Skip to main content

Online Remote Condition Monitoring of Tidal Stream Generators

Final Report Summary - REMO (Online Remote Condition Monitoring of Tidal Stream Generators)

Executive Summary:
Tidal stream power is a very attractive renewable energy source whose progress is being delayed by operation and maintenance problems, which result in very low operational availability times, as low as 25%. The goal of the REMO project is essentially to provide an enabling technology for tidal stream energy. It does this by reducing the projected life cycle maintenance costs of tidal stream energy by 50% and the generator downtime and associated lost generation revenue to a level comparable with wind turbines ie it will achieve an availability time exceeding 96%. Energy providers will then be attracted to investing in tidal stream energy, so that its full economic potential and environmental advantages are realised. In order to achieve this goal, the REMO project has developed a system for the remote and permanent monitoring of the structural integrity of all the rotating components in a tidal stream generator so that advanced warning of potential structural failure is obtained well in advance, avoiding potential damage and allowing all necessary component repairs or replacements to be performed at scheduled maintenance intervals, targeted to be less than 15 days per year (ie availability time of at least 96%), to minimise both maintenance costs and lost electricity revenue. The estimated maintenance cost reduction based on precedents with wind turbines will be 50%. This strategy reduces current projected costs of tidal stream energy production down to levels comparable with wind turbine electricity costs (0.058€/kWh), thus ensuring the economic viability of tidal generators.

REMO technology is based on the monitoring of structural vibrations throughout the entire frequency spectrum generated by the rotating components and so combines a suite of accelerometer sensors for the low frequency regime and acoustic emission sensors for the high frequency regime.
The REMO technology permanently monitors tidal wave generators and determines the vibrational signature of a healthy turbine and the evolution of that signature during the turbine life cycle. It then highlights any significant change in that signature that could be a symptom of a structural health problem or the build-up of marine fouling at any point in the life cycle and then issue an automatic warning.

Project Context and Objectives:
The project concept
In order to achieve this goal the REMO project has developed a system (See Fig 1.1a) for the remote and permanent monitoring of the structural integrity of all the rotating components in a tidal stream generator so that advanced warning of potential structural failure are obtained well in advance, avoiding potential damage and allowing necessary component repairs or replacements to be performed at scheduled maintenance intervals, targeted to be less than 15 days per year (ie availability time in excess of 96%, to minimise both maintenance costs and lost electricity revenue. The estimated maintenance cost reductions, based on precedents with wind turbines will be 50% [1,2] This strategy will reduce present projected costs of time stream energy production down to levels comparable with wind turbine electricity costs (0.058€/kWh) [3], thus ensuring the economic viability of tidal generators.

REMO technology is based upon the monitoring of structural vibrations throughout the entire frequency spectrum generated by the rotating components and so combines a suite of accelerometer sensors for the low frequency regime and acoustic emission sensors for the high frequency regime.

The REMO technology permanently monitors tidal wave generators and determine the vibrational signature of a healthy turbine and the evolution of that signature during the turbines life cycle. It then discovers any significant change in that signature that could be a symptom of a structural health problem at any point in the life cycle, including the build-up of marine fouling and then issue an automatic warning.

State of the art similarity analysis algorithms based on the Euclidian distance measure in multiple dimensions is used in both time and frequency domain to process all vibrational data involved in the state of health diagnosis [4].

The system has been validated by installing it on an in-service tidal stream generator mock up developed by one of the SMEs who will also be an end user of the proposed REMO technology.
Project Objectives
1. To develop waterproof sensors and electronic hardware for the acquisition of the complete vibrational spectrum of the rotational components in the drive train of a tidal stream generator.
Deliverable 2.1 in work package 2 was to develop accelerometer and acoustic emission sensor packs for underwater deployment. Within the REMO project the consortium has developed four acoustic emission sensors and four accelerometers which are water and pressure proof to a water depth of 50 meters according to the project requirements. The sensors were initially tested in the laboratory submerging them inside a water vessel. The test was conducted in two stages in order to check that both the case and the connector were both waterproof. The first waterproof test carried out in the lab showed that after 24 hours with the sensor casings submerged, the probe was still functional which meant that the sensor casings were waterproof. The second test showed that after a further 24 hours with all parts of the sensors unit fully submerged, the sensors were still functional which, when combined with the first test, concluded that both the sensor casing and connector were waterproof. The attachment to the tidal generator surface is different for each of these technologies. In the case of Acoustic Emission, the transducer is attached by using very strong magnetic holders whereas the accelerometers are attached using pot magnets screwed into them.
In work package 5 the REMO sensors were tested in a real environment. This experiment was performed in a water tank where the transducers worked under higher pressure conditions. The signals coming from the sensors were totally correct what meant the sensors were waterproof and in addition, they worked successfully under higher pressure conditions. On the other hand, the transducers were still attached to the tidal generator device for the whole experiment what meant that the attachment methods used was suitable for this case study. Further information regarding the signals acquire under water is gathered in deliverable 5.1.

It has been demonstrated that the hardware worked satisfactorily in the drive train. Deliverable 4.1 in work package 4 demonstrates the correct functioning of the software during the test carried out in the structural health monitoring mock-up. The post processing software was validated acquiring data continuously for a long period of time.

