Condition Monitoring of Forced-Draft Fan Using Vibration Analysis: A Case Study ^†

Abstract

The purpose of this paper is to present vibration-based condition monitoring of forced-draft fans used in sugar factories. The draft system’s uninterrupted operation is essential for the flawless operation of boilers. Considering its importance, a forced-draft fan was employed as a case study. The vibration and noise in the time and frequency domain, along with the overall vibration and noise levels, were measured from the driving and non-driving ends of forced-draft fans at different intervals of time so that errors in measurement could be avoided. These vibration data were analyzed to identify faults in the different components of the forced-draft fans, along with problems in their operation. The results of this analysis indicate that the fans under study produced more noise and vibration than the recommended standard value. Also, through signature analysis, it was found that the fans needed to be balanced and aligned properly. The problems observed were rectified, and recommendations are given for the proper maintenance of these fans. An effort was made to explore the relationship between patterns of the vibration spectrum and signs of failure in a forced-draft fan. It was found that vibration-based condition monitoring is an effective tool in sugar factories.

1. Introduction

Sugar factories need a continuous supply of steam for carrying out different processes in the manufacture of sugar. This is provided by steam boilers. The continuous and efficient operation of boilers requires their mountings and accessories to perform at full efficiency. The forced-draft (FD) fan is an important device in the efficient working of boilers. It is installed at the inlet of the boiler and used for creating positive pressure in the combustion chamber, which helps the fuel burn properly. Air is drawn from the atmosphere and forced into the furnace via a preheater. Since it is essential for the fan to perform its function uninterrupted, in this work, it was considered for study.

In sugar factories, there are many pieces of rotating machinery and equipment that need careful attention for their proper performance. Many authors have worked on monitoring the conditions of different pieces of equipment and diagnosing the root causes of faults and defects that occur in such equipment and rotating machinery. Industrial fans have many problems related to their shafts, blades, bearings, and housings. Viorel-Mihai et al. [1] investigated the causes of high vibrations in exhaust fans by comparing a mathematical model and an actual system and found that manual errors during the assembly of the isolation system were responsible for the high vibrations. Xie et al. [2] prepared a support vector machine model for predicting possible types of failure of centrifugal fan blades using detailed fatigue analysis and feature extraction from vibration signals. Zhang et al. [3] worked on developing a method in which vibration and acoustic data fusion are used for the detection of cracks in centrifugal fan blades using convolution neural networks. Ouerghemmi et al. [4] provided a method using motor current signature analysis and an extended park vector approach for the detection of a side cut V–belt fault used in the transmission of power from an induction motor to a centrifugal fan. Trebuna et al. [5] conducted a vibration analysis of two exhaust fans to diagnose the problem of excessive vibration in one of the fans using modal analysis and by determining the frequency response function. After analysis, recommendations were given to support the proper function of the fans, helping prevent future damage to the bearings and foundation of the fan. Du et al. [6] investigated the failure of an induced-draft fan’s shaft by analyzing the material and its stress using the finite element method and the effect of torsional vibration. It was found that high stress concentrations and torsional vibration at critical speed were responsible for the failure of the shaft. Dhamande et al. [7] presented a case study conducted to find faults in induced-draft fans using vibration analysis. Jagtap et al. [8] analyzed the failure of induced-draft fans using vibration analysis, ultrasonic monitoring methods, noise measurement, and wear debris analysis. Subramanian et al. [9] used many inspection and measurement methods for the diagnosis of faults in the fan blades of induced-draft fans. Sekhar et al. [10] utilized vibration signature analysis as a tool for detecting unbalance in the motors of conveyors. Nurbanasari et al. [11] investigated the catastrophic failure of forced-draft (FD) fan blades for a coal-fired power plant. It was found that blade failure occurred due to foreign particles of fly ash that collided with the surface of the blade. Dileep et al. [12] detected faults in FD fans using vibration monitoring. Ebersbach and Pengm [13] used time and frequency domain data to develop an expert system that identified twenty-seven types of faults in bearings, gears, belts, fans, coupling, and other mechanical parts. Rusinski et al. [14] investigated faults in fan blades and the effect of heavy vibrations on the housing of the main bearing. There are various environmental factors that affect the efficiency and operation of FD fans in sugar factories. These include dust and particulate matter produced during the crushing and milling processes, temperature variations due to different boiler sections, the moisture content in the air, leading to the corrosion of parts, the quality of the air, and emissions from the factory. These factors are responsible for reducing the performance of FD fans. They require regular cleaning and maintenance of different components, like filters, blades, ducts, etc. The use of corrosion-resistant materials or coatings on fan components, moisture control system, etc., can be effective in improving the performance of FD fans.

