure 4: Mixing process of oil and water by means of a stirrer When the stirrer starts to rotate, the water gathers around the rotor. After some minutes the homogenization is achieved. The temperature accelerates the mixing process, however, care should be taken to not increase the stirrer rotational speed to avoid bubbles generation. The same test has been carried out with particles. The results show that the particles are difficult to spread and therefore, agitation by means of the stirrer is not enough to achieve homogenization. In the process, the particles precipitate to the bottom of the reservoir where they adhere to the surface. An increase in the rotational speed of the stirrer is not effective to perform the particle mix in the oil. A more elaborated method is required to achieve a homogenization of the oil and the particles; therefore the effect of particles will not be included in the further analysis presented in this paper. 3.2 Temperature scenario In this scenario, the oil parameters have been recorded during 5 minutes at 40, 50, 60, 70, 80 and 90°C. A thermometer has been used as a reference for comparison with the temperature signals received from the oil-properties sensor. For all oils that have been tested, it was found that the deviations between temperatures measured by the thermometer and temperatures measured by the sensor are in the range of about +1 K. For the measurements of the viscosity of Oil-2, the sensor showed a high deviation for measurements at 40°C. Some variations were also observed at 50°C, but for fluid temperatures over 50°C the signal curves for the viscosity are nearly constant along the measured time, as it can be observed in Figure 5. The results indicate that the viscosity measurements by means of the oil-properties sensor under test is afflicted with a high error at temperatures lower than 50°C. Figure 5: Viscosity measurements for different temperatures With rising temperatures of the oil, the error in the viscosity measurements decreases and reaches values that are within the measurement range of the sensor, that is in maximum 50 mPa·s. In order to verify the accuracy of the viscosity measurements delivered by the sensor, a comparison with the oil manufacturer data was carried out. For comparing the results, the dynamic viscosity measured by the sensor was converted to kinematic viscosity based on the density given by the oil manufacturer. Both viscosity curves for Oil-2 are shown in Figure 6 and prove the previously given explanations about the measurement range of the sensor. Figure 6: Viscosity measurements in comparison with oil manufacturer's values For temperatures below 70°C the viscosity is out of the measurement range of the sensor and therefore the values show high deviations compared with the expected viscosity values. Concerning density, the measurements taken below 60°C show a high deviation from the oil manufacturer’s data, which is around 0.8 g/ cm 3 for Oil-2. The same result was given for the dielectric constant. Figure 4: Mixing process of oil and water by means of a stirrer concerning the IWES test bench the reader is referred to [9]. The monitoring of water and particles contained in the oil is a very important aspect due to the negative effect of these parameters on the oil condition. Additionally, water content sensors and particle counters should be tested in the IWES test bench. This means that a contamination unit performing an artificial water and particle oil mixture is essential to carry out these tests and to perform an appropriate assessment. In order to gain experience on the influence of water and particle contamination, some pre-tests for an oilproperties sensor have been carried out in a simplified laboratory set-up. The tests consist of: · Homogenization scenario: The oil mixing process with water and particles is evaluated. The main objective is to assess the effectiveness of the homogenization process for water, particles and bubbles production. · Temperature scenario: The oilproperties sensors is tested in the temperature range 40-90°C to observe the influence of temperature variation in the sensor output signal. · Water contamination scenario: The oil-properties sensor is tested considering the influence of water in the oil. Two tests are performed: First, the water content remains constant and the temperature is increased from 40-90°C. In the second test, the temperature is kept constant while the water content is increased. The purpose is to verify, if the sensor can detect the content of water by changes in the measured properties of the oil. The tests are carried out on a double-wall reservoir with a cover adapted to install the oil-properties sensor, a stirrer and a thermometer. The experimental setup is shown in Figure 3. The oil-properties sensor under test is designed to measure simultaneously the viscosity, density, dielectric constant and temperature. This multi-parameter sensor uses a tuning fork sensing element. The manufacturer of the oil-properties sensor states a wide