Mathematical model useful for determining and calibrating output of a linear sensor

• US 7,124,047 B2
• Filed: 09/03/2004
• Issued: 10/17/2006
• Est. Priority Date: 09/03/2004
• Status: Expired due to Fees

• Current Assignee: Eaton Corp. (Wisconsin) (Eaton Corporation PLC.)
• Original Assignee: Eaton Corp. (Wisconsin) (Eaton Corporation PLC.)
• Inventors: Herbert, David E., Zhang, Yongcang
• Primary Examiner(s): Barlow, John
• Assistant Examiner(s): KUNDU, SUJOY K
• Application Number:
  US10/934,058
• Publication Number:
  US 20060052959A1
• Time in Patent Office: 774 Days
• Field of Search: None
• US Class Current: 702/104
• CPC Class Codes: G01D 18/008 with calibration coefficients stored in memory
• CPC Class Codes: G01D 3/022 having an ideal characteristic, map or correction data stored in a digital memory

Abstract

The present invention relates to a method of analyzing accuracy of linear sensor output based on a theoretical prediction of the output model. The method comprises steps of: analyzing a statistically significant number of linear sensors to obtain operating profiles at two or more operating conditions; individually testing a linear sensor to obtain at least four data points; developing a theoretical prediction of the output model for each operating condition from the four data points; and comparing accuracy of the relationship between the theoretical prediction of the output model to sensor output in the operating profile corresponding to a given input for the same operating condition. Also disclosed are calibration algorithms associated therewith, as well as mechatronic units and further assemblies therefrom.

Claims

1. A method of analyzing accuracy of linear sensor output based on a theoretical prediction of the output model, the method comprising:

analyzing a statistically significant number of linear sensors to obtain a first operating profile for each sensor at a first operating condition, T₁;

analyzing the statistically significant number of linear sensors to obtain a second operating profile for each sensor at a second operating condition, T₂;

optionally confirming that each operating profile obtained represents a linear relationship between sensor input and sensor output using reciprocal slopes of output-input lines;

optionally confirming that operating profiles developed over a continuum of operating conditions represent a linear relationship using reciprocal slopes of each operating profile;

individually testing a linear sensor to obtain at least four data points, wherein the four data points correspond to test data obtained based on a first and second input at each of the first and second operating conditions and are represented as follows;

(I₁₁, O₁₁), (I₂₁, O₂₁), (I₁₂, O₁₂), and (I₂₂, O₂₂);

developing a theoretical prediction of the output model from the four data points and relaying the theoretical prediction of the output model to an electronic component, wherein the theoretical prediction of the output model for determining output at a given operating condition, T, and based on a given input, I, is developed in accordance with the following mathematical equation;

Output=O₀+I/b,

 

wherein b=b₁+[(b₁−

b₂)/(T₁−

T₂)](T−

T₁), where b₁=(I₁₁−

I₂₁)/(O₁₁−

O₂₁);

b₂=(I₁₂−

I₂₂)/(O₁₂−

O₂₂); and

O₀=(O₀₁+O₀₂)/2, where O₀₁=O₁₁−

I₁₁/b₁ and O₀₂=O₁₂−

I₁₂/b₂;

comparing accuracy of the relationship between the theoretical prediction of the output model for the first operating condition to sensor output in the first operating profile for the first operating condition corresponding to the given input; and

comparing accuracy of the relationship between the theoretical prediction of the output model for the second operating condition to sensor output in the second operating profile for the second operating condition corresponding to the given input.

2. The method of claim 1, further comprising the step of determining overall accuracy of the theoretical prediction of the output model by statistically interpreting worst deviation of the theoretical prediction of the output model from the operating profile for a corresponding operating condition.

3. The method of claim 1, wherein the statistically significant number of linear sensors comprises approximately 30 to about 300.

4. The method of claim 1, wherein the statistically significant number of linear sensors are individually selected from more than one manufacturing lot.

5. The method of claim 1, wherein the statistically significant number of linear sensors are analyzed during normal end-of-line testing after sensor manufacture.

6. The method of claim 1, wherein a selected individual linear sensor is tested to obtain the at least four data points during normal end-of-line testing after sensor manufacture.