2.To develop software for data presentation including
(i) Plots of sensor output voltage signal as a function of time

(ii) Fourier transforms of signal-time waveforms

(iii) Computation of various forms of signal averaging techniques such as root–mean–square voltage (VRMS), peak to peak average voltage.
The data acquired from the developed REMO system is displayed and analysed by the developed software. The software is capable of plotting the data both in time and frequency domain. Other features such as truncating and zoom are also available in the software. Various damage sensitive features such as VRMS, peak-to-peak, crest factor and kurtosis can be extracted from the time domain signals. All of them play a very important part in the detection and identification of defects.

(i) Time waveform analysis has demonstrated to be a powerful tool when it comes to analysing measurements from gearboxes. The variance in amplitude (showed by some of the signals acquired after damaging the gearbox) indicates that the gears do not mesh properly due to the defect generated. So this technique depicts at a glance that the status of the gearbox has worsened.
(ii) The time domain signal was then processed transform it into frequency domain for more detailed analysis and increased diagnostic capability. This kind of analysis helps the operator establish the defect source. According to the conversation the consortium had with potential end users of the REMO system, this one is the most important processing tool.
(iii) The features RMS, peak-to-peak, crest factor and kurtosis can be extracted.

In summary, the software developed during the REMO project has capabilities to detect changes in the status of the machine (time domain or feature extraction) and also identify the source of defect (frequency analysis). The software has been developed within work package 3. All the details about the software can be found in the deliverable 3.1. The continuous development and the result arising from the development is shown in deliverables 3.2 and 3.3.

3. To develop pattern recognition signal processing software based on similarity analysis and Euclidean distance.
The pattern recognition algorithm based on the Euclidean distance is employed on damage sensitive features extracted from the time domain signals. To improve the reliability of the detection capability, ISO standards have been implemented in the software package. The algorithm is able to distinguish between various types of defects whether they are gears, bearing or looseness.

The software and its performance are shown in the videos uploaded to the REMO webpage (

4 Validation of the system to specification in laboratory trials, with demonstration on an in service prototype tidal stream generator.
The stopper and drive train mock-up have been developed. The stopper has been developed by Degima and TWI is currently organising for the 175kg sample to be delivered at TWI for experimental work.

The system has been validated in a tidal stream generator and in a WEC (Wave Energy Converter) mock-up. The tidal stream generator mock-up has been the subject of many different tests in both the laboratory and real environments. Tests in the laboratory focused on demonstrating the capabilities of the system to identify the differences between a healthy and a faulty status using Acoustic Emission and Vibration.

Project Results:
Work Package Main Results – WP1

Partners Involved: WLB, Degima, Coservices, Tangent, Pulse Tidal, TWI, BIC, Engitec

Functional design specification and requirements of the REMO system
The hardware employed in the REMO project includes four accelerometers, four acoustic emission (AE) sensors, and one data acquisition unit. The equipment has been selected from commercial products as a basis and customized to ensure it remains waterproof. The hardware specifications are expected to be higher than those used in wind turbine monitoring, due to low rotation, long communication links and also the adverse conditions under sea. The general objectives of the REMO hardware are:

• Simplicity
• Low-cost
• Light weight and portability
• Shock-proof
• Fast and easy installation

Sensor specifications
The sensors (accelerometers and acoustic emission sensors) are required to be waterproof and pressure-proof to a water depth of 50 meters. They are also required to magnetically attach to the tidal stream generator. The sensors specifications are summarized in Table 7.

Both accelerometers and acoustic emission sensors are used for covering a wide signal spectrum. As a result, a significant amount of information from the tidal stream generator is monitored. The accelerometers measure vibrations (acceleration signals) in a lower frequency range. They are expected to have a 3dB bandwidth from 0.3 Hz to 100 Hz, with a mean output sensitivity of greater than 100 mV/g (g is the gravity), whereas the amplitude linearity is lower than 1%. The acoustic emission sensors measure the acoustic pressure in a higher frequency range. They are expected to have a 5 dB bandwidth from 100 KHz to 950 KHz, with a mean output sensitivity of greater than 10mV/mbar and amplitude linearity lower than 1%. The sensors are attached to the tidal stream generator using rare earth permanent magnetic clamps or magnetic bases, with an expected pull strength of over 10 KG.

Due to the long links between the data acquisition unit (located in the buoy above sea level) and the sensors (attached on the tidal stream generator under water), miniature preamplifiers are installed close to the sensors (see Table 8). Since the accelerometers usually have very high output impedance, the preamplifiers are also an impedance transformer with an output impedance of 50 Ohms which is typical for most data acquisition systems. Both sensors and the preamplifiers are waterproof and pressure-proof for tidal currents of 2.5m/s and water depths of 30-50 meters. The equipment is also expected to have a long life cycle compared to the generator maintenance intervals.

Definition of hardware platform and communication protocol
Since the sensor outputs are analogue voltages, a data acquisition unit is required to digitalize the analogue signal for further signal processing and data collection. The specifications of the data acquisition unit are summarized in Table 9.

There are four accelerometers and four acoustic emission sensors used in the REMO project, therefore the data acquisition unit must have at least eight analogue input channels. In addition, the power, tidal current speed, water temperature and temperature readings belonging to other locations (eg the gearbox) may need to be measured by other commercial sensors. Hence, four additional channels are required. If the accelerometers and the acoustic emission sensors are mounted on the marine structure for structural monitoring, a couple of additional channels are also required. Thus, a data acquisition unit with 12 - 16 analogue inputs is appropriate for the REMO project.

