Dynamic prediction of risk levels for manufacturing operations through leading risk indicators

• US 8,884,759 B2
• Filed: 09/11/2012
• Issued: 11/11/2014
• Est. Priority Date: 09/12/2011
• Status: Active Grant

• Current Assignee: Near-Miss Management, LLC
• Original Assignee: Near-Miss Management, LLC
• Inventors: Oktem, Ulku G., Pariyani, Ankur
• Primary Examiner(s): HOFSASS, JEFFERY A
• Application Number:
  US13/610,446
• Publication Number:
  US 20130063264A1
• Time in Patent Office: 791 Days
• Field of Search: 340/540, 340/999, 702/179, 702/182, 702/183
• US Class Current: 340/540
• CPC Class Codes: G05B 23/024 Quantitative history assessment, e.g. mathematical relationships between available data; Functions therefor; Principal component analysis [PCA]; Partial least square [PLS]; Statistical classifiers, e.g. Bayesian networks, linear regression or correlation analysis; Neural networks
• CPC Class Codes: G05B 23/0278 Qualitative, e.g. if-then rules; Fuzzy logic; Lookup tables; Symptomatic search; FMEA
• CPC Class Codes: G06F 16/122 using management policies backup systems G06F11/1446; file migration policies for HSM systems G06F16/185
• CPC Class Codes: G08B 29/02 Monitoring continuously signalling or alarm systems

Abstract

Provided are methodologies to properly assess and manage operational risks at operations sites, e.g., a manufacturing, production or processing facility, such as a refinery, chemical plant, fluid-catalytic-cracking units, or nuclear energy plant, or a biological or waste management facility, airport or even financial institutions, or at any facility in which operations are often accompanied by risk associated with many high-probability, low-consequence events, often resulting in near-misses. In some operations, processes are monitored by alarms, but the invention operates on either process data or alarm data. The methods are based upon measurement of one or more variables, and/or utilization and management of the concept of “hidden process near-miss(es)” to identify a change or escalation, if any, in probability of occurrence of an adverse incident. The methodologies combine a plurality of subsets (also useful independently) of dynamically calculated leading risk indicators for dynamic risk management.

Claims

1. A method for identifying a hidden process near-miss in a plant/facility operation, the method comprising:

collecting measured data associated with at least one process variable, although without related adverse incident;

collecting long-term process data for a period preceding the collecting measured data wherein the long term process data is automatically updated over time;

determining normal values or value ranges, or combinations thereof, based on long-term process data, and comparing the measured data associated with the at least one process variable with the collected long-term process data to determine deviation there between measured data relative to determined normal value(s), wherein a measure or extent, or both, of the deviation identifies at least one hidden process near-miss.

2. The method of claim 1, further comprising measuring process variables automatically at a predetermined frequency, and then dating and reporting the measurements for comparison in real-time, on-demand, periodically, or in combinations thereof.

3. The method of claim 1, further comprising automatically issuing a leading signal alert indicating when deviation is greater than a predetermined normal value, together with information regarding the identified at least one process variable associated with the hidden process near-miss.

4. The method of claim 1, further comprising correlating extent of the hidden process near-miss deviation from normal operation of the at least one process variable in terms of risk of occurrence of an adverse incident.

5. The method of claim 1, further comprising characterizing the identified hidden process near miss or an observed near miss by determining scale, criticality, and frequency of occurrence or occurrence of an abnormal event.

6. The method of claim 5, further comprising determining one or more criterion for the at least one variable based upon its characterization in terms of frequency of occurrence of at least one abnormal event.

7. The method of claim 6, further comprising determining at least one dynamic risk index for each of the at least one variables by combining or averaging the risk levels defined by the one or more criterion.

8. The method of claim 7, further comprising for each of the one or more variables determining a compounded risk score by identifying a maximum risk zone that was entered and an associated grade assigned to the variable because of its entry.

9. The method of claim 4, wherein risk of occurrence of an adverse event is a function of the extent of deviation relative to the identified hidden process near-miss.

10. A method for determining operational fitness at a plant/facility, the method comprising:

identifying near-misses based upon information extracted from data provided by one or more measurable criteria for automatic detection and classification of observed near-misses and hidden process near-misses, the criteria identified in the group consisting of;

abnormal events analysis, time segment analysis, dynamic risk analysis; and

real-time leading signals calculation(s);

compiling and processing the identified information from the measurable criteria for dynamically and automatically determining change or escalation, on-demand, in real-time, periodically or in combinations thereof of the identified information as compared with that of corresponding data from a previous time period or measurement, or within a designated test period; and

reporting same.

11. The method of claim 10, wherein the report provides information enabling plant/facility operating personnel to take preemptive corrective action, comprising revising operating procedures at the plant/facility or process design, or both, to avoid occurrence of an adverse incident.