7. The method of claim 1, wherein the first and second operating conditions vary essentially only in temperature.

8. The method of claim 1, wherein the first and second operating conditions are temperatures differing by at least 40°

C.

9. The method of claim 1, wherein the relationship between sensor input and sensor output in each operating profile is characterized using a best fit equation.

10. A calibration algorithm generating a calibration curve for a linear sensor upon input of four data points,

wherein the calibration algorithm is electronically stored within a memory;

wherein the four data points correspond to test data obtained from the linear sensor based on a first and second sensor input at each of the first and second operating conditions, the four data points individually being represented as follows;

(I₁₁, O₁₁), (I₂₁, O₂₁), (I₁₂, O₁₂), and (I₂₂, O₂₂); and

wherein the calibration algorithm determines output of a linear sensor at a given operating condition, T, based on a given input, I, and wherein the algorithm conforms to the following mathematical equation;

Output=O₀+I/b,

 

wherein b=b₁+[(b₁−

b₂)/(T₁−

T₂)](T−

T₁), where b₁=(I₁₁−

I₂₁)/(O₁₁−

O₂₁);

b₂=(I₁₂−

I₂₂)/(O₁₂−

O₂₂); and

O₀=(O₀₁+O₀₂)/2, where O₀₁=O₁₁−

I₁₁/b₁ and O₀₂=O₁₂−

I₁₂/b₂

wherein the output of the linear sensor determined from the calibration algorithm is relayed to an electronic component.

11. The calibration algorithm of claim 10, wherein the calibration algorithm is electronically stored within a mechatronic unit.

12. A method of generating a calibration curve for a linear sensor, the method comprising:

providing the calibration algorithm of claim 10;

testing an individual sensor to generate four data points, wherein the four data points correspond to test data obtained based on a first and second input at each of a first and second operating condition; and

inputting the four data points into the calibration algorithm to generate the calibration curve for the linear sensor.

13. The method of claim 12, wherein the individual sensor is tested after final assembly into a mechatronic unit.

14. A mechatronic unit comprising:

at least one linear sensor; and

at least one electronic component comprising the calibration algorithm of claim 10 operatively coupled therein for calibration of the at least one sensor.

15. The mechatronic unit of claim 14, wherein the mechatronic unit comprises electrohydraulic components.

16. The mechatronic unit of claim 14, wherein the mechatronic unit comprises at least one solenoid operated pressure control valve.

17. The mechatronic unit of claim 14, Wherein the mechatronic unit operates to control flow of hydraulic fluid in automatic transmissions of motor vehicles.

18. The mechatronic unit of claim 14, wherein the linear sensor comprises a voltage output derived from an input selected from pressure, strain, temperature, and torque.

19. The mechatronic unit of claim 14, wherein the sensor comprises a pressure sensor.

20. An automatic transmission for a motor vehicle comprising the mechatronic unit of claim 14.

21. A method of analyzing accuracy of linear sensor output based on a theoretical prediction of the output model, the method comprising:

optionally analyzing a statistically significant number of linear sensors to obtain a first operating profile for each sensor at a first operating condition, T₁;

optionally analyzing the statistically significant number of linear sensors to obtain a second operating profile for each sensor at a second operating condition, T₂;

optionally confirming that each operating profile obtained represents a linear relationship between sensor input and sensor output using reciprocal slopes of output-input lines;

optionally confirming that operating profiles developed over a continuum of operating conditions represent a linear relationship using reciprocal slopes of each operating profile;

individually testing a linear sensor to obtain at least four data points, wherein the four data points correspond to test data obtained based on a first and second input at each of the first and second operating conditions and are represented as follows;

(I₁₁, O₁₁), (I₂₁, O₂₁), (I₁₂, O₁₂), and (I₂₂, O₂₂);

developing a theoretical prediction of the output model from the four data points and relaying the theoretical prediction of the output model to an electronic component, wherein the theoretical prediction of the output model for determining output at a given operating condition, T, and based on a given input, I, is developed in accordance with the following mathematical equation;

Output=O₀+I/b,

 

wherein b=b₁+[(b₁−

b₂)/(T₁−

T₂)](T−

T₁), where b₁=(I₁₁−

I₂₁)/(O₁₁−

O₂₁);

b₂=(I₁₂−

I₂₂)/(O₁₂−

O₂₂); and

O₀=(O₀₁+O₀₂)/2, where O₀₁=O₁₁−

I₁₁/b₁ and O₀₂=O₁₂−

I₁₂/b₂;

optionally comparing accuracy of the relationship between the theoretical prediction of the output model for the first operating condition to sensor output in the first operating profile for the first operating condition corresponding to the given input; and

optionally comparing accuracy of the relationship between the theoretical prediction of the output model for the second operating condition to sensor output in the second operating profile for the second operating condition corresponding to the given input.