For Data transmission and communications, The available options are:
• Wi-Fi protocol is available for small distant transmissions, allowing a maximum of 100-200m communication. Although this may be sufficient in some situations, it may become unsuitable for turbines located at distances of 100 miles into the sea. Wi-Fi may be used to transmit data from the structure to a base station computer onsite (provided the distances are within range and signal strength is present) using networking or simply connecting to the internet through a gateway, and reading the data from further distances through the internet. Data rates of up to 11 Mbits/s can be achieved allowing mass data transmissions (This would be preferable because raw data could reach 100’s of Mega Bytes.) If this is connected to the internet (through a gateway), it may be used to host a website, however it would not be advisable as the operation of the system may be compromised if excess of users log in at the same time. If there is data connectivity via LAN (Ethernet) at the structure, it may be used to connect directly the REMO system in the turbine to the land based base station for processing and web hosting.

Figure 1 Wi-Fi protocol.

• GPRS may be used to transmit data from further distances (provided a mobile network is present in-situ). Data rates are limited to approximately 150 Kbits/s, which would not be recommended for transmission of large amounts of data. It can be used to transmit data from the MPS to a land based server but not vice versa. The suitable method would be to allow the MPS to search for a new configuration at the time of download which is initiated by the MPS at a predefined interval. This procedure is achieved by making a FTP connection to an FTP server via GPRS. The FTP server may be located anywhere in the world, because GPRS connects the MPS to the server via the internet. It may not be used to host a website and the hosting would need to be in a remote location.
Using GSM, sms messages may be sent and received by the MPS which may be used to either send a text based configuration alteration or to initiate the Download data-Upload configuration process.

If there is data connectivity via LAN (Ethernet) at the turbine, it may be used to directly connect the equipment in the turbine to the land based base station for processing and web hosting.

Figure 2 GSM/GPRS protocol.

Work Package Main Results – WP2

Partners Involved: Coservices, Tangent, TWI, BIC

Accelerometers and AE Packages for underwater deployment with magnetic coupling
Waterproofing methods
In order for the sensor to be used in underwater conditions it must be modified so that it can retain its functionality when submerged. One feature of the sensor which makes it suitable for underwater use is the way the internal workings are mounted in a solid resin tub, this means that the insides are partially protected from water should any get inside the device. However, the BNC connector socket is the main area where water can enter the sensor, because of this it must be waterproof as any water inside the sensor would cause the sensor to stop working. There were two approaches considered to waterproof the sensor.

The first of these involve using hot resin to create a waterproof seal around the BNC connection before applying heat shrink over the top in order to maximise water resistance (Figure 6); this process provides a solid case for the sensor connection. However this method may result in too much thermal transfer into the sensor and cause damage; it might also not be possible to remove the sensor from the casing once dried.

Figure 3 The first waterproofing method for the REMO sensor showing the sensor with a resin seal around the BNC connection protected by rubber heat shrink.

The second method considered involves applying heat shrink sleeving to the BNC connector and cable, applying a silastomer adhesive around the uncovered area; this would then be allowed to dry before applying heat shrink sleeving over the whole area (Figure 7). This method was suggested because it relies on the plastic melting properties of the heat shrink sleeving to seal the BNC to the sensor body rather than using hot resin which may cause damage to the internal workings of the sensor.

Waterproof testing experiment
In order to determine whether the waterproof sensor was suitable for this project, tests were conducted. Firstly, the sensor was connected to the decoupling box and attached to the data acquisition systems that were developed in task 2.3 for this project. A motor was run and the AE sensor held in place while measuring the readings from the motor and conducting a pencil lead break (PLB), a Vallen MAG4R was used to hold the sensor. PLB tests were used to demonstrate that the sensor is still responsive and functional.

Figure 4 Sensor test set up

After having established baseline readings for the sensor, waterproof tests were conducted in two stages. Firstly, the sensor was submerged in water (the connection shall remain above the water as shown in Figure 10), the sensor was left for 24 hours and then re-tested the next day to check if it is still functional. The test was conducted out of water in order to minimise the risk to the data collection equipment. Once the test has been completed, the next step was to fully submerge the sensor and the connection in the water vessel (Figure 11) for an additional 24 hours before repeating the tests conducted in the first half of this experiment. By performing two separate tests it allows for a basic level of detection for waterproof faults if they occur. it was possible to tell if the sensor casing has failed or if the BNC waterproof connection has failed.

Figure 5 Second waterproof test of the REMO sensor. Whole unit (sensor and magnetic holder submerged in approximately 10 cm water.)

Figure 6 Preliminary waterproof testing for the REMO sensor. AE sensor immersed in water (5cm depth) with waterproofed BNC connector above water level.
Work Package Main Results – WP3

Partners Involved: Coservices, Engitec

Data presentation and signal processing system

Test rig testing outside water using ENGITEC’s acquisition system

A set of tests took place on dry land using ENGITEC’s acquisition equipment for recording vibration and high-frequency acoustic emission data. The photographs in Figure 8 show the experimental setup during tests carried out in collaboration with TWI and Brunel at Granta Park, Cambridge.

Figure 7 Tests at TWI, Cambridge, UK carried out with ENGITEC’s own acquisition equipment and software.

Finally, in order to assess the capability of the software using faults developed naturally, tests were carried out using ENGITEC’s data acquisition system on large industrial wind turbines.

