Thermal Power Plant Condenser Fault Diagnosis Using Coordinated Condition Monitoring Approach

The increasing need for electricity has brought about the importance of maintaining power generating resources on a higher priority in India. Among the various resources, the thermal power plant is the primary resource of electricity generation. It is essential to maintain the thermal power plant continuously in an operating state. Unfortunately, this is not the case because the failure of equipment is inevitable even though it can be minimized by implementing suitable maintenance strategy. Garg [1] presented a methodology for analyzing the performance of industrial systems using uncertain data. The reliability, availability, and maintenance of thermal power plant have become more significant in recent years due to the growing demand for electricity from society. The optimum reliability and availability level are desirable not only to reduce the overall cost of production but also to reduce the risk of hazards [2]. The basic reliability and operational reliability are forecasted on the basis of the forecasted value of per unit contained in the model [3]. However, with thoughtful consideration for reliability, availability, and maintainability, the frequency of failures and similar consequences can be reduced considerably. Eti et al. [4] claimed that, if the attention is given during maintenance planning regarding the maintenance needs of the system, considerable savings are achieved in operational processes. 

As, the thermal power plant is a complex system consisting of various subsystems connected either in series, parallel, or mixed configuration. Failure of any subsystem may cause its unavailability, which will affect the performance of the system/unit/plant. Kumar et al. [5] developed the availability simulation model for power generation system of the thermal power plant and investigated the performance of the system under realistic working conditions. Dev et al. [6] stated that criticality level decides the importance of the system as well as a choice of appropriate maintenance and repair strategy so that the reliability and availability may be maintained up to the mark. According to Carazas et al. [7], reliability can be at least estimated during the plant design stages because its availability is strongly influenced by major factors such as the ability to diagnose the cause of the failure or the availability of equipment and skilled personnel to carry out the repair procedure.

The several approaches/tools have been used by earlier researchers to evaluate the reliability and availability parameters. Garg et al. [8] solved reliability redundancy allocation problem for non-linear source constraints. In their study, the artificial bee colony is developed, and further, the solutions are obtained with improvements. In addition, Garg [9] extended their study for obtaining the optimal solution of reliability redundancy allocation problem by using penalty based cuckoo search with non-linear constraints. Kumar et al. [10] classified the major systems of the thermal power plant into five different categories viz. boiler (boiler furnace and steam drum), boiler air circulation system, water circulation system, coal supply system and power generating system (steam turbine and a generator unit). In case of such condenser subsystem, condensate leaving the condenser is supply to the low-pressure heater, boiler feed pump through condensate extraction pump, which increases the pressure. Then, the temperature of the water is increased by passing it through a high-pressure heater and economizer. The boiler produces steam and stored in boiler drum which will further supplies to a steam turbine for electricity generation. Therefore, condenser subsystem is one of the essential prime systems of the thermal power plant. The reliability of condenser subsystem needs to maintain a high level.

In order to analyze the performance of the condenser subsystem for attaining the high availability and reliability, the reliability characteristics need to be studied in detail. This study reported a reliability analysis of the condenser system of Dahanu Thermal Power Plant (DTPP). In order to study the behavior of the system, the reliability of the condenser system at different time interval is determined. In addition, this study proposes a coordinated condition monitoring approach for fault detection of critical equipment of the plant. The approach is based on integrating three condition monitoring techniques, which includes vibration analysis, noise measurement, and ultrasound level. Through this study, the fault of the pump driving end bearing of the auxiliary cooling water pump (CWP) is diagnosed.  The details of the work done are discussed in subsequent sections.

The structure of the paper is organized as; Section 2 introduces reliability analysis for the condenser system, Section 3 provides details of coordinated condition monitoring approach based case study. The results and discussion are described in Section 4. Further, Section 5 highlighted the findings of the study.

The reliability and availability of the thermal power plant can be maintained high by adopting condition monitoring techniques for diagnosing the fault before the system breakdown. Bhargava et al. [12] predicted the reliability of thyristor using artificial intelligence techniques. The fault detection, fault diagnosis, and making maintenance-related decisions are made in the process of the condition monitoring program [13]. In recent years, various effective condition monitoring techniques have been developed, which includes vibration analysis, acoustic emission monitoring, wear debris analysis, thermography, temperature analysis, ultrasonic monitoring, non-destructive testing, visual inspection, motor condition monitoring, and motor current signature analysis. Use of such condition monitoring based maintenance strategy has not only been widely recommended by earlier researchers but also adopted by industries at large.

The studies reported by earlier researchers have intended to use integrated condition monitoring techniques for fault diagnosis [14]. Jagtap et al. [15] proposed an algorithm for fault identification of fault occurred in thermal power plant equipment using coordinated condition monitoring approach. The major reason to use such a program is to increase the accuracy of fault detection. Figure 2 shows the coordinated condition monitoring approach which contains vibration analysis technique, noise measurement, and ultrasound measurement. 

2.jpg

Figure 2. Coordinated condition monitoring approach

According to the approach, the current health of the selected system can be known by collecting and analyzing the condition monitoring data in the form of vibration parameter level, a noise parameter, and high-frequency ultrasound level.  If the monitored parameter exceeds the defined reference level, then the more attention is given on a high priority basis of such a critical system. By integrating all relevant information of process parameters of the selected system, the fault of the system is detected and diagnosed. The maintenance-related decisions are made, such as either to repair the failed component or to replace the component. In this study, the coordinated condition monitoring approach is adopted for diagnosing the fault of the auxiliary cooling water pump of the condenser system. The detailed study is discussed in the next section.

