In our methanol production facility, we operate an air booster compressor that is critical for the air separation unit. This compressor, a multi-shaft centrifugal type designed by a German manufacturer, has been in service since 2007. Over time, we observed a gradual increase in vibration levels, particularly in the large gear shaft, which eventually led to a catastrophic failure involving multiple fractures in the gear shaft teeth. This article details the first-person account of the incident, the analysis of the root causes, and the implementation of a domestic solution to restore operational integrity. Throughout this process, the term ‘gear shaft’ is emphasized as it is central to the failure mechanism and remediation efforts.
The air booster compressor is part of a system that handles compressed air for various processes. It features a driven gear shaft that powers a large gear, which in turn drives three smaller pinion shafts. These components are housed in a single gearbox, and the compressor includes five stages with impellers mounted on the pinion shafts. Key technical parameters are summarized in Table 1 to provide context for the operational environment and stresses involved.
| Parameter | Value |
|---|---|
| Height (mm) | 4000 |
| Width (mm) | 3000 |
| Length (mm) | 3000 |
| Total Mass (kg) | 41000 |
| Number of Compression Stages | 5 |
| Input Power (kW) | 16 |
| Air Volume Flow Rate (m³/h) | 156400 |
| Inlet Pressure (MPa) | 0.640 |
| Outlet Pressure (MPa) | 7.094 |
| Maximum Working Pressure (MPa) | 7.9 |
| Maximum Discharge Temperature (°C) | 106 |
| Sealing Gas Volume Flow Rate (m³/h) | 72 |
| Sealing Gas Pressure (MPa) | 0.5 |
| Number of Pinion Shafts | 4 |
| Large Gear Shaft Speed (rpm) | 1756 |
| Pinion Shaft 1 Speed (Stages 1+2, rpm) | 12566 |
| Pinion Shaft 1 First Critical Speed (rpm) | 6750 |
| Pinion Shaft 1 Second Critical Speed (rpm) | 8200 |
| Pinion Shaft 1 Third Critical Speed (rpm) | 19300 |
| Pinion Shaft 2 Speed (Stage 3, rpm) | 13404 |
| Pinion Shaft 2 First Critical Speed (rpm) | 8400 |
| Pinion Shaft 2 Second Critical Speed (rpm) | 17200 |
| Pinion Shaft 3 Speed (Stages 4+5, rpm) | 19148 |
| Pinion Shaft 3 First Critical Speed (rpm) | 11000 |
| Pinion Shaft 3 Second Critical Speed (rpm) | 13500 |
| Pinion Shaft 3 Third Critical Speed (rpm) | 27900 |
The initial signs of trouble emerged in 2014, when vibration levels in the large gear shaft began to rise slowly. By 2015, these vibrations exceeded 90 μm, prompting a minor overhaul in 2016 where the bearing on the drive side of the large gear shaft was replaced. However, this intervention did not resolve the vibration issues. During a major overhaul in June 2017, we discovered two fractured teeth on the drive gear shaft. Due to time constraints and lack of spare parts, we performed only basic repairs, including grinding the fractured areas, non-destructive testing, and low-speed dynamic balancing. The gear shaft was reinstalled after on-site measurements were taken for potential reproduction.
Post-repair, vibration levels remained high and continued to increase, reaching 100 μm by September 2017, which forced an emergency shutdown. Upon disassembly, we found seven additional fractured teeth on the gear shaft. The fracture surfaces exhibited distinct characteristics: smooth, fine fan-shaped zones at the working surfaces; concentric arcs in the propagation areas; improper oil injection positioning away from the meshing center; and dark gray fibrous fracture regions. These observations pointed to a fatigue failure mechanism, where the material’s endurance limit was exceeded, leading to initial fatigue cracks that propagated under cyclic loading. Subsequent fractures showed signs of impact failure, indicated by the rough, fibrous textures.
To understand the root cause, we conducted a detailed analysis of the gear shaft’s operational stresses and material properties. The gear shaft operates under high-speed conditions, with a linear velocity of 150 m/s, and is subject to significant bending and contact stresses. The fatigue life of a gear shaft can be modeled using the Basquin equation for high-cycle fatigue:
$$ N_f = \frac{C}{\sigma^m} $$
where \( N_f \) is the number of cycles to failure, \( \sigma \) is the stress amplitude, and \( C \) and \( m \) are material constants. For the original gear shaft material, LA-4340, typical values of \( m \) range from 6 to 12 for steel alloys. However, improper heat treatment or material defects can reduce \( C \), accelerating fatigue. Additionally, the contact stress on the gear teeth can be estimated using the Hertzian contact theory:
$$ \sigma_c = \sqrt{\frac{F}{\pi b} \cdot \frac{1}{\frac{1-\nu_1^2}{E_1} + \frac{1-\nu_2^2}{E_2}}} $$
where \( \sigma_c \) is the maximum contact stress, \( F \) is the normal force, \( b \) is the face width, \( \nu \) is Poisson’s ratio, and \( E \) is the modulus of elasticity. Misalignment or inadequate lubrication can increase \( F \), leading to localized stress concentrations that initiate cracks. In our case, the oil injection not being centered on the meshing point likely caused insufficient lubrication, increasing friction and thermal stresses, which compounded the fatigue issues.