Test rig experiments were carried out on dry land at TWI’s facility in Granta Park Cambridge. For data acquisition ENGITEC’s condition monitoring system was employed. For the AE measurements four PAC R50A sensors were employed. Two 25 kHz Wilcoxon accelerometers were used for vibration measurements.

AE signals were amplified using four PAC pre-amplifiers and one KRESTOS four-channel digital amplifier. Accelerometers were powered using a KRESTOS four-channel power supply. Decoupling was achieved through a custom-made decoupling box. Signals were acquired using two Agilent 2531 data acquisition cards.
Data was logged on using an AMPLICON IMPACT R Dual core industrial computer. Custom-made software was used for data logging of multiple channels simultaneously. Data was then analysed with the REMO software as previously.

The following defects were considered during testing; gear tooth damage only, impact and gear tooth damage simultaneously present, imbalance and gear tooth damage simultaneously present, imbalance, impact and gear tooth damage simultaneously present. The rotational speed at the output was ~11 RPM.

The rotational speed could not be increased to the maximum 22 RPM due to safety considerations. This meant that some of the GMFs and rotational frequencies had to be shifted.

Figure 26 shows the vibration power spectrum using Fast Fourier Transform for the raw signal captured with only the gear damage present. The gear damage is evident at 9.7 Hz and a second harmonic is seen at approximately 19.4 Hz. The peak at 21 Hz is the high speed shaft frequency whilst the peak at 64 Hz is the intermediate GMF.

Figure 8 Vibration measurement gear tooth damage only location 1. Power spectrum low frequency.

Figure 27 shows the power spectrum for the raw vibration signal captured with both imbalance and gear tooth damage present at the same time.

The imbalance caused some rich frequency content to appear between 10-20 Hz but it is not clearly visible. On this occasion, the gear damage seemed to dominate the signal.

Thus, its effect is far more evident when the test rig is submerged since the water’s resistance increases the influence of the steel plate and hence imbalance manifests itself more strongly in the power spectrum.

Figure 9 Power spectrum for the vibration signal captured with gear tooth damage and imbalance present. The effect of the imbalance is less evident and the gear tooth dominates in this case.

Figure 28 shows the raw vibration data for impact defect present together with gear tooth damage. The impacts are clearly seen in the pk-pk values as spikes.

Figure 10 Raw vibration signal for impact and gear tooth damage. The impact events are clearly seen as spikes in the pk-pk.
Work Package Main Results – WP4
Partners Involved: WLB, Degima, Coservices, Stirling Dynamics, TWI, BIC, Engitec

Structural Health Monitoring Module
Wave Energy Converter device

A mock-up of a Wave Energy Converter (WEC) stopper was built according to the project requirements. The mock-up is similar to the real stoppers that are built by Degima Figure 11depicts the mock-up used during the experiments.

Figure 11 Mock-up of a Wave Energy Converter.

For the project, the WEC selected a spar buoy similar to the one shown in Figure 12. A floating device ideal for catching and converting wave energy is one popular technology used. Floating devices are more able to deal with the great wave power density found in offshore locations. Additionally, because the devices can be tethered to float idly on top of the surface of the water, there is less restriction when it comes to their placement, and large arrays of floating devices can be constructed all across the open seas.

The real structure built by Degima is shown in Figure 12. In this photograph, the main parts of the buoy are identified: heave plate, spar and float.

Figure 12 PB40 Buoy built at Santander’s port, Spain.

It is also important to note the existence of the external mechanical stoppers and the pneumatic/rubber blocking devices, acting on the floater and against the spar (Figure 12). There are several types of PTOs in this kind of WECs: mechanical (ie rack and pinion), hydraulic rams or linear generators, however all of them have a constraint in the stroke of the PTO with the purpose of limiting the vertical displacement of the floater in extreme events (survival sea conditions for instance). In the case of mechanical or structural failure with these components, the floater will move freely and the stroke of the PTO could be overcome with unforeseen consequences and a potential risk of catastrophic failure of the WEC. Therefore, the end stoppers are an essential component for the survivability of the complete unit.

Figure 13 Test bench designed by TWI.

Test bench
The structure was placed in a custom test rig manufactured to test the REMO structure. TWI has several fatigue test machines but due to the complexity and dimensions of the sample under analysis, the custom test bed shown in Figure 13 and Figure 14.

Figure 14 Test bench at TWI facilities.

Experimental Procedure
The test was first run at the end of June and was stopped at the beginning of August without neither signs of crack initiation nor propagation. The method used to check for crack propagation was a visual inspection with a magnifying glass and a solution of soap and water. The solution is spread across the area where propagation is expected and if propagation occurs bubbles should appear. The test can be divided into seven stages which is beneficial in understanding the procedures followed throughout the experiment.
Figure 15 Location of the sensors.

TWI has developed software to process the AE activity acquired during the experiment. The software implemented in LabVIEW filters data acquired daily. It saves all the files acquired by the AE sensors 2 and 3 which cross a specific threshold level in a folder called Group 1. In another folder called Group 2, all files gathered by the AE sensors 0 and 1 which cross a specific threshold level are stored. This threshold level is modified from 45dB to 70dB using 5dB steps.

The filtering process carried out to discard unwanted data sets reduced the initial 77309 files to just a few hundred that pass the acceptance criteria for AE information Figure 16, Figure 17 and Figure 18 show the results of the filtering process using different threshold levels as an input parameter.

