Back to results
US 7,030,746 B2Granted
Method and system for generating automatic alarms based on trends detected in machine operation
AssigneeGeneral Electric Company
InventorsNathan Bowman Littrell, John Wesley Grant
Life of this patent
Estimating the term…
Abstract
A method to detect a change in a condition of a system including: sensing the condition of the system and capturing data values indicative of the sensed condition; periodically determining a difference between a first data value and a second data value; determining a characteristic representative of a series of differences representative of a predetermined period; comparing the characteristic to at least one threshold level, and generating an alarm if the characteristic is beyond the threshold level.
Claims27
A method to detect a change in a condition of a system comprising:
- a. sensing the condition of the system and capturing data values indicative of the sensed condition;
- b. periodically determining a difference between a first data value and a second data value;
- c. determining a characteristic occurring during a series of differences representative of a predetermined period;
- d. comparing the characteristic to at least one threshold level, and
- e. generating an alarm if the characteristic is beyond the threshold level.
2 more elements
A method to detect a change in a condition of a system comprising:
- a. sensing the condition of the system and capturing data values indicative of the sensed condition;
- b. periodically determining a difference between a first data value and a second data value;
- c. determining a characteristic representative of a series of differences occurring during a predetermined period, wherein the characteristic representative of the series of differences is a mean value of the differences;
- d. comparing the characteristic to at least one threshold level, and
- e. generating an alarm if the characteristic is beyond the threshold level.
2 more elements
A method to detect a change in a condition of a system comprising:
- a. sensing the condition of the system and capturing data values indicative of the sensed condition;
- b. periodically determining a difference between a first data value and a second data value;
- c. determining a characteristic representative of a series of differences occurring during a predetermined period, wherein the characteristic representative of the series of differences is a median value of the differences;
- d. comparing the characteristic to at least one threshold level, and
- e. generating an alarm if the characteristic is beyond the threshold level.
2 more elements
A method to detect a change in a condition of a system comprising:
- a. sensing the condition of the system and capturing data values indicative of the sensed condition;
- b. periodically determining a difference between a first data value and a second data value;
- c. determining a characteristic representative of a series of differences occurring during a predetermined period, wherein the characteristic representative of the series of differences is a standard deviation of the differences;
- d. comparing the characteristic to at least one threshold level, and
- e. generating an alarm if the characteristic is beyond the threshold level.
2 more elements
A method to detect a change in a condition of a system comprising:
- a. sensing the condition of the system and capturing data values indicative of the sensed condition;
- b. periodically determining a difference between a first data value and a second data value;
- c. determining a characteristic representative of a series of differences occurring during a predetermined period;
- d. comparing the characteristic to at least one threshold level;
- e. generating an alarm if the characteristic is beyond the threshold level, and
- f. temporarily storing the series of differences in a buffer and setting a buffer size to correspond to the predetermined period.
3 more elements
A method to detect a change in a condition of an operating system comprising:
- a. sensing the condition of the system and capturing data values indicative of the sensed condition;
- b. periodically determining a difference between a first data value and a second data value;
- c. determining a characteristic representative of a series of differences occurring during a predetermined period;
- d. comparing the characteristic to a plurality of threshold levels, wherein each level corresponds to a severity alarm level, and
- e. generating the severity alarm level if the characteristic is beyond the corresponding threshold level and if the system is at a predetermined operating mode during a period corresponding to the series of differences.
2 more elements
A method to detect a change in a condition of an operating system comprising:
- a. sensing the condition of the system and capturing data values indicative of the sensed condition;
- b. periodically determining a difference between a first data value and a second data value;
- c. determining a characteristic representative of a series of differences occurring during a predetermined period;
- d. comparing the characteristic to a plurality of threshold levels, wherein each level corresponds to a severity alarm level, and
- e. generating the severity alarm level if the characteristic is beyond the corresponding threshold level and if the system is at a predetermined operating mode during a period corresponding to the series of differences, and
- f. disabling the severity alarm level generation if a number of invalid differences exceeds a predetermined number.
3 more elements
A method to detect a change in a condition of an operating system comprising:
- a. sensing the condition of the system and capturing data values indicative of the sensed condition;
- b. periodically determining a difference between a first data value and a second data value;
- c. determining a characteristic representative of a series of differences occurring during a predetermined period, wherein the characteristic representative of the series of differences is an mean value of the differences;
- d. comparing the characteristic to a plurality of threshold levels, wherein each level corresponds to a severity alarm level, and
- e. generating the severity alarm level if the characteristic is beyond the corresponding threshold level and if the system is at a predetermined operating mode during a period corresponding to the series of differences.
2 more elements
A method to detect a change in a condition of an operating system comprising:
- a. sensing the condition of the system and capturing data values indicative of the sensed condition;
- b. periodically determining a difference between a first data value and a second data value;
- c. determining a characteristic representative of a series of differences occurring during a predetermined period, wherein the characteristic representative of the series of differences is a median value of the differences;
- d. comparing the characteristic to a plurality of threshold levels, wherein each level corresponds to a severity alarm level, and
- e. generating the severity alarm level if the characteristic is beyond the corresponding threshold level and if the system is at a predetermined operating mode during a period corresponding to the series of differences.
2 more elements
A method to detect a change in a condition of an operating system comprising:
- a. sensing the condition of the system and capturing data values indicative of the sensed condition;
- b. periodically determining a difference between a first data value and a second data value;
- c. determining a characteristic representative of a series of differences occurring during a predetermined period, wherein the characteristic representative of the series of differences is a standard deviation of the differences;
- d. comparing the characteristic to a plurality of threshold levels, wherein each level corresponds to a severity alarm level, and
- e. generating the severity alarm level if the characteristic is beyond the correspondina threshold level and if the system is at a predetermined operating mode during a period corresponding to the series of differences.
2 more elements
A method to detect a change in a condition of an operating system comprising:
- a. sensing the condition of the system and capturing data values indicative of the sensed condition;
- b. periodically determining a difference between a first data value and a second data value;
- c. determining a characteristic representative of a series of differences occurring during a predetermined period;
- d. comparing the characteristic to a plurality of threshold levels, wherein each level corresponds to a severity alarm level, and
- e. generating the severity alarm level if the characteristic is beyond the corresponding threshold level and if the system is at a predetermined operating mode during a period corresponding to the series of differences, and
- f. temporarily storing the series of differences in a buffer and setting a buffer size to correspond to the predetermined period.
3 more elements
A trend detection system comprising:
- a sensor monitoring a condition of an operating machine or process and generating sensor data;
- a computer controller receiving the sensor data and periodically storing said sensor data, said controller further comprising:
- a latch temporarily storing a first sensor data value and a second sensor data value;
- a difference logic function determining a difference value between the first and second sensor data values;
- a buffer storing a sequence of difference values inputted from the difference logic function;
- a logic function determining a characteristic of the sequence of difference values in said buffer;
- a comparator logic function determining whether said characteristic is beyond a predetermined threshold level, and
- a severity level alarm function generated if said comparator logic determines the characteristic is beyond the threshold level.
5 more elements
Dependencies are inferred from the claim text.