The various condition-monitoring techniques used for analyzing the status of FD fans, in addition to vibration monitoring, include monitoring the temperature, pressure, and airflow to determine problems like wear, unbalance, and clogging, which may lead to failure. Also, machine learning and data analytics are used to analyze historical data and identify the patterns in parameter variations related to the deterioration of fan performance. AI-based predictive maintenance, on the basis of real-time data, can also help achieve the optimum maintenance of FD fans.

This work aims to monitor the conditions of forced-draft fans using vibration measurements and to predict the possible reasons for high vibration in the FD fans used for boilers in sugar factories. It also aims to determine the types of faults that occur in FD fans, rectify them and propose remedial actions to prevent future failure of the fan. The data collected in the time and frequency domains were analyzed to identify any abnormal patterns or trends in the FD fan vibrations. To study the rotating machinery vibrations, fan measurements were taken at intervals of a month, in October, November, and December 2023. These measurements were correlated to identify the consequences of defects on the vibrations of the rotating machinery.

2. Methodology

The methodology used in this study is outlined in Figure 1. In this study, the vibrations of the forced-draft fans of three different boilers were analyzed. The methodology consisted of collecting vibration signals from the forced-draft fans using a fast Fourier transform (FFT) analyzer and analyzing them to detect any mechanical- or component-level defects. The vibration readings were collected after maintenance of the machines in the sugar factory and before the start of the next operating season. We decided to collect the vibration readings at monthly intervals. Accordingly, the readings were collected in October, November, and December. The key steps followed in this methodology were as follows:

(1)

The function, construction, and working principle of the forced-draft fans were studied.

(2)

The specifications of the system under study were obtained.

(3)

The vibrations of the system were measured, which included the following:

(i)

Wideband measurements of displacement, velocity, and acceleration (RMS);

(ii)

Measurements of the displacement, velocity, and acceleration spectrum;

(iii)

Measurements of rotational speed;

(iv)

Noise level measurements.

An accelerometer sensor was used to measure the vibration signals from the FD fan. The sensor was mounted on the bearings at both the driving and non-driving ends of the motor, as well as on both ends of the FD fan. Vibration readings were collected in the radial (X and Y) directions and axial direction from the bearings. All vibrations generated by the rotating components, such as the fan blades and shaft, were transmitted to the stationary parts of the bearing housing. The readings were taken during the day, typically between 1:00 PM and 3:00 PM.

(4)

The measurements were analyzed, which consisted of the following:

(i)

The measured values of the vibration amplitudes (RMS) were compared with the standard allowable values specified in ISO 10816 standard [15];

(ii)

The shaft frequency and its harmonics (1X, 2X, 3X, 4X, 5X, etc.) were identified;

(iii)

The bearing defect frequencies were identified;

(iv)

The gear mesh frequency, blade passing frequency, and any other relevant characteristic frequencies were identified.

The vibration measurement setup included a frequency range from 0 to 6400 Hz. The sound sensor collected the sound pressure data in Pascals at the source; however, it also captured background noise. To analyze the collected vibration data, the standard guidelines recommended in ISO 10816 were followed. Additionally, variation patterns in the data were examined to identify mechanical faults.

(5)

The vibration signatures were studied to identify mechanical faults, and changes in the vibration signature were observed across different measurements taken at various intervals.

(6)

The condition of the given system was assessed based on the vibration data.

3. Measurement and Analysis

An FD fan is a system that supplies ambient air to a boiler, usually via a pre-heater, which is useful for increasing the efficiency of the boiler. An external energy source, such as an electric motor, is used to operate the fan. Due to this external energy, FD fans are able to create positive pressure to feed air into the boiler. An FD fan is shown in Figure 2. It consists of a rotating assembly, commonly referred to as an impeller, which is positioned inside a steel casing. This casing includes both air intake and discharge zones, which are connected by air intake and discharge ductwork attached to the FD fan casing. During operation, the intake ductwork carries ambient air to the FD fan, while the discharge ductwork conveys the air to the boiler furnace, where it is mixed with fuel oil for combustion.