range of applications (e.g. engine oil, hydraulic, transmission, fuel). Figure 3: Experimental setup for initial tests of temperature and water influence to oil For gearbox applications for example, the manufacturer states a typical dynamic viscosity of 35 mPa·s at an operational temperature of 50°C. Moreover, the sensor manufacturer gives a measurement range for the dynamic viscosity between 0,5 mPa·s and 50 mPa·s with an accuracy of ±5%. The previous testing scenarios are carried out for three oils with an ISO viscosity grade (VG) of 320: · Oil-1: Mineral oil · Oil-2: Poly-alpha-olefin (PAO) · Oil-3: Polyalkylene glycol (PAG) Oil samples are taken before and after the test to analyze the difference between the quantity of water inserted into the oil and the quantity of water indicated by the laboratory report. 3. Results 3.1 Homogenization The homogenization is achieved by means of a stirrer. To verify the functionality of the mixing process, a sample of 0.5 l of oil and 1 ml of water were used. The behavior of the mixture oil - water at 40°C is observed when the stirrer is rotating at different rotational speeds. The water is colored in order to have a better visualization of the homogenization process. The results are shown in Figure 4. Motor Stirrer Double wall reservoir Thermometer Oil-properties sensor Figure 3: Experimental setup for initial tests of temperature and water influence to the oil T+S_3_16 05.04.16 09: 01 Seite 47 Aus der Praxis für die Praxis When the stirrer starts to rotate, the water gathers around the rotor. After some minutes the homogenization is achieved. The temperature accelerates the mixing process, however, care should be taken to not increase the stirrer rotational speed to avoid bubbles generation. The same test has been carried out with particles. The results show that the particles are difficult to spread and therefore, agitation by means of the stirrer is not enough to achieve homogenization. In the process, the particles precipitate to the bottom of the reservoir where they adhere to the surface. An increase in the rotational speed of the stirrer is not effective to perform the particle mix in the oil. A more elaborated method is required to achieve a homogenization of the oil and the particles; therefore the effect of particles will not be included in the further analysis presented in this paper. 3.2 Temperature scenario In this scenario, the oil parameters have been recorded during 5 minutes at 40, 50, 60, 70, 80 and 90 °C. A thermometer has been used as a reference for comparison with the temperature signals received from the oil-properties sensor. For all oils that have been tested, it was found that the deviations between temperatures measured by the thermometer and temperatures measured by the sensor are in the range of about +1 K. For the measurements of the viscosity of Oil-2, the sensor showed a high deviation for measurements at 40 °C. Some variations were also observed at 50 °C, but for fluid temperatures over 50 °C the signal curves for the viscosity are nearly constant along the measured time, as it can be observed in Figure 5. The results indicate that the viscosity measurements by means of the oil-properties sensor under test is afflicted with a high error at temperatures lower than 50 °C. With rising temperatures of the oil, the error in the viscosity measurements decreases and reaches values that are within the measurement range of the sensor, that is in maximum 50 mPa · s. In order to verify the accuracy of the viscosity measurements delivered by the sensor, a comparison with the oil manufacturer data was carried out. For comparing the results, the dynamic viscosity measured by the sensor was converted to kinematic viscosity based on the density given by the oil manufacturer. Both viscosity curves for Oil-2 are shown in Figure 6 and prove the previously given explanations about the measurement range of the sensor. For temperatures below 70 °C the viscosity is out of the measurement range of the sensor and therefore the values show high deviations compared with the expected viscosity values. Concerning density, the measurements taken below 60 °C show a high deviation from the oil manufacturer’s data, which is around 0.8 g/ cm 3 for Oil-2. The same result was given for the dielectric constant. 