Figure 16 45 dB Threshold.

Figure 17 50 dB Threshold.

Figure 18 55 dB Threshold.

The Figures above indicate the critical selection of a threshold level for the use of AE as a monitoring technique. Even though there was no macro evidence of crack propagation in the structure, Figure 18 shows a clear progressive AE activity pattern that can lead to micro crack initiation or de bonding of internal metallic crystals whose friction generates AE activity.

Machine Condition Monitoring Module
Acoustic emission module
The purpose of this module is to acquire the AE signals from the gearbox to carry out health monitoring. The software can acquire signals from four different AE sensors, however it is possible to record a single channel measurement if required by the user.

Figure 19. AE front panel.

Vibration Module

The vibration of the machine is captured and vibration spectra is processed, which provides alot of information about the condition of the component. Naturally, the spectrum, which gives such valuable information, must be obtained accurately.

The diagram below (Figure 20) is a simplistic explanation of how vibration data is acquired.

Time Acquisition= #Lines / Fmax # Samples= Sampling Rate x Time of acquisition

Figure 20. Acquisition of vibration data.

Complete Integrated system Tested with Laboratory Trials

Laboratory Test Description

The mock up dimensions are approximately 2.8m in height and 1.3m in depth. The drive train mock-up has been powered in the laboratory and different scenarios have been reproduced.

The motors were first run through a water check whether water is leaking through or not. The motors were able to work underwater after preparing the motor with appropriate sealant. This was used as part of WP5.

Figure 21 Drive Train Laboratory Tests.

Sensors Installation
In order to monitor the pair of gearboxes and motors, four AE sensor and four accelerometers were used. The picture below depicts where the sensors were placed.

Figure 22 Position of AE sensors and Accelerometers on both motors.

Data collection
As mentioned above, data has been collected for baseline signature and various defect scenarios. The various scenarios including baseline capture are described below.
In this case, baseline data was collected without defects introduced. This set of data was used as a signature with respect to the defect data, to enable defect detection and identification.. The test was carried out for one hour.

Gear Fault
On one of the gearboxes, the gears were damaged using a drill. This is illustrated below in Figure 23.

Figure 23 Gear Defect.

Test Rig Testing Outside Water Using ENGITEC’s Acquisition System and Analysis Software
Test rig tests on dry land were carried out at TWI’s facility in Granta Park Cambridge. For data acquisition ENGITEC’s condition monitoring system was employed. For the AE measurements four PAC R50A sensors were employed. Two 25 kHz Wilcoxon accelerometers were used for vibration measurements. The complete ENGITEC acquisition system is shown in the photographs in Figure 21.

Figure 21 shows the vibration power spectrum using Fast Fourier Transform for the raw signal captured with only the gear damage present. The gear damage is evident at 9.7Hz and a second harmonic is seen at approximately 19.4Hz. The peak at 21Hz is the high speed shaft frequency whilst the peak at 64Hz is the intermediate GMF.

Figure 24 Vibration measurement gear tooth damage only location 1. Power spectrum low frequency.

Figure 25 Spike energy of the above AE signal with no defects visible.
Work Package Main Results – WP5

Partners Involved: WLB, Degima, Coservices, Stirling Dynamics, TWI, BIC, Engitec

Field Trials

Tidal Turbine mock-up
The REMO mock-up has been designed to generate noise and vibration representative of a full scale tidal power system

3-phase motor/gearbox
Figure 26 shows a three phase motor connected to a three stage gearbox. This set is waterproof and hence used in trials underwater.

Figure 26 3-phase motor/gearbox.

The REMO mock-up PTO rotor has been designed to have sufficient drag to put the motor and gearbox assembly at rated power through still water. The rotor blades on the PTO rotor have been designed with apertures which can be opened by the removal of covers. This allowed the simulation of operation with a damaged or fouled blade, in order to test whether the REMO system is capable of detecting such a failure.

Water tank
Field trials have been carried out at TWI Ltd facilities in Middlesbrough. The tidal turbine mock up was submerged in a 2.8m diameter and 3m length cylindrical water tank.

Figure 27 Field trials in the water tank (TWI Ltd Middlesbrough).

Figure 28 Tidal turbine mock-up inside the water tank.

After having all the sensors attached to the gearbox, the cables were placed together to avoid tangling with the blades or pitch. Harnesses were placed behind each motor to balance the structure whilst it was being lifted and submerged. The mock-up was put into the water with the help of a rotating crane.

Experimental trials and results
Two AE and vibration sensors were attached to each gearbox in order to monitor the condition of the gearbox when it was operational underwater. Accelerometers were located on the output shaft bearing, one vertical and one horizontal. This location was selected as there must be a good mechanical path to the bearing as only vibration should travel along solid metal with no gaps or joints. AE sensors were located vertically and horizontally in the case of the gearbox, previous removing the coating and paints of the gearbox. Ultrasonic couplant was not necessary since water behaves as such.

Figure 29 Location of the sensors
a) Faulty Motor
b) Healthy Motor

Figure 30 Variation of the Maximum Value during Test 4.

