Electrostatic Monitoring Syste
http://www.CESI.com
Electrostatic Monitoring Syste
Gas Turbine Exhaust Gases Electrostatic Monitoring System
The measurement of electrostatic charge associated with the particulate transported by exhaust gases of a gas turbines or jet engines has been successfully applied in the aeronautical field as an early fault diagnosis tool for the possible degradation of internal “hot parts” (rubs, wear in blade coating and thermal barriers, hot spots in combustors, improper combustion etc.) The measuring technique is usually known as EDMS (Engine Distress Monitoring System) or as EEMS (Electrostatic Engine Monitoring System).
More.
WORKING PRINCIPLES AND GENERAL CONSIDERATIONS
The ELMO system is based on the monitoring of the electrostatic charge associated with particles produced by possible degradation of the internal components of gas turbines, which are transported into the exhaust gases; these particles are ionised due to the high temperatures produced in the combustion processes and degradation phenomena.
More.
HARDWARE CONFIGURATION
Apart from its sensors, ELMO system hardware consists of the same number of charge amplifiers, a signal conditioning unit, an acquisition unit and a data elaboration/display/storage unit.
More.
SOFTWARE CONFIGURATION AND DATA ANALYSIS
System management software was completely developed by CESI and mostly consists of the online acquisition/storage and display data programmes which communicate with each other and with remote units connected to the network by means of ad hoc developed interface software modules; the supplied software is complete with an off-line data elaboration programme.
More.
ELMO SYSTEM INSTALLED ON OPERATING GAS TURBINE
Preliminary test campaign, carried out on a full scale test-rig for experimentation of new combustors, allowed the assessment of sensitivity of the method to the release of particulate and the definition of the specifications, both software and hardware, necessaries to perform a reliable system for on-line monitoring of the electrostatic charge.
More.
CONCLUSIONS
Applicability and potentiality of the measurement of the electrostatic charge associated to exhaust gases as a method for monitoring distresses of internal parts of industrial gas turbines, was successfully evaluated by an extensive demonstrative programme, which started with preliminary tests carried out on an experimental combustor test-rig and carried on with the application on a significant number of gas turbines, different in rated power and constructive design, operating in power plant stations.
More.
The measurement of electrostatic charge associated with the particulate transported by exhaust gases of a gas turbines or jet engines has been successfully applied in the aeronautical field as an early fault diagnosis tool for the possible degradation of internal “hot parts” (rubs, wear in blade coating and thermal barriers, hot spots in combustors, improper combustion etc.) The measuring technique is usually known as EDMS (Engine Distress Monitoring System) or as EEMS (Electrostatic Engine Monitoring System).
More.
WORKING PRINCIPLES AND GENERAL CONSIDERATIONS
The ELMO system is based on the monitoring of the electrostatic charge associated with particles produced by possible degradation of the internal components of gas turbines, which are transported into the exhaust gases; these particles are ionised due to the high temperatures produced in the combustion processes and degradation phenomena.
More.
HARDWARE CONFIGURATION
Apart from its sensors, ELMO system hardware consists of the same number of charge amplifiers, a signal conditioning unit, an acquisition unit and a data elaboration/display/storage unit.
More.
SOFTWARE CONFIGURATION AND DATA ANALYSIS
System management software was completely developed by CESI and mostly consists of the online acquisition/storage and display data programmes which communicate with each other and with remote units connected to the network by means of ad hoc developed interface software modules; the supplied software is complete with an off-line data elaboration programme.
More.
ELMO SYSTEM INSTALLED ON OPERATING GAS TURBINE
Preliminary test campaign, carried out on a full scale test-rig for experimentation of new combustors, allowed the assessment of sensitivity of the method to the release of particulate and the definition of the specifications, both software and hardware, necessaries to perform a reliable system for on-line monitoring of the electrostatic charge.
More.
CONCLUSIONS
Applicability and potentiality of the measurement of the electrostatic charge associated to exhaust gases as a method for monitoring distresses of internal parts of industrial gas turbines, was successfully evaluated by an extensive demonstrative programme, which started with preliminary tests carried out on an experimental combustor test-rig and carried on with the application on a significant number of gas turbines, different in rated power and constructive design, operating in power plant stations.