3.1 Experimental set up

The data acquisition system was used for collecting the condition monitoring data, which is presented in Figure 3. The accelerometer was placed on diving end (DE) and non-driving end (NDE) of motor bearing for ACWP 1B. The Fast Fourier transform analyzer recorded the vibration data (i.e., velocity and acceleration parameter) of ACWP 1B. The vibration measurements were taken along three directions, i.e., along the horizontal direction, vertical and axial direction.

The acceleration trend of ACWP 1B was analyzed to check the deviation in the vibration parameter, which is as shown in Figure 4.

It was observed that the acceleration trend of pump NDE in the horizontal direction was increased from 0.4 g level to 1.6 g level. The root causes which are affecting the acceleration parameters need to be carried out to avoid any future malfunction of ACWP 1B. At the same time, the overall vibration level (velocity parameter) of ACWP 1B were compared with reference standards and checked for threshold level. Figure 5 (a) to (h) shows the vibration spectrum before the maintenance obtained for velocity and acceleration parameter along the horizontal, vertical, and axial direction.

3.jpg

Figure 3. Data acquisition system used for ACWP 1B

4.jpg

Figure 4. The trend of ACWP 1B (acceleration)

51.jpg

52.jpg

53.jpg

54.jpg

55.jpg

56.jpg

57.jpg

58.jpg

Figure 5. Vibration spectrum before the maintenance

It was observed from Figure 5 that the velocity and acceleration parameters for pump DE bearings were increased to a significant level and recorded as 5.23 mm/sec and 1.5g respectively, which is tabulated in Table 4.

Table 4. Vibration parameters of ACWP 1B

It is noted that the threshold level of vibration parameters of ACWP 1B was closer to the standard reference level (i.e.7 mm/sec). Therefore the fault of the system needs to identify and diagnosed correctly. The next section discussed the fault diagnosis using coordinated condition monitoring approach.

3.2 Fault diagnosis using coordinated condition monitoring approach

The coordinated condition monitoring approach assisted in identifying the fault of the system. The approach identifies the fault of the system accurately. Moreover, it provides a base for equipment maintenance decision. Three condition monitoring parameters are monitored using three techniques, which includes vibration monitoring, noise monitoring, and ultrasonic sound measurement.  The acceleration spectrum of pump DE bearing was obtained using vibration analysis technique, which is as shown in Figure 6. Increased vibration parameters revealed that the fault of bearing has occurred. Therefore it is necessary to take corrective action before the system stop working.

6.jpg

Figure 6. Acceleration spectrum of Pump DE bearing

The major reasons behind the increasing trend of bearings are that most common rolling element bearing faults include damage of the inner race, the outer race, and the rolling elements. Generally, during the early stages of the fault, the surface is only locally affected, and vibrations are generated as a result of the repetitive impacts of the moving components on the defect.

It was observed from Figure 6 that "g" value was high on pump bearings and which causes to excite the high frequencies. These high frequencies are generated due to the presence of bearing defect frequencies. Thus it was suspected that pump bearings had been damaged. To confirm the bearing fault has occurred, the noise measurement and ultrasonic sound measurement were taken. At the same time, the sound level meter recorded as 95 dB of sound level closer to pump DE bearing, which is as shown in Figure 7. Simultaneously the ultrasonic meter showed the ultrasound level of 15 dB by blowing 5 led lights on the display panel of ultrasound meter. The final stage of fault diagnosis employed for comparing the observation recorded using three condition monitoring techniques. The vibration spectrum identified that measured "g" value was high on pump bearings and due to which the high frequencies got excited. These high frequencies happened due to the occurrence of bearing defect frequencies. Also, the noise level measured was 95 dB, which was more than 90 dB reference level and ultrasound of 15 dB (blow of 5 Led lights on the display panel) was observed using the ultrasonic meter. Therefore it confirmed that the pump motor DE bearing had been damaged, which needs to be replaced as early. The decision was taken to disassemble the pump motor and replace the bearing.

7.jpg

Figure 7. Sound level measurement

During the maintenance work, it was found that both motor bearing in a damaged condition, which is as shown in Figure 8. Furthermore, no Damage was observed on pump internals and pump impeller, which is as shown in Figure 9.

8.jpg

Figure 8. Motor bearing in a damaged condition

9.jpg

Figure 9. Pump internals and pump impeller

Table 5. Vibration parameters of ACWP 1B

The maintenance work was carried out by replacing the pump bearing. The performance of ACWP 1B was analyzed after replacement of pump bearing. The vibration monitoring data was recorded using vibration analysis technique. The vibration spectrum after the maintenance is plotted in Figure 10 (a) to (h). The details of post-maintenance observations viz. velocity spectrum values and acceleration spectrum values obtained after maintenance using vibration analysis has been tabulated in following Table 5. The Pump bearing was replaced, and the vibration level reduced to a normal level.

10-1.jpg

10-2.jpg

10-3.jpg

10-4.jpg

10-5.jpg

10-6.jpg

10-7.jpg

10-8.jpg

Figure 10. Post maintenance vibration spectrum of ACWP-1B

The reliability analysis of condenser system of Dahanu thermal power plant is determined and reported in this study. The preventive maintenance time interval is evaluated for planning the maintenance activity of the critical system of the plant. This study proposed a coordinated condition monitoring approach, including vibration analysis technique, noise monitoring, and ultrasound measurement for fault diagnosis of ACWP 1B of the condenser system. It was found that the pump driving end bearing was damaged. The important measures were taken and the pump bearing replaced with a new bearing. Through this study, the fault of pump motor bearing is avoided at an early stage of operation. Furthermore, an unplanned breakdown of the plant is also avoided. Improvement in reliability and availability has been achieved for the condenser system. This study strongly recommends to use coordinated condition monitoring approach with more than two / three condition monitoring techniques to improve timeliness and accuracy of fault diagnosis for other critical systems of the thermal power plant.