The REMO system has been tested in the water tank located at TWI facilities in Middlesbrough. During field trials, the system was proven to work underwater. As shown in the last heading, the system was able to detect changes in the rotation speed and in the condition of the gearbox in both AE and vibration signals. Although crest factor value exhibits non expected behaviour, RMS, counts and information entropy values show high sensitivity for both speed variation and gearbox fault progression using AE signals. Using vibration signals, RMS is the best parameter as it does not increase with time like the maximum value and does not depend on the threshold value selected like counts parameter. In addition, RMS is also the most sensitive parameter as it increases 80% from the beginning of the test. The algorithm for fault detection developed during WP3 has been proven to work as well as the hardware developed during WP2 and 4. Thus, the REMO system has been validated both in dry and underwater environments. The instrumentation has been tested and after an initial operational issue in the underwater environment, additional work was carried out to achieve correct signal readings from the machinery.
Work Package Main Results – WP6

Partners Involved: WLB, Degima, Coservices, Tangent, Stirling Dynamics, TWI, BIC, Engitec
Each partner has provided a brief description for exploitable knowledge developed within their part of the project implementation.

For this purpose, the Exploitation Manager regularly reminded the consortium members to update information in this document.

This section gives an overview, of how the knowledge was disseminated.

Generally, knowledge was exploited through the following media:

• Project Web Site
• The Project Coordinator maintains the project webpage, which is accessible under The project webpage will continue after the project has concluded with the project webpage to maintain interest in the project.
• Project Flyer. The partners generated a flyer to disseminate the project to the potential end users.
• Awareness campaign. The SME partners have prepared awareness campaigns with SME service companies and end-users. This included:
 Press releases
 Brochures
 Presentations
• Project Video. The partners have prepared a video containing information about the project’s technical outcome in order to raise awareness of the tidal industry.
In this document all events that have taken place are listed.
 Conferences. The projects partners have organize/attend and will organize/attend national and international conferences to disseminate results of the project
In this document all events that are planned or have taken place are listed:
REMO User Manual

A detailed operational user manual was developed as part of WP 6, screenshots of the user manual developed are provided below:

Figure 31 Software configuration picture (From REMO operational manual).

Figure 32 Software configuration picture (From REMO operational manual).

REMO Project Video

The REMO project video can be reached in the project website.

Figure 33 REMO Logo.

Figure 34 REMO promotional video (Partners Logos).

Figure 35 REMO promotional video (Field Trials).

REMO Project Website

The project website has been widely used during the project duration as a communication and dissemination tool for the project partners:

Figure 36 REMO website main page.

Potential Impact:
Europe’s dependency on electricity produced from fossil fuels and nuclear power plants has remained unchanged in recent decades despite the introduction of large amounts of renewable energy predominantly produced from wind farms and photovoltaic parks and to a lesser extent, from biofuels and biomass. Due to the continuous growth in electricity needs of Europe, power production from renewable energy sources has primarily been used up to match the growing energy demand. Onshore wind energy, photovoltaic, biofuel and biomass projects have multiplied rapidly during the last thirty years. In 2012, the renewable energy sector was estimated to employ more than 1.2 million people throughout Europe. Moreover, the economic value of renewable energy for Europe was estimated at €130 billion or approximately 1% of the Gross Domestic Product (GDP) of the EU-27 .

Locations where onshore wind energy and photovoltaic projects are getting closer to saturation point in Europe. Repowering old wind farm sites using larger multi-MW wind turbines provides a possibility to achieve higher power outputs in the future, but the licensing procedure is not straightforward in most Member States. Photovoltaic parks have grown rapidly during the last decade and in some Member States the total power output is of comparable scale to that achieved by wind farms (eg Greece). However, like wind farms, photovoltaic face competition for land usage and the locations where large-scale parks can be constructed is diminishing. Offshore wind energy is an alternative to onshore wind farms which has received significant attention in recent years particularly in North-Western States. However, CAPEX and OPEX costs for offshore wind farms remain very high compared to other sources of energy production such as fossil fuels and nuclear. Onshore wind farms have an average construction cost of €1200-1800 per kW installed. Offshore shallow water (depth <30m) wind farms on the other hand have an average construction cost of €2500 3000 per kW installed which is two to three times higher than that of onshore wind farms . For the floating wind turbine projects demonstrated to date construction costs of €10,000 20,000 per kW installed have been reported . It is obvious that there is a strong need to considerably reduce the construction and installation costs of floating wind turbines in order to make them financially viable in the medium to long term. The ocean energy sector has an initial target construction cost at this stage of development of €4000-5000 per kW installed.

Tidal and wave power production is still in its infancy in Europe with very few projects currently on-going. Most of the projects are small-scale although there are plans for utility sized plants in the short-to-medium term . Figure 1 shows the milestones for tidal and wave energy according to the SI OCEAN project initiative .

The EU has set very ambitious objectives to reduce the total greenhouse emission of Member States by 80-95% of the 1990 equivalent . It is obvious that renewable power production from wind energy, photovoltaics, biofuels and biomass alone cannot achieve this target. It is therefore imperative that the tidal and wave energy contribution to the European energy mix becomes more evident. The European Commission estimates that by 2050 tidal and wave energy production could reach up to 100GW . On a global scale the International Energy Agency (IEA) estimates that the potential of ocean energy could reach up to 748GW of installed capacity whilst by 2030 up to 160,000 jobs could be created directly and up to 5.2 billion tonnes of CO2 could be saved by 2050 .

Figure 37 Milestones for tidal and wave energy industry up to 2050 [Ref. Sian George Presentation on Vision Paper & SI OCEAN].

