Have a UROP opening you would like to submit?
Please fill out the form.
Instrumentation for Condition Monitoring of Critical Pump Equipment
2: Mechanical Engineering
Yumeng Cao, firstname.lastname@example.org
Machinery health monitoring and predictive maintenance is a booming field for both academic research and industrial application. A comprehensive solution to monitor the equipment and extract information for decisions will enable field engineers to predict Remaining Useful Lifetime (RUL) of the machine, prevent serious failures and make maintenance decision in a cost-effective manner. Such predictive maintenance management program would involve a number of measurement techniques (e.g., vibration monitoring, thermography, tribology), signal processing (e.g. power spectral density, Fourier transformation, wavelet transform) and machine learning (e.g. neural networks, support vector machine, Bayesian network, etc.). Pumps are ubiquitous in industrial applications where fluids are involved. For chemical, refinery and power generation industries, the number of pumps being used can easily go into the hundreds of thousands range per factory. Nearly half of all operating costs are related to repair and maintenance tasks. With the development of sensing technology for Internet of Things (IoT) and machine learning techniques, Prognostics and Health Management (PHM) techniques can be developed with pump equipment health monitoring system and algorithms for maintenance decision making. The successful application of data-driven methods for predictive maintenance depends on good quality data gathered by a monitoring system with suitable sensors installed. MIT Mechatronics Research Lab (MRL) is looking for UROP students with interest in sensor instrumentation and pump testbed design. The specific goal is to design and implement a pump system testbed with multiple sensing modalities and controllable failure modes for test data generation. The sensor selection, instrumentation and test bed design will consider both the data analytics requirements and physical constrains of existing industrial systems. The student will be involved in experimental data collection and pre-processing to verify sensor performance and ensure good data quality. As a multidisciplinary project, the specific tasks will involve, mechanical design, manufacturing, printed circuit board (PCB) level electronics, data acquisition (DAQ) and basic programing for signal processing. Students will be trained to use a number of lab tools for experiments including National Instrument DAQ systems and SIOS laser interferometer. The expected outcome is a testbed designed and implemented for pump system data fault mode data collection. Throughout this research experience, students are expected to gain significant technical knowledge, implementation skills and hands-on experience with mechatronic system design, sensor instrumentation and PCB design for real-world industrial applications. Opportunities for publications and patents are available for students with outstanding performance. This role involves regular communication meetings with project members and principle investigator.
Solidworks, Matlab, PCB (Multisim/Ultibaord preferred), machine shop access (preferred), signal processing (preferred), in-person research on-campus required