3.3 Water contamination scenario For the water contamination scenario, a quantity of 0,375 ml water was mixed with the oil, resulting in a water content of 600ppm with respect to volume. Measurements were carried out for 40, 50, 60, 70, 80 and 90 °C. The results from this test were compared with the results obtained in the temperature scenario. The temperature 48 Tribologie + Schmierungstechnik 63. Jahrgang 3/ 2016 Figure 5: Viscosity measurements for different temperatures Figure 6: Viscosity measurements in comparison with oil manufacturer’s values T+S_3_16 05.04.16 09: 01 Seite 48 Aus der Praxis für die Praxis measurements were not affected by the insertion of water. Moreover, slightly changes were detected in the viscosity and density. However, the density at higher temperatures showed a higher deviation. Concerning dielectric constant, it can be observed different values for the same temperature. For both cases the dielectric constant decreases with an increase in the temperature. The different sensor parameters as a function of temperature with water and without water are illustrated in Figure 7 for Oil-2. A second test was carried out at a constant temperature of 70 °C for different water contamination levels. Figure 8 illustrates the temperature as a function of the injected water. The results show that the water content does not have a significant influence on the temperature measurements. Concerning viscosity, no tendency could be clearly identified. For Oil-1, the viscosity tends to increase. The measurements show a standard deviation between 8 - 18 %. The viscosity measured by the oil-properties sensor without water showed a deviation of approximately 8 % in regard to the value stated by the oil manufacturer. The results are shown in Figure 9. For Oil-2, a slightly decrease in the viscosity was observed. The standard deviation of the measurements was between 1 - 6 %. The viscosity measured by the oil-properties sensor without water showed a deviation of approximately 6 % in regard to the value stated by the oil manufacturer. The results are shown in Figure 10. For Oil-3, the viscosity increased for water contents lower than 400 ppm. For higher water content, the viscosity decreased. The standard deviation of the measurements was between 6 - 15 %. The viscosity measured by the oil-properties sensor without water showed a deviation of approximately 61 % in regard to the value stated by the oil manufacturer. The results are shown in Figure 11. The sensor showed relatively stable signals for Oil-1 and Oil-2 with water content lower than 806 ppm. The measure- Tribologie + Schmierungstechnik 63. Jahrgang 3/ 2016 49 a) b) c) d) Figure 7: Sensor parameters under the influence of water as a function of the oil temperature. a) Temperature, b) Viscosity, c) Density, d) Dielectric constant T+S_3_16 05.04.16 09: 01 Seite 49 Aus der Praxis für die Praxis ments for Oil-3 exhibited a higher deviation especially for values higher than 806 ppm. The results show that the oilproperties sensor is sensitive to the type of oil. Furthermore, for high water content i. e. 1200 ppm, the sensor did not provide a stable signal oscillating constantly especially for Oil-3 as illustrated in Figure 12: The density measurements showed a different behavior for each oil type. In the case of Oil-1, the standard deviation in the measurements was approximately 8 %. For Oil-2 the maximum standard deviation was around 5 %. For Oil-3, the standard deviation oscillates between 4 -9 %. The variations of density due to water injection remained within the standard deviation limits; therefore no clearly indication of a change in the density due to the water content can be deducted from the sensor measurement signals. Figure 14 illustrates the density measurements. Oil samples were taken before and after the tests. The laboratory analysis was a standard analysis. After the tests, the total amount of water inserted into the oil was 1200 ppm. It was expected that the infrared spectrum analysis showed some indication of water contamination; however the results for the 3 oils concerning water contamination showed normal results. Slightly changes in the viscosity, not larger than 3 mm 2 / s where observed. This is within the range of the method’s accuracy therefore no change in the viscosity could indicate water contamination. For water content analysis of the test oils, a Karl Fischer analysis was performed to identify more exactly the real water content in the sample. Concerning additives, changes in phosphor and sulfur were observed in the 3 oils. Oil-1 and Oil-2 showed a slightly increase in the phosphor content. Oil-3 showed a phosphor decrease. Concerning sulfur, Oil-1 presented an increase. Oil-2 and Oil-3 showed a slightly decrease. It should be noted that this changes could be caused for two reasons. First the accuracy of the method and second the effect of water in Extreme Pressure (EP) additives like phosphor and sulfur. The change in percentage can be observed in Table 2. 50 Tribologie + Schmierungstechnik 63. Jahrgang 3/ 2016 Figure 9: Influence of water on viscosity measurements for Oil-1 Figure 8: Influence of different water contamination levels on temperature measurements Figure 10: Influence of water on viscosity measurements for Oil-2 Figure 11: Influence of water on viscosity measurements for Oil-3 T+S_3_16 05.04.16 09: 01 Seite 50 Aus der Praxis für die Praxis 4 Discussion 4.1 Homogenization From the results of the test, we can conclude that a simple mechanical system is enough to achieve water homogenization and perform a test for oil sensors. However, the mixture of particles requires a more complex process. The particles stick together and cannot be easily separated. Previous spreading of the particles is required and a more efficient agitation method would be necessary. One recommendation would be to use ultrasonic agitation as suggested in [10]. However several aspects as bubble formation due to increased temperature, viscosity, adhesion and deposition of the particles need to be further analyzed. 