Motor vehicle fuel analyzer

• US 5,654,497 A
• Filed: 06/03/1996
• Issued: 08/05/1997
• Est. Priority Date: 03/03/1992
• Status: Expired due to Fees

• Current Assignee: Lockheed Martin Energy Systems, Inc. (French Alternative Energies and Atomic Energy Commission)
• Original Assignee: Lockheed Martin Energy Systems, Inc. (French Alternative Energies and Atomic Energy Commission)
• Inventors: Hoffheins, Barbara S., Lauf, Robert J.
• Primary Examiner(s): BROCK, MICHAEL J
• Application Number:
  US08/655,664
• Time in Patent Office: 428 Days
• Field of Search: 73/23.31, 73/23.2, 73/23.41, 73/31.05, 395/21, 395/22, 395/24, 395/27, 364/498
• US Class Current: 73/23.2
• CPC Class Codes: G01N 33/2829 Mixtures of fuels

Abstract

A gas detecting system for classifying the type of liquid fuel in a container or tank. The system includes a plurality of semiconductor gas sensors, each of which differs from the other in its response to various organic vapors. The system includes a means of processing the responses of the plurality of sensors such that the responses to any particular organic substance or mixture is sufficiently distinctive to constitute a recognizable "signature". The signature of known substances are collected and divided into two classes based on some other known characteristic of the substances. A pattern recognition system classifies the signature of an unknown substance with reference to the two user-defined classes, thereby classifying the unknown substance with regard to the characteristic of interest, such as its suitability for a particular use.

Claims

1. A system for classifying mixtures of hydrocarbon fuels according to octane or cetane number comprising:

an array of gas sensors having different response characteristics to the vapors of different known and unknown fuels such that the set of responses for each of said fuels forms a distinctive signature pattern with respect to octane or cetane number of said fuels, said array comprising at least four sensors;

a data acquisition system coupled to said sensor array for deriving said distinctive signature patterns of said known fuels, said data acquisition system comprising at least an analog-to-digital convertor; and

a data analysis system comprising at least a neural network coupled to said data acquisition system for comparing the pattern of a sampled fuel mixture to the signature pattern of said known fuels whereby said sampled signature pattern of octane or cetane number is classified with respect to said known patterns of octane or cetane numbers by said neural network.

2. The system of claim 1 wherein said data acquisition system and said data analysis system are incorporated in a digital computer, and wherein said analog-to-digital convertor is a multiple channel analog-to-digital circuit within said digital computer, said analog-to-digital circuit being connected to said gas sensors to receive said set of responses.

3. The system of claim 2 wherein said data acquisition system further includes a bit pattern conversion program running in said digital computer.

4. The system of claim 2 wherein said neural network is a software neural network program operating within said digital computer.

5. The system of claim 2 wherein said data acquisition system further includes a computer program controlling the acquisition of gas sensor data by said analog-to-digital convertor.

6. The system of claim 1 wherein at least one of said gas sensors is an independently heated gas sensor.

7. The system of claim 1, wherein the differences in said response characteristics of at least some of said gas sensors result from differences in their preparation.

8. The system of claim 1, wherein the differences in said response characteristics of at least some of said gas sensors result from differences in their operating temperature.

9. The system of claim 1, wherein at least one of said gas sensors is a semiconductor sensor.

10. The system of claim 1, wherein at least one of said gas sensors is a solid state sensor.

11. The system of claim 1, wherein at least one of said gas sensors is a homogeneous semiconductor gas sensor.

12. The system of claim 1, wherein the differences in said response characteristics of at least some of said gas sensors result from differences in their composition.

13. The system of claim 1, wherein said neural network is a software neural network program stored on a programmable memory chip.

14. The system of claim 1, wherein said data analysis system is an application-specific integrated circuit comprising at least a neural network processor.

15. The system of claim 1 wherein said array of gas sensors is fabricated onto a single substrate.

16. A system according to claim 1, wherein said array of gas sensors is fabricated into a single sensor head.

17. The system of claim 16 further including a plurality of said sensor heads, each in separate locations dispersed about an environment, and means responsive to signals derived from all of said sensors of said heads to determine the identities fuel mixtures present at each of said locations.

18. The system of claim 1 wherein said known fuel mixtures are taken from the group of liquid fuels comprising gasoline, diesel fuel, aviation fuel, and gasohol.

19. The system of claim 18 wherein said array of gas sensors additionally have different response characteristics to the vapors of different known and unknown fuels such that the set of responses for each of said fuels additionally forms a distinctive signature pattern with respect to oxygenate or deicer content and wherein said neural network can additionally classify said sampled signature pattern with respect to said know patterns of oxygenate or deicer content.

20. The system of claim 1 wherein said gas sensor array comprises metal oxide gas sensors.

21. The system of claim 1 wherein said data acquisition system further includes a bit pattern conversion means for converting said signature patterns into bit patterns.

22. The system of claim 1 wherein said neural network is a Hamming network.

23. The system of claim 1 wherein said neural network is a Boltzmann network.

24. The system of claim 1 wherein said array of gas sensors includes at least ten sensors and wherein at least one of said sensors is dissimilar from the other sensors.

25. A method identifying mixtures of hydrocarbon fuels according to octane or cetane number comprising the steps of:

selecting an array of gas sensors, each of said sensors having different response characteristics to the vapors of different known fuels such that the set of sensor responses to each of said known fuels forms a distinctive signature pattern with respect to the octane or cetane number of said fuels, said array of sensors comprising at least four sensors;

deriving said signature patterns of said known fuels by analog-to-digital conversion means; and

analyzing the signature pattern of a sampled fuels with respect to the signature patterns of said known fuels by neural network means whereby said sampled signature pattern of octane or cetane numbers is classified with respect to said known signature patterns of octane or cetane numbers.

26. The method of claim 25 including the further step of controlling said analog-to-digital conversion means by a data acquisition computer program.

27. The method of claim 25 wherein said neural network means is a software neural network program operating within a digital computer.

28. The method of claim 25 wherein said known fuel mixtures are taken from the group of liquid fuels comprising gasoline, diesel fuel, aviation fuel, and gasohol.

29. The method of claim 25 including the further step of converting said signature patterns to bit patterns by a bit pattern conversion means.