More.
ntroduction
The possibility of extending this monitoring technology to industrial gas turbines, which are equipped with larger exhaust ducts, run with different fuels and work in different operating conditions respect to aeronautic engines, required specific experimental verifications and confirmations. This article describes a synthesis of data and the most noteworthy experiences collected by CESI in a large demonstrative programme carried out in field, on industrial gas turbines. Response and sensibility of a developed instrumental monitoring system, based on the above mentioned method, to the machine’s working conditions and to various possible degradation mechanisms, has been evaluated. A consistent number of tests were undertaken, initially in a full scale combustor test rig and then directly on gas turbines of different construction and rating power, operating in electrical power plants, where it has been possible to compare the indications given by the electrostatic charge measurements with the results of inspections on the machines during programmed or following the event of faults maintenance. Collected data confirm this method’s potential advantages as it provides a warning on the existence of degradation phenomena in internal components much earlier than the conventional monitoring techniques such as those based on vibration analysis or “gas path” performance techniques. Also described is the improved hardware and software of the ELMO (Electrostatic Monitoring) system developed by CESI, which permit continuous monitoring of industrial gas turbines. Based on results obtained up to now, there is definition of the work to be done so that the method can automatically and punctually recognise the main types of faults and distresses of these machines.
WORKING PRINCIPLES AND GENERAL CONSIDERATIONS
The force lines of the electrostatic field generated by travelling charges interact with one or more special sensors mounted on the exhaust pipe walls and induce opposite charges which may be detected and converted into a signal that is useful for characterising the particulate’s presence. The charge sensors are passive elements able to operate in high working temperatures, mounted facing each other inside the exhaust pipe wall in a non-invasive manner using appropriate adaptors; number and position of sensors are chosen to provide a complete visibility of the whole section of gas path. The table and Figure show the main characteristics of the sensors made by CESI.
Experience has demonstrated that any gas turbine in healthy state produces a basic level of electrostatic charge, which varies with the working conditions, the type of fuel, the combustion stability, the climatic conditions and the normal deterioration and wear of internal components. This background charge level is used to define the reference “baselines” (charge levels normalised to the machine’s working and operating conditions) and the “acceptance thresholds” with respect to which “excursions” in charge levels related to anomalous increase of particulate in exhaust gas are highlighted. Rubs between machine’s rotating and stationary parts, leaks of oil from the bearing seals, hot spots in the combustion chambers, loss of blade coating and thermal barrier, increase of combustion unburned, seal abrasion, intake or loss of gas path internal component parts are all typical sources of “debris” which generate increases in the charge levels. Obviously not all particulate production are correlated to machine malfunctions. For instance, online compressor washing or activation of fogging systems for inlet air cooling can assist removal of deposits along the fluid path and produce increases in charge levels with no relation to malfunctions. This type of event must be identified and recognised in order to avoid false alarms. The main advantages of the on-line electrostatic charge monitoring method applied to industrial gas turbines respect to other conventional methods (based for instance on machine performance analysis or temperature spreading in turbine exhaust) are the use of a limited number of sensors (one or two), the early warning, the direct monitoring of internal component deterioration and the real-time evaluation of gas path conditions.
HARDWARE CONFIGURATION
Apart from its sensors, ELMO system hardware consists of the same number of charge amplifiers, a signal conditioning unit, an acquisition unit and a data elaboration/display/storage unit. The charge amplifiers are variable gain amplifiers, equipped with band-pass filter, housed in a protective casing located maximum 5-6 m apart from the sensors: these devices transform the original charge signal into a voltage signal suitable amplified to be transported with a coaxial cable up to a maximum distance of 150-200 m. The charge signals and other engine data signals which identify the machine’s working conditions (basically the 1 x revolution signal from the turbine rotor and the power delivered by the turbine) are feed in the signal conditioning unit, equipped with optoinsulator modules which carry out galvanic insulation. These modules interface with the 16 bit DSP acquisition board unit (max acquisition frequency up to 153,6 kHz), equipped with 8 single-ended analogue input channels, 8 anti-aliasing analogue filters and 8 digital filters, software programmable. The board is feed into one expansion slot of an industrial PC which sets the on-line data elaboration/display/storage unit. Monitoring of a second gas turbine within the same plant is possible by an additional acquisition board feed into the same PC. With the exception of the charge amplifiers, all the other ELMO hardware components are housed on a rack of an instrumental cabinet which is generally located in the station’s computer room.