3.1. Forced-Draft Fan of Boiler 1

The FD fan used for boiler 1 has a volume flow rate of 69,120 m³/h and is a centrifugal type. This fan rotates in the clockwise direction when viewed from the non-drive end. The fan is powered with a 100 HP electric motor running at a speed 555 rpm. The rotor shaft is supported by bearing number 526.

3.2. Vibration Measurement

The vibrations generated by the rotating components, such as the fan blades mounted on the shaft, are transmitted to the shaft, then to the bearings, and finally, to the casing or housing of the rotating machinery. Therefore, the vibrations were measured by mounting the accelerometer on the bearing casings. As per the ISO 10816 standards, the vibrations of the fans were analyzed using velocity measurements, and the sensor used was an accelerometer. Also, the measurements were taken in the radial (X and Y) directions and the axial direction at the bearing location. The average readings were recorded using probes at various locations, i.e., at the bearings on the driving and non-driving sides of the fan. Three readings each from the vertical (X), horizontal (Y), and axial directions were taken at the bearings. The instruments used for vibration analysis included a 4-channel FFT spectrum analyzer data collector and balancer. The accelerometer used had a sensitivity of 100 mV/g, and a microphone with a sensitivity of 46.17mV/Pa was used to measure the noise. The major specifications of the FFT analyzer included 24-bit resolution A/D convertor, a dynamic range of less than 100 DB, a Hanning window, and an instrument error limit of 5%. The parameter settings of the instrument were as follows: unit: mm/s; frequency range: 12,800; FFT window: Hanning; averages: 2; range: 0–6400; sampling frequency: 16,384; length: 4000 ms.

Figure 3a, Figure 3b, and Figure 3c show graphs of the velocity spectrum of the FD fan in the vertical, horizontal, and axial directions, respectively. In these graphs, frequency is plotted on the X-axis, while vibration velocity, measured in mm/s, is represented as amplitude on the Y-axis. The spectrum shows peaks corresponding to the 1X, 2X, 3X, 4X, and 5X harmonics, which are useful for identifying mechanical defects. To determine the mechanical faults, we needed to observe the amplitude of vibration corresponding to the rotational or shaft speed (1X) and its harmonics (multiples such as 2X, 3X, etc.). As presented in Figure 1, the shaft rotated at 555 rpm, i.e., 9.25 Hz. When the 1X shaft frequency has a higher amplitude compared to all other harmonics, it indicates that there is an unbalance in the machine. When the 2X frequency has the maximum amplitude, it indicates misalignment. When all the harmonics have nearly equal amplitudes of vibration, it indicates looseness in the system.

Table 1. Harmonics of shaft frequency in mm/s for FD fan of high-pressure boiler 1.

Measurement Interval		1	2	3
1X	H	0.32	0.28	0.10
	V	0.07	0.09	0.10
	A	0.20	0.22	0.24
2X	H	0.18	0.16	0.04
	V	0.05	0.05	0.04
	A	0.03	0.06	0.09
3X	H	0.12	0.05	0.01
	V	0.01	0.01	0.01
	A	0.03	0.06	0.14
4X	H	0.09	0.03	0.01
	V	0.02	0.00	0.01
	A	0.01	0.03	0.04
5X	H	0.09	0.04	0.01
	V	0.01	0.01	0.01
	A	0.02	0.03	0.07

provides a tabular representation of Figure 3, showing the harmonics of the shaft frequency (1X, 2X, 3X, 4X….etc.) for three different readings collected at three different time intervals in the vertical (V), horizontal (H), and axial (A) directions. The average displacement (Disp), velocity (Vel), and acceleration (Acc) readings of vibration in the horizontal, vertical, and axial directions are given in

Table 2. RMS velocity values of low-speed FD fan of high-pressure boiler 1.

Measurement Interval		1	2	3
Horizontal	Disp	4.15	12.26	10.26
	Vel	0.53	0.59	0.63
	Acc	0.15	0.13	0.12
Vertical	Disp	1.06	31.96	42.21
	Vel	0.30	0.29	0.28
	Acc	0.42	0.37	0.40
Axial	Disp	2.92	33.88	25.21
	Vel	0.41	0.64	0.60
	Acc	0.12	0.10	0.11
Limiting Range		0–100 microns	1.8–4.5 mm/s	Up to 1 g

. It includes three sets of vibration measurements taken at different time intervals for the low-speed FD fan of boiler 1, along with the corresponding limiting ranges as per the ISO 10816 standard. The recorded sound level of this FD fan was 92.75 decibels.

