Mirna Urquidi-Macdonald, Ph.D.,
Professor of Engineering Science and Mechanics
- Fuel Cells, Semi-Fuel Cell, and Batteries
- Corrosion detection and assessment of underground pipelines
- Battery life prediction
- Assessment of cathodic protection of underground structures
- Pitting Corrosion and Failure analysis
- Nuclear Reactors
- Modeling of Clinical and Pharmacological Human Systems. Heart failure predictions and Pharmaco-dynamics.
University Service and Awards
College of Engineering
- 1993-2001 - Faculty Advisor for the Society for Hispanic Professional Engineers and PSU Hispanic Students Organization (SHPE)
- 2002 - Society of Women Engineering (SWE) Faculty Advisor
- 2002 - Highest Service award from the Society of Hispanic Professional Engineers 2003
- 1997-2000 - Member of the Commission for Women
Outreach and Community Service and Awards
- 1996 - Faculty / Staff Diversity Recgnition Award, Sponsored by the Penn State Multicultural Resource Center
The development of advanced lithium/water and lithium/air batteries (fuel cells), as sponsored by the Office of Research and Development in Washington DC. This work has resulted in a detailed description of the Li/H2O interface, including the first description of the formation of Lithium hydride as the barrier oxide film. Li/O2 batteries possess extraordinarily high specific energies (>11,000 W hr/kg), making them of great interest as primary electrical energy sources. We are exploring the design of super cathodes for fuel cells applications.
The development of methods for assessing the effectiveness of non-interruptible power supplies in cathodically protected, buried, gas pipelines. Methods are based on impedance techniques developed to interpret electrochemical impedance data measured on an actual test pipeline. It was found that the net, operating in the pattern recognition mode, could distinguish between the over-protected and under-protected conditions, even though the connections between the pipe and sacrificial anode were inaccessible. This work was sponsored by the American Gas Association. More recently, we developed this technique further and a prototype is planned to be developed as a joint effort between Penn State and the University Federico II in Italy.
Exploration of the process that leads to the degradation of composite intercalation cathodes for lithium/solid polymer electrolyte batteries. My specific role in this project has been to develop both deterministic (i.e. mechanism-based) and pattern recognition model to describe and predict the degradation process. In particular, the model approach has been particularly successful in predicting cycle life of lithium batteries, as evaluated on battery data supplied by NASA.
The development of methods for assessing the effectiveness of non-interruptible power supplies in cathodically protected, buried, gas pipelines. Pattern recognition was developed to interpret electrochemical impedance data measured on an actual test pipeline. It found that the net, operating in the pattern recognition mode, could distinguish between the over-protected and under-protected conditions, even though the connections between the pipe and sacrificial anode was inaccessible. This work was sponsored by the American Gas Association.
The deterministic prediction of pitting damage on condensing heat exchangers. This work involved the development of individuals pits and the assembly of these events to define the damage functions. It is believed that this was the first deterministic prediction of corrosion damage. Comparison of the predicted failure times with those measured at Battelle Columbus Laboratory showed that the predictions were quite accurate. We believe that this predictive technology will eventually revolutionize the prediction of system life times, and will displace damage tolerance analysis as the principal method for assessing the development oif corrosion damage in engineering (including energy) structures.
The development of deterministic models for predicting the rates of growth cracks in stainless steel components in the coolant circuits of water-cooled nuclear reactors. The latest model in this series, the Coupled Environment Fracture Model (CEFM), requires calibration with only a single datum and is now used extensively in codes (e. g. DAMAGE PREDICTOR) for modeling the development of damage in operating boiling water reactors (BWR). An unexpected conclusion of this model, the validity of which has now been demonstrated experimentally, is that reactions occurring on the external surface and which consume the current exiting the crack mouth control the crack growth rate. This finding led to the prediction that inhibition of the charge transfer reactions on the external surface should inhibit crack growth and this is, indeed, found to be the case. This idea is now being explored world-wide as a means of protecting components in the coolant circuits of nuclear powered reactors.
The study of corrosion fatigue in oil platforms. The project aims at applying ANNs to correlate fatigue crack growth rates with local environmental conditions of the Gulf of Mexico in order to reduce the cost of maintaining off shore oil production facilities.
During my sabbatical at the Gerontology Research Center (GRC of the NIH), I worked on developing maps (using artificial neural nets) between drug dose and effect for abciximab, a drug that is used before angioplasty and which, if administrated at high doses, can have very undesirable side effects. We produced two publications and are in the process of submitting an invention disclosure related to our findings. I believe the technology we produced, whereby a drug can be prescribed on a patient-by-patient basis rather than on the basis of a population of patients, is unique. I am now working with a second database that includes social issues, health issues, clinical history, and genetics in the area of heart function and blood pressure. We are in the process of using wavelet analysis on beat-to-beat heart signatures in order to represent the evolution of the Power Spectral Analysis, instead of the non-time-dependent signature obtained from a Fast Fourier Transform. The success of this type of work critically depends on recognizing the time dependency of the different variables, categorizing the importance of the independent variables, and developing mapping that can be represented in a common framework. Particularly powerful in the area of Bio-Complex Systems and Pharmaco-Informatics is the ability to combine different simulation techniques to interpret data and understand the results obtained from the arsenal of experimental/analytical techniques that are currently in use. The simulation techniques include traditional models (i.e., Statistics, Probability, Cluster analysis), and smart modeling (Genetic Algorithm, Neural Network Pattern Recognition, fuzzy logic, knowledge based modeling, wavelet representation, etc.).
Neural Networks data mining and electrochemistry applied to material science; life prediction of components; batteries and fuel cells; and pharmaco modeling.