4.2 Temperature influence For all measurements, it was found that the deviations between temperatures measured by the thermometer and temperatures measured by the sensor are in the range of about +1 K that is within the measurement accuracy of the oilproperties sensor. However the viscosity and density measurements showed a significant deviation from the value provided by the oil manufacturer. On the other hand, the sensor starts to deliver reliable values for temperatures higher than 60 °C that corresponds to the accuracy of the sensor specified by the manufacturer. The dielectric constant measurements could not be compared with oil manufacturer’s data, therefore only a trend analysis was carried out. From the test results it can be said that the dielectric constant exhibited a decrease of approximately 5 % with temperature. 4.3 Water Influence According to the results, water does not produce any changes in the temperature Tribologie + Schmierungstechnik 63. Jahrgang 3/ 2016 51 Figure 12: Oil-3 viscosity measurements for water content of 1200 ppm and a temperature of 70 °C Figure 13: Influence of water on the dielectric constant measurements Figure 14: Influence of the water content on the density measurements Table 2: Percentage additive change for samples with water contamination Phosphor Sulfur Oil-1 +3,6% +2,6% Oil-2 +1,1% -2,8% Oil-3 -5,9% -3,6% T+S_3_16 05.04.16 09: 01 Seite 51 Aus der Praxis für die Praxis output signal of the sensor. During the test with constant water content and changing temperature, it could be observed that dielectric constant decreases as in the temperature test. This could be explained as the saturation level of the oil changes with temperature, which could affect the output signal of the sensor, which is also sensitive to temperature changes. However, when the temperature was kept constant at 70 °C, the dielectric constant increased. The correlation between dielectric constant and water content has been already observed in [11]. This correlation or cross-effect could cause difficulties to assess the changes in the dielectric constant due to water ingress. Density changes due to water contamination were not possible to detect as the variation of the density was within the standard deviation of the measurement. Even if common practice in CMS is to perform trend analysis, it is necessary to know the interaction of the measured parameters and carry out a correlation analysis for a measurement value correction. According to the laboratory report on the oil-sample analysis, Oil-1 and Oil-2 needed a second sampling to verify their condition; however for Oil-3, the status of the sample was normal. This is expected as the Oil-3 is a PAG and it can dissolve more water than mineral and PAO oils: Alarms levels for a PAO and mineral oil are around 600 ppm and around 15 000 ppm for a PAG [12]. From the oil samples, slight changes in the oil additives were identified. However, it is not possible to draw definite conclusions about these changes as the accuracy of the method does not allow to associate these changes to water contamination. Nevertheless, water diminishes the oil lubrication capabilites affecting EP additives e. g. phosphore and sulfur, which usually act to minimize friction under high pressure and high temperature conditions as occurs in gearboxes [13]. Therefore, a correlation between water ingress and EP changes should not be completely discarded. 5 Conclusion Testing oil sensors involves numerous challenges from the testing procedures to the assessment of the results. From the experiments executed in this study it can be concluded that: • Water homogenization is possible with a simple mechanical system, however particle homogenization requires a special procedure to distribute the particles uniformly and inject them in a measurement pipe. The particles should not be added to the walls of the reservoir or pipes. Proper materials, sealing systems, exhaustive cleaning and filtering will be necessary to ensure the creation of realistic conditions being present in wind turbine gearboxes. • The sensor under test did not provide accurate viscosity and density measurements for temperatures lower than 60 °C. This inaccuracy can be problematic for wind turbines which also operate at oil temperatures lower than 60 °C. • The influence of water was identified by an increase in the dielectric constant, however as the temperature has counter effect decreasing the dielectric constant, it is necessary to have a temperature correction of the measurement values of the dielectric constant in order to isolate the pure effect of water contamination. • The accuracy of the sensor is dependent on the type of oil. The sensor showed a better performance