In Table 1, it can be observed that all the harmonic amplitudes in the three directions were within acceptable limits. Similarly, Table 2 shows that all measured vibration parameters in the three directions also fell within the permissible range. The vibration amplitudes in the horizontal direction across different harmonics were nearly equal, which indicates possible looseness in the nuts and bolts. Additionally, the recorded sound level was slightly higher than the recommended value. It is recommended to tighten the nuts and bolts to eliminate looseness and reduce the vibration and noise level.

3.3. Forced-Draft Fan for Boiler 2

The specifications of the FD fan of boiler 2 include a 250 HP electric motor operating at 1380 rpm. It is equipped with bearing number 617. Figure 4a, Figure 4b, and Figure 4c show the velocity spectrum graphs of the FD fan in the vertical, horizontal, and axial directions, respectively.

Table 3. Harmonics of shaft frequency in mm/s for FD fan of boiler 2.

Measurement Interval		1	2	3
1X	H	0.01	1.8	0.02
	V	0.01	0.58	0.00
	A	0.00	0.36	0.00
2X	H	2.10	0.56	0.70
	V	0.40	0.27	0.15
	A	0.64	0.13	0.99
3X	H	0.12	0.63	0.13
	V	0.02	0.13	0.01
	A	0.04	0.12	0.08
4X	H	0.69	0.09	0.15
	V	0.21	0.22	0.06
	A	0.13	0.13	0.29
5X	H	0.09	0.4	0.23
	V	0.00	0.11	0.02
	A	0.04	0.06	0.06

provides a tabular representation of Figure 4 based on three readings collected at different time intervals.

Table 4. RMS velocity values of FD fan of boiler 2.

Measurement Interval		1	2	3
Horizontal	Disp	15.68	24.60	34.17
	Vel	2.96	2.93	3.31
	Acc	0.90	0.82	0.57
Vertical	Disp	3.11	62.09	32.71
	Vel	0.78	1.26	1.04
	Acc	0.90	1.06	0.46
Axial	Disp	5.26	55.65	46.96
	Vel	1.45	1.45	2.26
	Acc	0.70	0.72	0.59
Limiting Range		0–100 microns	1.8–4.5 mm/s	Up to 1 g

provides the average readings of vibration for the three different readings collected at different intervals of time.

The sound level of this fan was recorded as 114.6125 decibels, which is significantly higher than that of the FD fan of boiler 1. In Table 3, it can be observed that all harmonics readings were within the acceptable range. As shown in Table 4, the average readings of all measured parameters in all directions within an acceptable limit, except acceleration in the vertical direction. This suggest an increase in the velocity in the vertical direction, indicating growing looseness. Additionally, Table 4 shows that the horizontal velocity was higher compared to the vertical and axial directions. The horizontal and axial readings were greater than the vertical readings, which indicates imbalance and misalignment, according to the ISO 10816 standard. Furthermore, the sound level of this component exceeded the recommended value. It is recommended to align the shafts of the motor and the fan and to balance the fan blades to address these issues.

3.4. Forced-Draft Fan of Boiler 3

This forced-draft fan system uses a fan to deliver air to the furnace. This is a centrifugal type of fan that has a volume flow rate of 80,640 m³/h. The fan rotates in the clockwise direction at a speed of 1480 rpm (as viewed from the non-drive end). It requires a 100 HP motor for operation. The bearing used in this fan is the SN516 type.

Figure 5a, Figure 5b, and Figure 5c present velocity spectrum graphs of the FD fan of boiler 3 in the vertical, horizontal, and axial directions, respectively.

Table 5. Harmonics of shaft frequency in mm/s for FD fan of high pressure boiler 3.