12. A method of claim 10, comprising further identifying near-misses based upon information extracted from data provided by the abnormal events analysis, wherein one or more observed near-misses or hidden process near-misses, or a combination thereof, is identified when one or more variables crosses a predetermined operational shutdown limit without a resulting shutdown, or when the frequency of one or more variables or the risk score or effective risk score associated with an abnormal event exceeds more than a predetermined fixed value or cutoff value in a predetermined time period, or combinations thereof, wherein the predetermined time period is a moving time period updated automatically.

13. A method of claim 10, further comprising, generating one or more real-time leading signals by monitoring trajectory of a variable and determining previously unforeseen percentage deviation(s) from running averages of the process values of that variable at predetermined time intervals;

and when percentage deviation from long-term data is greater than a predetermined cut-off value, assigning a violation tag, and generating a leading signal therefrom to forewarn plant/facility operators of one or more potential adverse incidents or shutdowns.

14. A method of claim 13, further comprising identifying near-misses based upon information extracted from data provided by the one or more real-time leading signals, wherein a variable exceeds predetermined discrete standard deviation levels, as compared with long-term operation data, training data or test data in terms of normal, or when risk status of an equipment or interlock changes or escalates above predetermined normal(s) or long-term level(s) as measured by a real-time risk indicator.

23. A method of claim 10, comprising further determining alarm fitness when operations and process variables are monitored at the plant/facility by a plurality of alarms in an alarm system, wherein one or more selected alarm-monitored process variables are each monitored by one or more alarms or an alarm system, the method comprising:

identifying near-misses based upon information extracted from data provided by one or more measurable criteria for automatic detection and classification of observed near-misses and hidden process near-misses, the criteria identified in, but not limited to, the group consisting of alarm frequency analysis, risk-based alarm data analysis, abnormal events analysis, time segment analysis, and dynamic risk analysis;

compiling and processing the identified information from the selected criteria for dynamically and automatically determining change or escalation, on-demand, in real-time, periodically or in combinations thereof of the identified information as compared with that of corresponding data from a previous moving time period or measurement, or within a designated test period; and

reporting same.

24. The method of claim 23, wherein the report provides information enabling plant/facility operating personnel to take preemptive corrective action, comprising revising operating procedures at the plant/facility or process design, or both, to avoid occurrence of an adverse incident.

25. A method of claim 23, wherein identifying information in near-miss data in the alarm based-monitoring system, provided by determining frequency and the alarm analysis criteria, further comprises determining measurable difference(s) from a predetermined fixed value or cutoff value, or when an alarm flood or an alarm attack occurs.

26. A method of claim 23, further comprising conducting one or more time segment analyses based on the time(s) when an alarm is activated or when the alarm is inactivated or deactivated, or a combination thereof, for the same variable, providing an orderly arrangement of the alarm data based upon deviation of the one or more variables from normal or long-term data, thereby identifying the highest risk alarm and associated variable.

27. A method of claim 23, further comprising classifying alarm fitness by conducting an advanced frequency analysis using a comparative view of significance levels to describe the identified near-miss data, identifying or classifying frequency of the comparative data to permit interpretation, and analyzing the data within a given region and between regions for one or more time periods.

28. A method of claim 27, further comprising identifying notable alarms using criteria identified in but not limited to the group consisting of alarm frequency analysis and time segment analysis.

15. A method for identifying a hidden process near-miss in an alarm-monitored plant/facility operation, the method comprising:

collecting measured alarm data associated with at least one alarm-monitored process variable, although without related adverse incident;

collecting long-term alarm data for a period preceding the collecting measured alarm data; and

determining normal values or value ranges, or combinations thereof, based on long-term alarm data comparing the measured alarm data associated with the at least one alarm-monitored process variable with the collected long-term alarm data to determine deviation between the collected alarm data relative to determined normal value(s), wherein a measure and extent of the deviation identifies and characterizes at least one alarm data-related hidden process near-miss.

16. The method of claim 15, further comprising measuring alarm-data for an alarm-monitored process variable automatically as alarm-data becomes available, and then dating and reporting the measured alarm-data for comparison in real-time, on-demand, periodically, or in combinations thereof.

17. The method of claim 15, further comprising correlating extent of the alarm-data related hidden process near-miss deviation from normal alarm-monitored operation of the at least one process variable in terms of risk of occurrence of an adverse incident.

18. The method of claim 15, further comprising characterizing the identified alarm-monitored hidden process near miss or an observed near miss by determining scale, criticality, and frequency of alarm occurrence or of occurrence of an abnormal event.

19. The method of claim 18, further comprising determining one or more criterion for the at least one alarm-monitored variable, based upon its characterization in terms of frequency of occurrence of one or more abnormal events.

20. The method of claim 19, further comprising determining at least one dynamic risk index for each of the at least one alarm-monitored variables by combining or averaging the risk levels defined by the one or more criterion.

21. The method of claim 20, further comprising for each of the one or more alarm-monitored variables determining a compounded risk score by identifying, time spent beyond the normal alarm level for the variable, a maximum risk zone that was entered and an associated grade assigned to the variable because of its alarm value.

22. The method of claim 17, wherein quantitatively a risk of occurrence of an adverse event is a function of the extent of deviation relative to an identified hidden process near-miss.