Patents by Inventor Sriram Ganapathy
Sriram Ganapathy has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Patent number: 10089977Abstract: Exemplary embodiments of the present invention provide a method of system combination in an audio analytics application including providing a plurality of language identification systems in which each of the language identification systems includes a plurality of probabilities. Each probability is associated with the system's ability to detect a particular language. The method of system combination in the audio analytics application includes receiving data at the language identification systems. The received data is different from data used to train the language identification systems. A confidence measure is determined for each of the language identification systems. The confidence measure identifies which language its system predicts for the received data and combining the language identification systems according to the confidence measures.Type: GrantFiled: July 7, 2015Date of Patent: October 2, 2018Assignee: INTERNATIONAL BUSINESS MACHINES CORPORATIONInventors: Sriram Ganapathy, Mohamed K. Omar, Robert Ward
-
Publication number: 20170011734Abstract: Exemplary embodiments of the present invention provide a method of system combination in an audio analytics application including providing a plurality of language identification systems in which each of the language identification systems includes a plurality of probabilities. Each probability is associated with the system's ability to detect a particular language. The method of system combination in the audio analytics application includes receiving data at the language identification systems. The received data is different from data used to train the language identification systems. A confidence measure is determined for each of the language identification systems. The confidence measure identifies which language its system predicts for the received data and combining the language identification systems according to the confidence measures.Type: ApplicationFiled: July 7, 2015Publication date: January 12, 2017Inventors: Sriram Ganapathy, Mohamed K. Omar, Robert Ward
-
Patent number: 9231263Abstract: A system and method that monitor the rate of a voltage drop of fuel cells in a fuel cell stack to determine whether the voltage drop is a result of cathode reactant starvation or anode reactant starvation. The method looks at a falling voltage of a fuel cell to determine whether the rate of the fall in voltage indicates that hydrogen starvation of the anode of the fuel cell is occurring. The method also looks at the actual voltage of the fuel cell that is falling to determine whether it is a below a predetermined minimum voltage threshold also indicating that hydrogen starvation of the anode of the fuel cell is occurring. If hydrogen starvation is occurring, the method performs power limiting of the fuel cell stack either based on the rate or the voltage level.Type: GrantFiled: January 21, 2013Date of Patent: January 5, 2016Assignee: GM Global Technology Operations LLCInventors: Derek R. Lebzelter, Balasubramanian Lakshmanan, Sriram Ganapathy, Sanja Sljivar-Lovria, Bruce J. Clingerman, Michael J. Kiefer
-
Patent number: 9105892Abstract: A system and method for converting a fuel cell stack power request signal to a stack current set-point that considers stack performance parameters. The method includes obtaining a power-current relationship curve of the fuel cell stack to provide stack parameters including exchange current density and mass transfer coefficient. The method then calculates a slope for the stack using the parameters from the power-current relationship curve estimation that includes calculating a cell voltage at two predetermined stack current densities. The method then calculates a change in current in response to the power request signal, the stack voltage, the stack current and the calculated slope, and uses the change in current to update the current set-point for the stack.Type: GrantFiled: March 15, 2010Date of Patent: August 11, 2015Assignee: GM Global Technology Operations LLCInventors: Sriram Ganapathy, Kiran Mallavarapu, Stefan W. Lixfeld, Manish Sinha
-
Patent number: 8986899Abstract: Methods and systems for enhancing the performance of fuel cells in fuel cell vehicles. Some embodiments comprise a current control module for controlling the application of a load to the fuel cell and a voltage monitoring module for monitoring one or more voltages within the fuel cell. The current control module may be configured to apply a delay period before applying a load to the fuel cell after the fuel cell has reached an open circuit voltage. In other embodiments, a fixed delay period may be applied before applying a load to the fuel cell after the fuel cell has reached an open circuit voltage or, for example, an incremental set of delay periods that increase as measured temperature decreases.Type: GrantFiled: October 29, 2012Date of Patent: March 24, 2015Assignee: GM Global Technology Operations, LLCInventors: Daniel I. Harris, Sriram Ganapathy, Michael Cartwright, Pinkhas A. Rapaport
-