Measurement Interval		1	2	3
1X	H	1.61	0.69	1.02
	V	1.70	1.80	1.71
	A	0.01	0.01	0.89
2X	H	0.25	0.43	0.64
	V	0.29	0.49	0.47
	A	0.89	0.14	0.59
3X	H	0.05	0.00	0.14
	V	0.04	0.00	0.06
	A	0.01	0.00	0.09
4X	H	0.04	0.04	0.02
	V	0.05	0.49	0.07
	A	0.59	0.52	0.03
5X	H	0.01	0.02	0.03
	V	0.04	0.00	0.04
	A	0.03	0.00	0.08

provides a tabular representation of Figure 5. The average vibration readings are given in

Table 6. RMS velocity values of FD fan of high-pressure boiler 3.

Measurement Interval		1	2	3
Horizontal	Disp	11.16	34.57	31.45
	Vel	2.90	2.23	2.53
	Acc	0.82	0.71	0.53
Vertical	Disp	13.53	108.20	55.27
	Vel	2.87	2.90	2.84
	Acc	0.90	0.96	0.87
Axial	Disp	11.61	70.08	67.35
	Vel	5.11	3.30	2.28
	Acc	0.45	0.40	0.42
Limiting Range		0–100 microns	1.8–4.5 mm/s	Up to 1 g

, which include three measurements of the displacement, velocity and acceleration at the first, second, and third time intervals. The sound pressure level of this component was measured at 123.073 decibels.

In Table 5, it can be observed that all the harmonics readings were within a safe range. However, the vibration at the 2X running speed exceeded that of the 1X level by 150%, indicating misalignment. It is possible that coupling damage may occur if the FD fan continues to operate at the same speed. The average reading also shows high axial vibration velocity, although it is following a decreasing trend. In Table 6, it can be observed that all three measuring parameters, that is, displacement, velocity, and acceleration, were within the acceptable limits, except displacement in the vertical direction during the second measurement interval, which exceeded the permissible range. The sound pressure level of this fan was significantly higher than that of the previous two FD fans. It is recommended to align the shaft of the motor and fan to address the misalignment issue.

4. Result and Discussion

Based on the vibration condition monitoring of the three FD fans (for boilers 1, 2, and 3), it is evident that the FD fans of boilers 1 and 2 were in better condition compared to that of boiler 3 during the first measurement interval conducted in October. In the first reading of the FD fan of boiler 3, the axial velocity amplitude was found to be above the allowable limit, prompting detailed spectral analysis. It was found that the 1X and 2X components had significantly higher amplitudes compared to the 3X, 4X, and 5X components, indicating angular or parallel misalignment. When verified with the harmonics of the horizontal and vertical directions, the 1X component was observed to be high, while for the axial direction, 1X and 2X were both high. Hence, parallel misalignment was absent, but angular misalignment was present. This was further verified by the phase measurement across the coupling. It can be concluded that the misalignment was due to a coupling problem.

Maintenance was carried out on all three fans as per the recommendations given above, after the first measurement and before the start of the new production cycle in the sugar factory. The second set of measurements was taken in November, and the results are presented in Table 1, Table 2, Table 3, Table 4, Table 5 and Table 6 for all three FD fans. As observed in Table 6, the displacement amplitude during the second reading of vibration in the vertical direction of the FD fan of boiler 3 was found to be more than the allowable limit. Hence, the spectra shown in Figure 6, Figure 7 and Figure 8 were analyzed to determine the reason by examining the frequency spectrum of velocity in the vertical, horizontal, and axial directions corresponding to the three different intervals.

It was found that the 1X, 2X, 3X, and higher harmonics had nearly the same amplitude, which shows that there was mechanical looseness in the system. Hence, all the nuts and bolts of the foundation of the bearing were tightened, which reduced the overall level of vibration displacement. Continuous monitoring of FD fans is necessary for their proper function. The outcomes of this case study indicate that vibration monitoring helps to identify problems in machinery and rectify them in real time to avoid any future failures, which in turn could disturb the production process.

5. Conclusions

In this study, condition monitoring of FD fans of boilers used in a sugar factory was carried out. It was observed that the continuous operation of these fans is crucial for the proper functioning of the boilers. The failures that occur in FD fans are shaft, bearing, housing, coupling, and blade failures, which are due to unnoticed mechanical faults, such as unbalance, looseness, misalignment, coupling problems, etc. Vibration analysis proves to be a promising tool for identifying such problems in rotating machinery. It can assist maintenance engineers in effectively extending the useful life of rotating machinery by enabling early fault detection and timely corrective actions.