Govind Saraswat Email & Phone Number

Austin, Tx, Us

Austin, Texas, United States

Developing robust and renewable whole home energy solutions.

Greater Denver Area

• Research includes power system modeling and analysis, measurement-based operation and control, machine learning and optimization. Performed large scale Power Hardware-in-loop (PHIL) experiments for coordinating 50+ DERs for providing ancillary services. • Developed scalable load forecasting and system identification algorithms for power systems with high penetration of distributed energy sources. Used random forest classifiers and deep neural network for non-intrusive load monitoring for household loads. • Analyzed the effects of cyber-attacks on state-of-art state estimation algorithms by designing realistic cyber-attack scenarios.

Greater Minneapolis-St. Paul Area

• Created cyber-secured distributed algorithms for cyber-physical networks with applications to smart micro grids. Implemented the algorithms on RaspberryPi clusters using NodeJS framework as a distributed controller layer for providing ancillary services to the grid.• Fine-tuned the algorithm to handle real world delays and asynchrony achieving high-bandwidth (200 kHz) state updates for application toward secondary frequency regulations. • Interfaced the distributed controller layer with high power commercial scale PV and battery inverters using Modbus communication protocol for sending power dispatch commands in micro grids for demand response applications.

Hyderabad Area, India

Bengaluru Area, India

• Developing high quality software in C++/Tcl for performing EM reliability analysis of microprocessors.

Santa Clara, Ca

• Developed a new reliability analysis method using log-normal probability distribution and failure acceleration models to analyze the reliability of microprocessors due to Electromigration (EM). Method was able to accurately predict the failure rate due to EM thus eliminating the inherent approximate nature of other state-of-art methods. A US patent was awarded for the same.• Implemented the method as a highly parallel C++ tool which can analyze designs with more than billion nodes. The implementation requires graph tracing of chip layout to match patterns for which model parameters are known. • Interfaced with the manufacturers (outside team) and other cross-function teams to understand and procure EM models for latest technology node and implement them quickly to enable fast EM verification.

University Of Minnesota - Twin Cities

1. Kalman filter based detection and identification of participatory modes of a flexure based nano-imaging systemAchieved high-bandwidth tracking of contribution of different modes of a nano-measurement probe (AFM) using a receding horizon Kalman filter. Further designed a linear unbiased estimator to detect and quantify the presence of higher modes in the probe. This resulted in an order of magnitude improvement over traditional amplitude based tracking. Filter was designed using physics based modeling and characterization of the probe.2. Feedback control and sensitivity analysis of frequency shift of an oscillatory probe based measurement systemProposed a new scheme of imaging using frequency shift of AFM probe as the feedback signal. Performed simulations achieving high resolution imaging; two orders of increase in SNR compared with conventional amplitude based feedback.3. Real-time probe-based quantitative determination of material properties at the nanoscale Developed an equivalent linear model for non-linear probe-sample interactions relating the sample material properties to the equivalent parameters. Built and implemented an RLS algorithm on an FPGA (Xilinx Virtex2p30) to estimate the parameters in real-time. Demonstrated the effectiveness of the method by investigating properties of a polymer blend. 4. Optimal control law design for a stochastic hybrid system to increase transport Developed an optimal control law using dynamic programming to maximize transport of a stochastic hybrid system. Extracted ‘useful work’ with the help of thermal noise while incorporating realistic constraints of the measurement system.5. Stochastic modeling of directed intracellular transport Hypothesized a biased random walk based model for intracellular transport and developed a simulation model which showed directed transport in two dimensions. Analytically verified the results by providing exact solution of corresponding stochastic differential equation (SDE).

Implemented BSIM4 (Berkeley Simulation) model for MOSFET in 'PSpice' and verified the results with another circuit simulator 'Spectre'. Programming was done in C++ on Microsoft .NET Framework.

Doctor Of Philosophy (Ph.D.), Electrical And Electronics Engineering

Bachelor Of Technology, Electrical Engineering