22. A method of analyzing accuracy of linear sensor output based on a theoretical prediction of the output model, the method comprising:

obtaining a first operating profile for each sensor at a first operating condition, T₁;

obtaining a second operating profile for each sensor at a second operating condition, T₂;

optionally confirming that each operating profile obtained represents a linear relationship between sensor input and sensor output using reciprocal slopes of output-input lines;

optionally confirming that operating profiles developed over a continuum of operating conditions represent a linear relationship using reciprocal slopes of each operating profile;

individually testing a linear sensor to obtain at least four data points, wherein the four data points correspond to test data obtained based on a first and second input at each of the first and second operating conditions and are represented as follows;

(I₁₁, O₁₁), (I₂₁, O₂₁), (I₁₂, O₁₂), and (I₂₂, O₂₂);

developing a theoretical prediction of the output model from the four data points and relaying the theoretical prediction of the output model to an electronic component, wherein the theoretical prediction of the output model for determining output at a given operating condition, T, and based on a given input, I, is developed in accordance with the following mathematical equation;

Output=O₀I/b,

 

wherein b=b₁+[(b₁−

b₂)/(T₁−

T₂)](T−

T₁), where b₁=(I₁₁−

I₂₁)/(O₁₁−

O₂₁);

b₂=(I₁₂−

I₂₂)/(O₁₂−

O₂₂); and

O₀=(O₀₁+O₀₂)/2, where O₀₁=O₁₁−

I₁₁/b₁ and O₀₂=O₁₂−

I₁₂/b₂;

comparing accuracy of the relationship between the theoretical prediction of the output model for the first operating condition to sensor output in the first operating profile for the first operating condition corresponding to the given input; and

comparing accuracy of the relationship between the theoretical prediction of the output model for the second operating condition to sensor output in the second operating profile for the second operating condition corresponding to the given input.

23. A method of analyzing accuracy of linear sensor output based on a theoretical prediction of the output model, the method comprising:

optionally obtaining a first operating profile for each sensor at a first operating condition, T₁;

optionally obtaining a second operating profile for each sensor at a second operating condition, T₂;

optionally confirming that each operating profile obtained represents a linear relationship between sensor input and sensor output using reciprocal slopes of output-input lines;

optionally confirming that operating profiles developed over a continuum of operating conditions represent a linear relationship using reciprocal slopes of each operating profile;

individually testing a linear sensor to obtain at least four data points, wherein the four data points correspond to test data obtained based on a first and second input at each of the first and second operating conditions and are represented as follows;

(I₁₁, O₁₁), (I₂₁, O₂₁), (I₁₂, O₁₂), and (I₂₂, O₂₂);

developing a theoretical prediction of the output model from the four data points and relaying the theoretical prediction of the output model to an electronic component, wherein the theoretical prediction of the output model for determining output at a given operating condition, T, and based on a given input, I, is developed in accordance with the following mathematical equation;

Output=O₀+I/b,

 

wherein b=b₁+[(b₁−

b₂)/(T₁−

T₂)](T−

T₁), where b₁=(I₁₁−

I₂₁)/(O₁₁−

O₂₁),

b₂=(I₁₂−

I₂₂)/(O₁₂−

O₂₂); and

O₀=(O₀₁+O₀₂)/2, where O₀₁=O₁₁−

I₁₁/b₁ and O₀₂=O₁₂−

I₁₂/b₂;

optionally comparing accuracy of the relationship between the theoretical prediction of the output model for the first operating condition to sensor output in the first operating profile for the first operating condition corresponding to the given input; and

optionally comparing accuracy of the relationship between the theoretical prediction of the output model for the second operating condition to sensor output in the second operating profile for the second operating condition corresponding to the given input.