Preventing Torque Converter Fatigue Failure: A Case Study on Hydraulic Variable Speed Gearboxes​

Preventing Torque Converter Fatigue Failure: A Case Study on Hydraulic Variable Speed Gearboxes

‍

Problem

A large midstream company had experienced multiple torque converter turbine wheel failures in a hydraulic variable speed gearbox on a critical compression train. The machine was being operated well within the rated torque limit of the hydraulic gearbox and the driver motor. The company and the gearbox OEM had struggled to determine the root cause of the failure and were concerned about recurrence due to the large commercial implications of downtime on the machine.

Solution

After learning about Mechademy’s success at another company-owned facility, the company opted to deploy Turbomechanica, Mechademy’s cloud-based predictive analytics platform, to monitor its motor-driven centrifugal compressor trains, including the problematic hydraulic gearboxes.​

Mechademy implemented physics-based asset models to evaluate the performance of the centrifugal compressor, hydraulic gearbox, and driver motor as individual components and as part of the entire compression train. This holistic approach allowed for real-time monitoring of machine performance, as well as the power balance across the connected machines.​

Additionally, machine learning models were built using historical data to analyze vibration signatures (both radial proximitors and casing accelerometers), dry gas seal health, and motor winding temperatures.​

Within one week of going live, Turbomechanica detected that the hydraulic gearbox was operating on its working oil cooler thermal limit line for extended periods. This undesirable mode of operation, where the torque converter turbine rotates counter to the torque converter impeller, was identified as a major source of heat generation that likely contributed to the historical gearbox failures.​

The platform also identified that this operational condition occurred specifically when the incoming gas supply exceeded demand, leading to a low-speed, high-power mode of operation within the gearbox. By correlating this to heat generation, the root cause of the turbine wheel failure was identified with confidence.

Implementation

Due to the company’s cybersecurity requirements, the platform had to adhere to strict data handling protocols. All data had to be “pushed” to vendor systems in a unidirectional manner to prevent any cybersecurity risks. The Turbomechanica platform’s data connector service met this requirement seamlessly, and with no additional hardware or sensors needed, the platform was fully operational within two weeks, providing immediate results.

Results

With the root cause identified, the company implemented changes to its control logic and set points to avoid prolonged operation on the oil cooler’s thermal limit. This intervention prevented further gearbox failures, reduced downtime, and safeguarded the station’s critical operations.