Further analysis revealed that the gear shaft’s design and manufacturing tolerances contributed to the failure. The helical gears required high precision, but wear and deformation over time altered the meshing conditions. The cumulative damage from fatigue can be expressed using Miner’s rule:
$$ D = \sum \frac{n_i}{N_{f,i}} $$
where \( D \) is the total damage, \( n_i \) is the number of cycles at stress level \( i \), and \( N_{f,i} \) is the fatigue life at that stress. When \( D \) reaches 1, failure occurs. Our operational data indicated that vibration-induced stress variations pushed \( D \) beyond the threshold, especially after the initial fractures altered the load distribution.
Given the urgency and high cost of importing a replacement gear shaft (with lead times up to 8 months), we opted for a domestic solution. We collaborated with a local manufacturer to reproduce the gear shaft using domestic materials and expertise. The original material, LA-4340, was replaced with 35CrMoV, a Chinese-grade alloy steel with comparable mechanical properties. Key material properties are compared in Table 2.
| Property | Original Material (LA-4340) | Domestic Material (35CrMoV) |
|---|---|---|
| Tensile Strength (MPa) | ≥930 | ≥980 |
| Yield Strength (MPa) | <td≥835 | </td≥835</td≥785 |
| Elongation (%) | ≥12 | ≥14 |
| Impact Toughness (J) | ≥63 | ≥55 |
| Hardness (HB) | 269-302 | 269-302 |
The manufacturing process involved precise measurements and simulations to account for deformations in the gearbox and other components. Due to the severe damage, direct measurement of original design parameters was unreliable. We used advanced computational tools to determine optimal modification factors, such as the profile shift coefficient, and applied gear tooth modifications to improve contact patterns. The goal was to achieve over 85% contact area under load, compared to the previous 50%. The gear shaft’s geometry was optimized using CAD models, and the manufacturing included strict quality controls, such as carburizing and grinding to enhance surface hardness and fatigue resistance.
During the installation, we ensured proper alignment and lubrication. The oil injection system was adjusted to target the meshing center, reducing the risk of localized heating and stress. Post-installation, we conducted rigorous testing, including vibration analysis and thermal monitoring. The results showed significant improvements: vibration levels dropped to 37 μm, and temperatures stabilized at 80°C, within acceptable limits. The domestic gear shaft performed comparably to the original, with no recurrence of fractures during subsequent operations. A follow-up inspection in 2019 confirmed that all parameters met the manufacturer’s specifications, validating the success of the domestic solution.
In conclusion, the failure of the gear shaft was primarily due to fatigue and impact fractures exacerbated by operational stresses, misalignment, and lubrication issues. The domestic reproduction of the gear shaft using 35CrMoV material not only resolved the immediate problem but also provided a cost-effective and timely alternative to imported parts. This experience underscores the importance of proactive maintenance, accurate failure analysis, and the potential of local manufacturing in addressing critical equipment failures. The successful implementation prevented substantial economic losses and set a precedent for similar applications in the industry.
To further generalize the findings, we can derive a reliability model for gear shafts under similar conditions. The failure rate \( \lambda(t) \) can be modeled using a Weibull distribution:
$$ \lambda(t) = \frac{\beta}{\eta} \left( \frac{t}{\eta} \right)^{\beta-1} $$
where \( \beta \) is the shape parameter and \( \eta \) is the scale parameter. For fatigue-dominated failures, \( \beta \) typically ranges from 1 to 3. By monitoring vibration and stress data, we can predict remaining useful life and schedule maintenance accordingly. This approach enhances the longevity of gear shafts and minimizes unplanned downtime.
In summary, the gear shaft is a critical component in centrifugal compressors, and its failure can have cascading effects on plant operations. Through detailed analysis and innovative solutions, we managed to turn a challenging situation into an opportunity for improvement. The insights gained from this case can be applied to other high-speed rotating machinery, ensuring safer and more efficient operations in the future.