Tidal and wave energy production offers several advantages over other conventional renewable energy sources as it does not require access to valuable land resources and can produce electricity continuously at a predictable rate. There is a substantial potential of tidal and wave energy available in Europe which is yet to be exploited but CAPEX and OPEX remain unacceptably high in order to be sufficiently competitive .

According to the Industry Vision Paper published by the European Energy Association9 in 2013 tidal and wave energy has started growing at a steady pace. Over the last seven years more than €600 million has been invested in tidal and wave energy by the private sector but public financing has not been satisfactory so far. Nonetheless, within the EU, the installed capacity tripled between 2009 and 2013 with more than 10MW of installed capacity, predominantly related to demonstration and pilot systems.

Some of the devices installed have a rated capacity of over 1MW. Approximately 2GW of potential development projects are in the planning stage within the EU by various operators. The growth of tidal and wave energy is based on a Three Point Plan as described in the same paper and shown in Figure 22.

Figure 38 Three Point Plan employed in order to unlock Europe’s tidal and wave energy potential7.

With both wind energy and photovoltaic power nearing saturation and the growth of electricity production from biofuels in doubt, alternative renewable energy sources which have not been sufficiently exploited commercially so far but have a proven potential will gradually grow in significance. The European Commission has highlighted ocean energy as the key for the European energy mix in the 21st century. The contribution from ocean energy in the reduction of greenhouse emissions was critical in achieving the targets set for 2050. Figure 23 shows the immediate objectives for ocean energy projects by 2020 within Europe. The Member States who currently are actively pursuing such projects are the UK, Ireland, France, Spain and Portugal. Denmark has also expressed a strong interest but no projects are planned by 2020. Of course, other Member States are also looking into ocean energy but with less specific objectives at this stage including, Greece, Germany, Italy, Croatia, Netherlands and Belgium. Once tidal and wave power converter technologies become more mature, it is almost certain that the aforementioned countries will follow closely in the exploitation of ocean energy.

Figure 39 Member Stated looking into exploiting ocean energy by 20207.

Figure 24 shows some of the key industrial companies who are likely to play a critical role in the exploitation of ocean energy in Europe in the short to medium term.

Figure 40 Some of the industrial entities who are likely to be heavily involved in the exploitation of ocean energy in the short to medium term7.

The strategic risks that need to be addressed relate to the affordability, manufacturability, instability, operability, survivability, reliability and predictability of ocean energy projects. All these risks need to be mitigated in order to allow ocean energy to take off and become sufficiently competitive to other energy sources, renewable or not.

Figure 25 shows the number of devices currently being researched in the UK, under scaled down demonstration, full scale demonstration and first farm development along with the associated costs .

Figure 41 Ocean energy activity within the UK10.

Figure 26 shows the current global activity with respect to ocean energy development10.

Figure 42 Ocean energy activity around the world10.

Some researchers (eg Prof. Gregorio Iglesias Rodriguez of Plymouth University ) have proposed potential methods of reducing the cost of tidal and wave energy projects further by taking advantage of synergies with other technologies. For example, tidal devices could potentially be combined with offshore wind turbines to reduce the overall cost and take advantage of a single electrical infrastructure, hence reducing overall CAPEX and OPEX for both projects. Moreover, such integration could reduce financial risk by increasing the probability of the timely Return of the Investment (ROI) and increasing financeability (
Tidal turbines share several similarities with wind turbines. Tidal turbines, like wind turbines can be of the gearless or geared types. Most of existing designs are of the geared type as they are generally cheaper, lighter and use fewer magnets. In a similar fashion to wind turbines, tidal turbines operate in a turbulent environment. This means that tidal turbine gearboxes are required to be capable of operating under a very harsh variable load environment apart from the fact of being submerged underwater several metres below the surface of the sea. Long-term experience with large-scale multi-MW wind turbines has shown that variable loading cause a high level of wear and fatigue in gearbox components. Gearbox problems are one of the most important causes of downtime and very few wind turbines reach the end of their design lifetime without having changed a gearbox at least once during this period. Tidal turbines face even harsher operational conditions, meaning gearbox problems are likely to develop over prolonged in-service operation.

Energy is the lifeblood of the European economic growth. In the complex and highly competitive environment of energy production, the renewable energy sector, particularly the wind energy industry, is the one energy sector which stands out in terms of ability to reduce greenhouse gas emissions and pollution, exploit local and decentralised energy sources, stimulate world-class high-tech industries and support the sustainable economic growth of the region. The EU has compelling reasons for setting up an enabling framework to promote wind energy. It is largely indigenous, it does not rely on uncertain projections on the future availability of fuels and its predominantly decentralised nature makes the European economy and prosperity less vulnerable to external factors associated with the reliable supply of imported fossil fuels. It is undisputed that wind energy and renewable energies in general will constitute a key factor for Europe’s a sustainable future. For those reasons, the European wind energy industry is currently exhibiting rapid growth rates in an effort to meet the increasing environmental, societal and economic demands set for modern power generation production within the EU.

Wind energy currently provides approximately 8% of the overall European electricity production. However, with the current growth trends exhibited, the wind energy industry is capable of delivering 12% of the overall European power production by 2020 and more than 20% by 2030. On the other hand tidal energy is still in its infancy but with an excellent potential for growth.

Employment in the European tidal energy sector is expected to grow rapidly over middle to long term with both direct and indirect employment in manufacturing, installation and maintenance. It is evident that the importance and level of influence of the tidal energy sector to the European economy and society will depend on the level of investment in this sector over the next two decades.

