Thermal modeling and error detection in a data processing configuration

• US 20060231639A1
• Filed: 05/16/2005
• Published: 10/19/2006
• Est. Priority Date: 04/14/2005
• Status: Active Grant

• Current Assignee: Lenovo International Limited (Lenovo Group Ltd.)
• Original Assignee: International Business Machines Corporation
• Inventors: Harper, Richard Edwin, Piazza, William Joseph
• Granted Patent: US 7,526,400 B2
• Time in Patent Office: Days
• US Class Current: 236/94
• CPC Class Codes: G01K 15/00 Testing or calibrating of thermometers
• CPC Class Codes: G01K 7/42 Circuits effecting compensation of thermal inertia; Circuits for predicting the stationary value of a temperature
• CPC Class Codes: G06F 1/20 Cooling means
• CPC Class Codes: G06F 1/206 comprising thermal management

Abstract

A method of operating a data processing system includes predicting values for a set of thermal sensors based on an assumed inlet air temperature and a selected configuration state. Additional values are predicted for a plurality of assumed inlet air temperatures and a plurality of configuration states. Actual thermal sensor values are then determined. A measure of the difference between each set of predicted values and the set of actual values is determined and, based thereon, a most likely configuration state is identified. The plurality of thermal sensors preferably includes an inlet air thermal sensor suitable for determining air temperature at an inlet to the system. The configuration states includes a state in which one of the thermal sensors is malfunctioning. The measure of difference is determined by squaring a difference between each actual thermal sensor values and a corresponding predicted thermal sensor value and summing the squares.

Claims

1. A method of operating a data processing system having a plurality of sensors, comprising:

predicting a set of values for the plurality of sensors corresponding to an assumed ambient condition and a selected configuration state;

repeating said predicting for a plurality of assumed ambient conditions and a plurality of configuration states to generate a plurality of sets of predicted values;

determining a set of actual values of said sensors;

determining information indicative of differences between each set of predicted values and the set of actual values;

identifying a most likely configuration state and actual ambient condition based on the determined information.

2. The method of claim 1, wherein the plurality of sensors includes an inlet air temperature sensor suitable for determining air temperature at an inlet to the data processing system.

3. The method of claim 1, wherein the plurality of configuration states include a configuration state in which one of the sensors is malfunctioning.

4. The method of claim 3, wherein the malfunctioning sensor reports a temperature value that differs from the actual temperature, and wherein the difference between the temperature reported by the malfunctioning sensor and the actual is relatively constant.

5. The method of claim 4, wherein determining the information comprising determining a figure of merit for each of the sets of predicted values, wherein each figure of merit is determined by squaring a difference between each of the actual sensor values and a corresponding predicted sensor value and summing the squares.

6. The method of claim 5, wherein identifying the most likely configuration state comprises selecting the configuration state having the lowest figure of merit.

7. The method of claim 1, further comprising, initiating corrective action responsive to said identifying the most likely configuration state.

8. A data processing system having a plurality of thermal sensors, comprising:

a plurality of interconnected server blades within a chassis;

a management module, including a management processor;

management processor means for predicting a set of values for the plurality of thermal sensors corresponding to an assumed inlet air temperature and a selected configuration state;

management processor means for repeating said predicting for a plurality of assumed inlet air temperatures and a plurality of configuration states to generate a plurality of sets of predicted values;

management processor means for determining a set of actual values of said thermal sensors;

management processor means for determining information indicative of differences between each set of predicted values and the set of actual values;

management processor means for identifying a most likely configuration state based on the determined information.

9. The data processing system of claim 8, wherein the plurality of thermal sensors includes an inlet air temperature thermal sensor suitable for determining air temperature at an inlet to the data processing system.

10. The data processing system of claim 8, wherein the plurality of configuration states include a configuration state in which one of the thermal sensors is malfunctioning.

11. The data processing system of claim 10, wherein the malfunctioning thermal sensor reports a temperature value that differs from the actual temperature, and wherein the difference between the temperature reported by the malfunctioning thermal sensor and the actual is relatively constant.

14. The data processing system of claim 8, further comprising, initiating corrective action responsive to said identifying the most likely configuration state.

12. The data processing system of claim 12, wherein determining the information comprising determining a figure of merit for each of the sets of predicted values, wherein each figure of merit is determined by squaring a difference between each of the actual thermal sensor values and a corresponding predicted thermal sensor value and summing the squares.

13. The data processing system of claim 12, wherein identifying the most likely configuration state comprises selecting the configuration state having the lowest figure of merit.

15. A computer program product comprising computer executable instructions, stored on a computer readable medium, for operating a data processing system having a plurality of thermal sensors, comprising:

instructions for predicting a set of values for the plurality of thermal sensors corresponding to an assumed inlet air temperature and a selected configuration state;

instructions for repeating said predicting for a plurality of assumed inlet air temperatures and a plurality of configuration states to generate a plurality of sets of predicted values;

instructions for determining a set of actual values of said thermal sensors;

instructions for determining information indicative of differences between each set of predicted values and the set of actual values;

instructions for identifying a most likely configuration state based on the determined information.

16. The computer program product of claim 15, wherein the plurality of thermal sensors includes an inlet air temperature thermal sensor suitable for determining air temperature at an inlet to the data processing system.

17. The computer program product of claim 15, wherein the plurality of configuration states include a configuration state in which one of the thermal sensors is malfunctioning.

18. The computer program product of claim 17, wherein the malfunctioning thermal sensor reports a temperature value that differs from the actual temperature, and wherein the difference between the temperature reported by the malfunctioning thermal sensor and the actual is relatively constant.

19. The computer program product of claim 18, wherein determining the information comprising determining a figure of merit for each of the sets of predicted values, wherein each figure of merit is determined by squaring a difference between each of the actual thermal sensor values and a corresponding predicted thermal sensor value and summing the squares and wherein identifying the most likely configuration state comprises selecting the configuration state having the lowest figure of merit.

20. The computer program product of claim 15, further comprising, initiating corrective action responsive to said identifying the most likely configuration state.