for the mineral and the PAO oil as for the PAG oil. • The water content in the oil was not clearly identified by the sensor. Standard laboratory analysis give the water content based on infrared spectroscopy which only identify the water content larger than 1000 ppm, which in a gearbox with PAO or mineral oils is in the alert range. Therefore a Karl Fischer analysis needs to be performed to identify the water content in the oil samples. • Changes in the additives caused by water could not be identified in the standard oil-sample analysis in the present study; however it is known from the literature that water can affect EP additives. A more accurate measurement could help to identify more precisely the correlation between these two parameters. References [1] S. Sheng, „Investigation of oil conditioning, real-time monitoring and oil sample analysis for wind turbine gearboxes,“ in AWEA Project Performance and Reliability Workshop, San Diego, 2011. [2] J. Stover and C. Jensen, „The Road Map to Effective Contamination Control in Wind Turbines,“ in Condition Monitoring Workshop, 2011. [3] J. Ukonsaari and H. Moller, „Oil Cleanliness in Wind Power Gearboxes,“ ELFORSK, Stockholm, 2012. [4] T. Meindorf, „Condition Monitoring Facetten moderner, umfassender Zustandsüberwachung und ihr Kundennutzen,“ in Ölsensoren-Symposium 2013, Brannenburg, 2013. [5] A. Toms, „Oil debris monitoring: Part of an effective gearbox monitoring strategy,“ in 69 th STLE annual meeting and exhibition, Lake Buena Vista, 2014. [6] I. E. C. IEC, IEC 61400-4 : Wind turbines - Part 4: Design requirements for wind turbine gearboxes, Geneva: International Electrotechnical Comission IEC, 2012. [7] K. Fischer, F. Besnard and L. Bertling, „Reliability Centered Maintenance for Wind Turbines Based on Statistical Analysis and Practical experience,“ IEEE Transaction on Energy Conversion, vol. 27, no. 1, pp. 184-195, 2012. [8] D. Coronado and C. Kupferschmidt, „Assessment and Validation of Oil Sensor Systems for On-line Oil Condition Monitoring of Wind Turbine Gearboxes,“ in 2 nd International Conference on System-Integrated Intelligence: Challenges for Product and Production , Bremen, 2014. 52 Tribologie + Schmierungstechnik 63. Jahrgang 3/ 2016 T+S_3_16 05.04.16 09: 01 Seite 52 Aus der Praxis für die Praxis [9] D. Coronado, C. Kupferschmidt, C. Wirth and M. Ernstorfer, „Design and Construction of a Test Bench for Assessment and Validation of Oil Sensor Systems for Online Oil Condition Monitoring of Wind Turbine Gearboxes,“ in STLE Annual Meeting and Exhibition, Lake Buena Vista, 2014. [10] Internationa Standard Organization ISO, ISO 16889 - Hydraulic fluid power - Multi-pass method for evaluating filtration performance of a filter element, Geneva: ISO, 2008. [11] Y. Dingxin, Z. Xiaofei, H. Zheng and Y. Yongmin, „Oil Contamination Monitoring Based on Dielectric Constant Measurement,“ in International Conference on Measuring Technology and Mechatronics Automation, 2009. [12] Siemens AG, „Getriebeschmierung für Stirnrad-, Kegelrad-, Kegelstirnradgetriebe, Planetengetriebe und Getriebemotoren,“ Siemens AG, Bocholt, 2012. [13] N. Canter, „Special Report: Trends in extreme pressure additives,“ Tribology & Lubrication Technology, 2007. Tribologie + Schmierungstechnik 63. Jahrgang 3/ 2016 53 Bestellcoupon Tribologie und Schmierungstechnik „Richtungsweisende Informationen aus Forschung und Entwicklung“ Getriebeschmierung - Motorenschmierung - Schmierfette und Schmierstoffe - Kühlschmierstoffe - Schmierung in der Umformtechnik - Tribologisches Verhalten von Werkstoffen - Minimalmengenschmierung - Gebrauchtölanalyse - Mikro- und Nanotribologie - Ökologische Aspekte der Schmierstoffe - Tribologische Prüfverfahren Bestellcoupon Ich möchte Tribologie und Schmierungstechnik näher kennen lernen. Bitte liefern Sie mir ein Probeabonnement (2 Ausgaben), zum Vorzugspreis von 7 39,-. So kann ich die Zeitschrift in Ruhe prüfen. Wenn Sie dann nichts von mir hören, möchte ich Tribologie und Schmierungstechnik weiter beziehen. Zum jährlichen Abo-Preis von 7 184,- Inland bzw. 7 198,- Ausland. Die Rechnungsstellung erfolgt dann jährlich. Das Jahresabonnement ist für ein Jahr gültig; die Kündigungsfrist beträgt sechs Wochen zum Jahresende. Firma, Abteilung Straße, Nr. Name, Vorname PLZ, Ort Ort/ Datum, Unterschrift: (ggf. Firmenstempel) Coupon an: expert verlag, Abonnenten-Service, Postfach 2020, 71268 Renningen oder per Fax an: (0 71 59) 92 65-20 Im expert verlag erscheinen Fachbücher zu den Gebieten Weiterbildung - Wirtschaftspraxis - EDV-Praxis - Elektrotechnik - Maschinenwesen - Praxis Bau / Umwelt/ Energie sowie berufs- und persönlichkeitsbildende Audio-Cassetten und -CDs (expert audio ) und Software (expert soft ) Bitte fordern Sie unser Verlagsverzeichnis auf CD-ROM an! expert verlag Fachverlag für Wirtschaft & Technik Telefon (0 71 59) 92 65-0 · Telefax (0 71 59) 92 65-20 Wankelstraße 13 · D-71272 Renningen E-Mail expert@expertverlag.de Postfach 20 20 · D-71268 Renningen Internet www.expertverlag.de T+S_3_16 05.04.16 09: 01 Seite 53