Nonetheless, there is a growing need to stimulate further technological progress in tidal turbine design which will lead to improvements in operational reliability in order to further promote the growth of tidal energy and consolidate its future as one of the key renewable energy sources along with wind energy, photovoltaic and concentred solar power.

The cost of the power generated by tidal turbines is mainly associated with the average annual power output (or capacity factor) and the availability of the tidal turbine itself. The power output of a tidal turbine is not constant but is directly linked with the flow speed of the water and the capacity factor of the tidal turbine.

Tidal turbine reliability is a critical factor in the success of tidal power generation. Poor reliability directly affects both the revenue stream through increased operation and maintenance costs and reduced availability to generate power due to turbine downtime. Indirectly, the acceptance of tide-generated power by the financial and developed communities as a viable enterprise is influenced by the risk associated with the capital equipment reliability. Therefore, an increased level of risk associated with the efficiency of tidal energy production is generally accompanied by increased finance fees or interest rates.

Operation and maintenance costs will constitute a sizeable share of the total annual costs of a tidal turbine. For a new machine, O & M costs have an average share over the lifetime of turbine of more than 30% of the total cost per kWh produced. Therefore, O & M costs are increasingly attracting attention of manufacturers seeking to develop new designs requiring fewer regular service visits and less downtime. In order for the tidal energy industry to achieve the growth targets set by the EU for the forthcoming years, tidal turbine numbers have to increase substantially, while at the same time their operation and maintenance costs will need to be reduced by a noticeable factor to permit the European tidal energy industry to retain its advantage over growing overseas competition. Therefore, one of the highest priorities for the European tidal energy industry is currently the significant improvement of the reliability of tidal turbines, involving a solid reduction in current inspection and maintenance costs mainly associated with unpredicted failures of critical components.

The simultaneous increase in tidal turbine numbers and their size, in combination with the remote locations where tidal farms are typically built offshore, means that maintenance costs will see an unprecedented increase in the future, harming the long-term growth prospects and sustainability of the European tidal energy industry as a whole unless proper action is taken on time.

The average cost of the electricity produced by tidal turbine operators will also be a factor which will directly influence the extent to which the tidal energy industry will manage to penetrate the European and global power generation market.

To address these pressing problems, the tidal energy industry will have to successfully implement new and efficient maintenance strategies and inspection methodologies which will reduce (to an absolute minimum) equipment-related failures and associated costs. The application of the REMO methodology in the field could potentially lead to an overall decrease of up to 60% in inspection costs and substantially reduce the need for corrective maintenance. The widespread application of REMO on tidal turbines will also permit the evolution of maintenance strategies towards condition-based and predictive maintenance strategies, leading to the optimisation of maintenance logistics, particularly in relation to the appropriate allocation of available maintenance personnel, equipment and financial resources.

The widespread implementation of the systems and techniques developed within REMO will assist the European tidal energy industry to significantly improve its reliability record by delivering the technology required in order to minimise tidal turbine failures and practically eliminate the need for corrective maintenance resulting from them. The achievements and results produced within the REMO project will have a significant number of positive impacts as it will contribute to the improvement of the tidal energy industry with regards to its: a) reliability and efficiency, b) competitiveness in comparison to other power energy sources and c) sustainable growth.

REMO has the potential to become a significant component on the road towards achieving the efficiency, reliability and business targets set for the European tidal energy industry and will help push forward the European Roadmap on Renewable Energy sources even further.

The application of an advanced condition monitoring such as REMO will lead to a significant reduction in the number of on-site visits carried out by maintenance engineers in order to inspect wind turbines for damage. This is of particular importance where the offshore location of tidal turbine farms are concerned due to the fact that adverse weather conditions can delay the conduct of any scheduled inspection or maintenance for several weeks.

REMO will allow several environmental benefits to be drawn by supporting the growth of the tidal energy industry. It will assist in the minimisation of failures and associated costs while it will considerably improve the reliability of wind turbines. Moreover, this project will contribute to the optimisation of the efficiency of preventive maintenance thus extending the operational lifetime of assets and reducing the need for new components which contribute to the depletion of valuable natural resources.

The rise in greenhouse gas emissions from energy is unsustainable. By 2030, global greenhouse emissions could more than double due to rising use of fossil fuels, notably in developing countries. At the same time most climate experts suggest that carbon dioxide emissions need to be halved if the worst impacts of climate change are to be avoided. The IPPC states in its Fourth Assessment report that CO2 emissions will need to reach their peak by 2015 at the latest and immediately start declining for the world to stay below a two degree Celsius increase in average temperature. New nuclear fusion and carbon capture will not be available within that timeframe. Therefore, the need for a long-lasting solution that is environmentally benign, economically sound and can be put quickly and efficiently into place is more urgent than ever. Tidal energy fills all these criterias and it could become one of the most important tools that Europe possesses in decarbonising power generation while maintaining the economic growth and prosperity of the region.

Apart from the fact that it does not emit any carbon dioxide emissions, Ocean Energy does not deplete natural resources in the way that fossil fuels do, nor does it cause environmental damage through resource extraction, transportation, or waste management.

List of Websites:
More information can be found in the REMO project website
For more information about the project contact:
Dr Slim Soua, Project coordinator
Granta Park, Great Abington
CB21 6AL
+44 (0) 1223 899000