Patent number: 8920995Abstract: The present disclosure relates to systems and methods that may be used to predict a performance metric of a fuel cell. A system consistent with the present disclosure may include sensors in communication with the fuel cell stack, a performance metric prediction system, and a control system. The performance metric prediction system may determine a current density based on inputs provided by the sensors at a plurality of time periods, calculate a first parameter while the current density is below a lower threshold, and calculate a second parameter while the current density is above an upper threshold. The first parameter and the second parameter may be used to selectively adjust a fuel cell polarization curve over time. Based upon the polarization curve, a performance metric of the fuel cell stack may be predicted. The control system may implement a control action based upon the performance metric.Type: GrantFiled: March 15, 2013Date of Patent: December 30, 2014Assignee: GM Global Technology Operations LLCInventors: Derek R. Lebzelter, Sriram Ganapathy, Daniel T. Folmsbee
-
Publication number: 20140272652Abstract: The present disclosure relates to systems and methods that may be used to predict a performance metric of a fuel cell. A system consistent with the present disclosure may include sensors in communication with the fuel cell stack, a performance metric prediction system, and a control system. The performance metric prediction system may determine a current density based on inputs provided by the sensors at a plurality of time periods, calculate a first parameter while the current density is below a lower threshold, and calculate a second parameter while the current density is above an upper threshold. The first parameter and the second parameter may be used to selectively adjust a fuel cell polarization curve over time. Based upon the polarization curve, a performance metric of the fuel cell stack may be predicted. The control system may implement a control action based upon the performance metric.Type: ApplicationFiled: March 15, 2013Publication date: September 18, 2014Applicant: GM GLOBAL TECHNOLOGY OPERATIONS LLCInventors: DEREK R. LEBZELTER, SRIRAM GANAPATHY, DANIEL T. FOLMSBEE
-
Publication number: 20140205924Abstract: A system and method that monitor the rate of a voltage drop of fuel cells in a fuel cell stack to determine whether the voltage drop is a result of cathode reactant starvation or anode reactant starvation. The method looks at a falling voltage of a fuel cell to determine whether the rate of the fall in voltage indicates that hydrogen starvation of the anode of the fuel cell is occurring. The method also looks at the actual voltage of the fuel cell that is falling to determine whether it is a below a predetermined minimum voltage threshold also indicating that hydrogen starvation of the anode of the fuel cell is occurring. If hydrogen starvation is occurring, the method performs power limiting of the fuel cell stack either based on the rate or the voltage level.Type: ApplicationFiled: January 21, 2013Publication date: July 24, 2014Applicant: GM GLOBAL TECHNOLOGY OPERATIONS LLCInventors: Derek R. Lebzelter, Balasubramanian Lakshmanan, Sriram Ganapathy, Sanja Sljivar-Lovria, Bruce J. Clingerman, Michael J. Kiefer
-
Patent number: 8722264Abstract: A method for providing a current density set-point for a fuel cell stack in response to a power request from the stack where the set-point is determined based on system parameters that identify the life and degradation of the stack. The method includes dividing a current density range of the fuel cell stack into a predetermined number of sample regions, and selecting the sample regions in order from low to high during the current set-point analysis. The method calculates an average cell voltage for the current density of the selected sample region, and stack power from the average cell voltage. The method then determines whether a power request signal is less than the stack power for the selected sample region and greater than the calculated power for the previous sample region, and if so, calculates the current density set-point at the requested power based on these values.Type: GrantFiled: February 6, 2008Date of Patent: May 13, 2014Assignee: GM Global Technology Operations LLCInventors: Sriram Ganapathy, Kiran Mallavarapu, Balasubramanian Lakshmanan, John P. Salvador, Frank X. Leo
-
Patent number: 8722263Abstract: A method for determining when to inject hydrogen gas into the anode side of a fuel cell stack associated with a fuel cell vehicle when the vehicle is off. The method includes estimating the concentration of hydrogen gas in the anode side of the fuel cell stack using a gas concentration model and determining if the estimated concentration of hydrogen gas is below a first predetermined threshold. If the estimated hydrogen gas is less than the threshold, then hydrogen gas is injected into the anode side from a hydrogen source. While the hydrogen gas is being injected, the method compares the estimated concentration of the hydrogen gas in the anode side to a desired concentration, and generates an error signal there between. If the error signal is greater than a second predetermined threshold, the algorithm continues to inject the hydrogen into the anode side of the fuel cell stack.Type: GrantFiled: April 5, 2010Date of Patent: May 13, 2014Assignee: GM Global Technology Operations LLCInventors: Jun Cai, Joseph Nicholas Lovria, Sriram Ganapathy, Jaehak Jung, John P. Salvador