SOFTWARE CONFIGURATION AND DATA ANALYSIS
System management software was completely developed by CESI and mostly consists of the online acquisition/storage and display data programmes which communicate with each other and with remote units connected to the network by means of ad hoc developed interface software modules; the supplied software is complete with an off-line data elaboration programme.
The specialist sets the system acquisition parameters (sampling, signal sensitivity, alarm thresholds, data storage rate, etc.), defines the reference baselines for the charge signals based on the analysis of data acquired for a time period considered sufficient to characterise the machine’s behaviour, sets the communication protocols between the various software and finally sets the on-line data display pages for the operator. As shown in Fig. 6, data relative to several machines equipped with ELMO systems can be displayed on the same display unit. Following suitable filtering and analogue-digital conversions, the system carries out a calculation of the following fundamental components of signals coming from the charge amplifiers:
activity level (AL): expressed in pC represents the high frequency content of the charge signal and gives an indication of the amount of fine particulate (particles of dimensions less than 40 µm) present in exhaust gases;
event rate (ER): expressed as a dimensionless percent quantity, measures the unfiltered original data exceeding, in a given time window, established threshold levels; this component relates to the number of large particles (approximately bigger than 40 µm) present in the exhaust gases at any time; events are separated according to amplitude and polarity (positive/negative events of first/second threshold);
shaft order analysis: charge level components that are synchronous or entire multiples of the rotational speed of the gas turbine highlight the release of particles at frequency correlated to machine rotation (for instance blade rubs).
The instantaneous values of these components perform the system’s first level analysis, while the calculation of the time averaged values, normalised to the machine’s operating conditions, perform the second level analysis. From a consistent and significant amount of averaged data, collected along all possible different working conditions of the healthy gas turbine, a suitable algorithm calculates the reference “baselines” (see Fig. 6) for the charge signals that represent the intervals of statistical variation considered as acceptable; when thresholds are surpassed the system sends automatic (graphical or acoustical) alarm signals to the plant operator. In particular, alarm situations are distinguished by “excursion” (average levels of fundamental quantities which surpass the reference baseline threshold) related to phenomena of progressive degradation of the machine’s internal components and by alarm situations for “single event” (instantaneous levels of fundamental quantities which surpass threshold level limits) related to sudden and more unexpected degradation phenomena (as, for instance, the extreme case of loss of blade).
ELMO SYSTEM INSTALLED ON OPERATING GAS TURBINE
Preliminary test campaign, carried out on a full scale test-rig for experimentation of new combustors, allowed the assessment of sensitivity of the method to the release of particulate and the definition of the specifications, both software and hardware, necessaries to perform a reliable system for on-line monitoring of the electrostatic charge. CESI developed the ELMO system and continued the validation activities on full scale gas turbine in service in electrical power station. Having the double objective of consolidating and perfecting the software algorithms for collection, analysis and automatic storage of data and to increase the cases of machine events considered to be of diagnostic importance, the monitoring activity has been extended to some industrial gas turbines of different size (power rating from 45 MW up to 240 MW) and construction type (aeroderivative and heavy-duty gas turbines). Larger machines (power rating higher than 120 MW) normally need the installation of two charge sensors (immediately downstream the last stage of rotating blades of the turbine) in order to gain visibility of the entire transversal section of gas path: that is both due to the larger size of exhaust duct and to the existence of constructive elements within the duct itself (hot bearing supports, as show in the following figures).
Internal view of a exhaust duct Installation on 120 MW gas turbine
During the first installations, assessment of sensitivity and calibration of the instrumental measurement chains were carried out by means of forced injection of powders (of known weight and fineness) of different materials (metallic and ceramic) usually used for the construction of the hot parts of the gas turbines (see the following figures).
Forced injection of particulate and system response
The considerable data base collected from a monitoring extended to a medium-long time periods has permitted to verify the reliability and the functionality of the hardware and software components, improved and optimised along the experimentation, and also to characterise the different gas turbines.
Generally speaking it has been pointed out that the background levels of electrostatic charge and their sensitivity to the variation of the working conditions change according to machine type and size and are affected by the combustion parameters setting up, the fuel composition and the climatic conditions.