-
Publication number: 20140120438Abstract: Methods and systems for enhancing the performance of fuel cells in fuel cell vehicles. Some embodiments comprise a current control module for controlling the application of a load to the fuel cell and a voltage monitoring module for monitoring one or more voltages within the fuel cell. The current control module may be configured to apply a delay period before applying a load to the fuel cell after the fuel cell has reached an open circuit voltage. In other embodiments, a fixed delay period may be applied before applying a load to the fuel cell after the fuel cell has reached an open circuit voltage or, for example, an incremental set of delay periods that increase as measured temperature decreases.Type: ApplicationFiled: October 29, 2012Publication date: May 1, 2014Applicant: GM GLOBAL TECHNOLOGY OPERATIONS LLCInventors: Daniel I. Harris, SRIRAM GANAPATHY, MICHAEL CARTWRIGHT, PINKHAS A. RAPAPORT
-
Patent number: 8642220Abstract: A system and method for determining the maximum allowed stack current limit rate for a fuel cell stack that considers cell voltage. The method includes estimating a fuel cell stack voltage based on a fuel cell resistance value, stack variables and a current request signal. The fuel cell resistance value can be modeled based on stack temperature and stack relative humidity. The stack variables can include exchange current density and mass transfer coefficient. The method then uses the estimated fuel cell voltage and a look-up table based on estimated voltage to determine a current rate limit value for changing the current of the stack. The method then adds the current rate limit value and the current request signal to obtain the current set-point.Type: GrantFiled: September 25, 2009Date of Patent: February 4, 2014Assignee: GM Global Technology Operations LLCInventors: Yanyan Zhang, Sriram Ganapathy, Loren Devries, Bruce J. Clingerman
-
Patent number: 8450018Abstract: A system and method for determining when to trigger reconditioning of a fuel cell stack and when to disable the reconditioning of the fuel cell stack. In one embodiment, the stack reconditioning is triggered when a maximum stack power estimation falls below a first predetermined power threshold. The reconditioning of the stack can be disabled so it is not performed when the trigger occurs if the reconditioning process does not raise the maximum power estimation above a second predetermined power threshold or the time from one reconditioning trigger to a next reconditioning trigger is less than a predetermined time threshold, or both.Type: GrantFiled: February 18, 2010Date of Patent: May 28, 2013Assignee: GM Global Technology Operations LLCInventors: Manish Sinha, Sriram Ganapathy, Tayoung Choi, David A. Arthur
-
Patent number: 8428957Abstract: A technique of spectral noise shaping in an audio coding system is disclosed. Frequency decomposition of an input audio signal is performed to obtain multiple frequency sub-bands that closely follow critical bands of human auditory system decomposition. The tonality of each sub-band is determined. If a sub-band is tonal, time domain linear prediction (TDLP) processing is applied to the sub-band, yielding a residual signal and linear predictive coding (LPC) coefficients of an all-pole model representing the sub-band signal. The residual signal is further processed using a frequency domain linear prediction (FDLP) method. The FDLP parameters and LPC coefficients are transferred to a decoder. At the decoder, an inverse-FDLP process is applied to the encoded residual signal followed by an inverse TDLP process, which shapes the quantization noise according to the power spectral density of the original sub-band signal. Non-tonal sub-band signals bypass the TDLP process.Type: GrantFiled: August 22, 2008Date of Patent: April 23, 2013Assignee: QUALCOMM IncorporatedInventors: Harinath Garudadri, Sriram Ganapathy, Petr Motlicek, Hynek Hermansky
-
Patent number: 8420271Abstract: A system and method for detecting a low performing cell in a fuel cell stack using measured cell voltages. The method includes determining that the fuel cell stack is running, the stack coolant temperature is above a certain temperature and the stack current density is within a relatively low power range. The method further includes calculating the average cell voltage, and determining whether the difference between the average cell voltage and the minimum cell voltage is greater than a predetermined threshold. If the difference between the average cell voltage and the minimum cell voltage is greater than the predetermined threshold and the minimum cell voltage is less than another predetermined threshold, then the method increments a low performing cell timer. A ratio of the low performing cell timer and a system run timer is calculated to identify a low performing cell.Type: GrantFiled: July 14, 2009Date of Patent: April 16, 2013Assignee: GM GLobal Technology Operations LLCInventors: Tayoung Choi, Sriram Ganapathy, Jaehak Jung, David R. Savage, Balasubramanian Lakshmanan, Pamela M. Vecasey
-
Patent number: 8405399Abstract: A system for estimating parameters of a fuel cell stack. The system includes a stack health monitor for monitoring minimum cell voltage, stack voltage and current density of the fuel cell stack. The stack health monitor also indicates when a predetermined minimum cell voltage threshold level has been achieved. The system further includes a controller configured to control the fuel cell stack, where the controller determines and records the average fuel cell voltage. The controller generates and stores artificial data points proximate to the one or more predetermined minimum cell voltage threshold levels each time the minimum cell voltage drops below the one or more predetermined minimum cell voltage threshold levels so as to provide an estimation of the fuel cell stack parameters including a minimum cell voltage trend and a minimum cell voltage polarization curve.Type: GrantFiled: July 20, 2010Date of Patent: March 26, 2013Assignee: GM Global Technology Operations LLCInventors: Yanyan Zhang, Sriram Ganapathy, Kiran Mallavarapu
-
Patent number: 8214174Abstract: An algorithm for determining a polarization curve of a fuel cell stack. When the fuel cell stack is running and certain data validity criteria have been met, the algorithm goes into a data collection mode where it collects stack data, such as stack current density, average cell voltage and minimum cell voltage. When the stack is shut down, the algorithm uses a cell voltage model to solve a least squares problem to estimate predetermined parameters that define the polarization curve. If the estimated parameters satisfy certain termination criteria, then the estimated parameters are stored to be used by a system controller to calculate the polarization curve of the stack.Type: GrantFiled: January 31, 2007Date of Patent: July 3, 2012Assignee: GM Global Technology Operations LLCInventors: Sriram Ganapathy, John P. Salvador, Balasubramanian Lakshmanan, Frank Leo, Jason R. Kolodziej
-
Patent number: 8206860Abstract: A system and method for maintaining the voltage of fuel cells in the fuel cell stack below a predetermined maximum voltage. The method determines a desired voltage set-point value that defines a predetermined maximum fuel cell voltage value and uses the voltage set-point value and an average fuel cell voltage to generate an error value there-between. The method generates a minimum gross power prediction value using the modified voltage set-point value to prevent the fuel cell voltages from going above the predetermined maximum fuel cell voltage value and generating a supplemental power value based on the minimum gross power prediction value and the error value to determine how much power needs to be drawn from the stack to maintain the fuel cell voltage below the predetermined maximum voltage value. The method uses the supplemental power value to charge the battery or operate an auxiliary load coupled to the stack.Type: GrantFiled: January 13, 2010Date of Patent: June 26, 2012Assignee: GM Global Technology Operations LLCInventors: Sriram Ganapathy, John P. Salvador, Tayoung Choi, Darrell W. Burleigh
-
Patent number: 8192879Abstract: An algorithm for determining the maximum net power available from a fuel cell stack as the stack degrades over time using an online adaptive estimation of a polarization curve of the stack. The algorithm separates the current density range of the stack into sample regions, and selects a first sample region from the far left of the estimated polarization curve. The algorithm then calculates the cell voltage for that current density sample region, and determines whether the calculated cell voltage is less than or equal to a predetermined cell voltage limit. If the calculated cell voltage is not less than the cell voltage limit, then the algorithm selects the next sample region along the polarization curve. When the calculated cell voltage does reach the cell voltage limit, then the algorithm uses that current density for the sample region being analyzed to calculate the maximum power of the fuel cell stack.Type: GrantFiled: February 6, 2008Date of Patent: June 5, 2012Assignee: GM Global Technology Operations LLCInventors: John P. Salvador, Sriram Ganapathy, Kiran Mallavarapu, Frank X Leo, Balasubramanian Lakshmanan
-
Patent number: 8195407Abstract: A system and method for estimating the amount of hydrogen and/or nitrogen in a fuel cell stack and stack volumes at system start-up and shut-down. The method defines the fuel cell stack and stack volumes as discrete volumes including an anode flow-field and anode plumbing volume, a cathode flow-field volume and a cathode header and plumbing volume. The method estimates the amount of hydrogen and/or nitrogen in the anode flow-field and anode plumbing volume, the cathode flow-field volume and the cathode header and plumbing volume when the fuel cell system is shut down, during a first stage when the hydrogen partial pressure between the anode and cathode is not in equilibrium and during a second stage when the hydrogen partial pressure between the anode and cathode is in equilibrium by considering various flows into and out of the volumes.Type: GrantFiled: March 10, 2010Date of Patent: June 5, 2012Assignee: GM Global Technology Operations LLCInventors: John P. Salvador, Joseph Nicholas Lovria, Sriram Ganapathy, John C. Fagley