Activity level calculated at rated load on Siemens V64.3A (left) and Fiat 701F (right) gas turbines
The above figires show, for instance, the time variation of activity levels (labelled as ELMO sx or EDMS dx, where dx and sx indicate the right side or left side of the exhaust duct), superimposed on load signal levels, as detected on a 120 MW and 240 MW power rated gas turbines along the full load operation; the activity levels highlight physiological variations related to combustion parameter variation setting up. The figure below (left) shows an example of increasing in the rate of events (labelled as CNT-2A) induced by the start up of a plant “fogging” system in a 120 MW power rated gas turbine. The figure below (right) on the other hand, illustrates the increase in activity levels measured on a 45 MW power rated gas turbine correlated to the variations in the fuel mixture. These latter cases constitute examples of possible and important “physiological” variations in charge levels which are to be recognised and distinguished from the “pathological” ones in order to avoid false alarms.
Effects of “fogging” on a 120 MW rated power gas turbine Effects of fuel mixture variation on a 45 MW power rated gas turbine
Along the monitoring period, a considerable number of “diagnostic” cases are also pointed out by the system, which revealed the effective potentialities offered by the electrostatic charge method for an early warning of incoming anomalies, all of which are outlined with difficulty or not quickly with alternative conventional monitoring techniques. For instance,in the follwoing figures are compared the charge signals components calculated along two different start-up transients registered on the same machine (a 240MW rated power gas turbine); in the second start-up there is a notable increase in activity levels, especially in the sensor located on the right side of the exhaust duct (green line in figure right): shaft order analysis pointed out the fundamental contribution of 1X component (synchronous with the rotating speed) to the increase of the total charge signal level and allowed diagnosis of a partial rubs of rotating blades on stationary parts, which was confirmed by inspection carried out on the machines during the stop for maintenance (figure in the middle). The presence of “partial” rub marks mainly on the abrasive tiles located to the right side of the first stage of the turbine was coherent with the response of the two charge sensors mounted on the two sides of exhaust duct.
Normal start-up on 240 MW gas turbine Anomalous start-up with rubbing
Signs of “partial rubbing”
Another example of “ pathological ” distresses detected by the ELMO system on a 45 MW rated power gas turbine is shown in the following figure. The very low and constant activity level normally recorded at full load, started to increase. Following the warning given by the ELMO system, the plant operators planned a short outage of a GT a few days later and performed an endoscopic inspection which outlined the existence of abnormal erosion of the thermal barrier on combustor chambers and coating on the turbine blades (figures below).
Activity level on 45 MW rated power gas turbine (model LM6000PB)
CONCLUSIONS
Transferring this technology from aeronautical to industrial gas turbines has given rise to complete revision and engineering both of hardware components (sensors, signal conditioners and data acquisition systems) and management software of the instrumental system; this is due to the differences in construction, functionality and operating management in the two cases. In-field experimentation has permitted verification of the reliability of the instrumental apparatus and their suitability in actual working conditions for continuous and long-term operativity. The large data collected, on the other hand, allowed optimisation of data analysis methods. In particular, there is the automatic data processing for estimate the reference “baselines”, that is, the charge levels associated with exhaust gases considered as “physiological” and normalised to the working conditions of the machine, and for setting up the warning signals for “anomalous” events, based on suitable threshold levels estimated on the basis of statistical methods.
Some of the gas turbines equipped for the experimental campaign have developed typical distresses phenomena (blades rubs, anomalous thermal barrier erosion in the combustion chambers, loss of coating on blades, damages on the turbine rotating blades) punctually pointed out at the early stage by the ELMO system and later confirmed by the inspection controls. The system can presently establish whether the ionised particulate’s source is located in the stationary or rotating parts of the gas turbines, but it is not able to identify with certainty the specific component or machine part which produces during its degradation or damage the signalled increasing in charge level. In order to improve this functionality, it is necessary to have a larger and more diversified fault events data record than the present one; to obtain this, further efforts must be made to extend the application of the method to a higher number of machines, in order to shorten the validation time. Monitoring of the electrostatic charge remains, however, a valid system to provide the plant operator with an early warning on possible development of distresses in the gas path, and opportune for confirming this is a comparison with generally slower and less punctual signalling supplied by other monitoring systems of a conventional type usually installed on the machines (vibration analysis, exhaust temperature and gas path analysis).
