WO2014177902A1 - Method for monitoring a pumping device - Google Patents
Method for monitoring a pumping device Download PDFInfo
- Publication number
- WO2014177902A1 WO2014177902A1 PCT/IB2013/001277 IB2013001277W WO2014177902A1 WO 2014177902 A1 WO2014177902 A1 WO 2014177902A1 IB 2013001277 W IB2013001277 W IB 2013001277W WO 2014177902 A1 WO2014177902 A1 WO 2014177902A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- actual measured
- data sets
- damage
- measured data
- difference
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0218—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
- G05B23/0224—Process history based detection method, e.g. whereby history implies the availability of large amounts of data
- G05B23/0227—Qualitative history assessment, whereby the type of data acted upon, e.g. waveforms, images or patterns, is not relevant, e.g. rule based assessment; if-then decisions
- G05B23/0235—Qualitative history assessment, whereby the type of data acted upon, e.g. waveforms, images or patterns, is not relevant, e.g. rule based assessment; if-then decisions based on a comparison with predetermined threshold or range, e.g. "classical methods", carried out during normal operation; threshold adaptation or choice; when or how to compare with the threshold
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0259—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterized by the response to fault detection
- G05B23/0267—Fault communication, e.g. human machine interface [HMI]
- G05B23/0272—Presentation of monitored results, e.g. selection of status reports to be displayed; Filtering information to the user
Definitions
- the invention relates to a method for monitoring a pumping device for a fluid medium by means of a computing device, comprising the following steps:
- Presently usual methods for monitoring pumping devices can be divided in four categories.
- the basic output of most of the monitoring packages provides alarms based on hitting a specified high or low point of interest, or respectively based on exceeding or falling below a certain threshold.
- Alarms can be set to protect the system from operating overloaded, underloaded, or e.g. running too hot.
- Another aspect is to ensure the protection from excessively high or low pressures, such as e.g. well-head pressures, which can be critical for ESPs, if they are used for oil production from an oil well.
- well-head pressures which can be critical for ESPs
- the so-called "trend diagnostics exception” reporting belongs to the third category. Based on performance trend changes a software provides the recognition of an exceptional status. The identified exceptions are transmitted to the operating personnel, whereat a decision may be taken, if any action would help prevent or reduce the chance of a shutdown or poor operating conditions of the ESP in the near future. Thus the maintenance of the production can be ensured, as well as the proper operation of the pump system.
- the program highlights wells having significant change in downhole pressure, power draw, temperature change, system loading, power quality, pump design operation outside of suggested range, pump discharge change, system lift efficiency and other defined exception criteria.
- Diagnostics analysis brings together measured data, proprietary data model and expert knowledge to contribute to the diagnosis of problems in an ESP application. It applies "fuzzy logic" to identify patterns in measured data. The operating personnel is notified of potential problems along with the measure of how closely the measured data matches the pattern for a problem.
- a difference of at least one of the actual measured data sets from the normal state datasets captured during the operation of the pumping device is determined.
- Deriving a difference between actual measured data sets and normal state data sets provides a very fast possibility to detect anomalies related to the pumping device. Especially in comparison with "trend diagnostics exceptions” and “diagnostics analysis” the effort for determining a deviation between actual measured data sets and normal state data sets is very small, since unlike these reporting systems the result, thus the difference, is available very quickly, whereas the comparison of the actual measured data sets with defined exception criteria, respectively the identification of patterns in the measured data is very time
- the established difference is used for the detection of types of damage differing from the standard types of damage.
- new respectively unknown types of damages can be detected.
- those damages can be detected, which have not been previously defined. Therefore the method is particularly reliable since also unforeseen events can be detected.
- the threshold value can be set to both small values, where high reliability is ensured, whereat the warning message is communicated when small changes of measured data have to be supervised exactly, and it can be set to high values, if the tolerance of the measured data is e.g. of low relevance for the maintenance of the operation and thus the warning message is communicated infrequently.
- the operating personnel can be informed about a present anomaly particularly fast.
- Several alarms are useful to warn both visually and audibly when the threshold is reached. This will ensure more substantially, that any faults or damages are observed.
- the actual measured data set is assigned to the normal state data sets by means of an interface.
- the operating personnel can respond particularly effective to alarms. If the present alarm is a false alarm, the interface enables the operating personnel to classify the actual measured data set, which is related to the false alarm, as a normal state data set. In other words, the operating personnel can overrule the computing device by means of the interface whereat the actual measured data set, which were incorrectly assigned to the false alarm firstly, can be stored to the normal state data sets by the operating personnel.
- the data characterizing the at least one type of damage by means of the interface are assigned to a data set characterizing the standard types of damage.
- the operating personnel also has the possibility to add new types of damages to the data set characterizing the standard types of damage.
- the operating personnel can add data sets of previously unknown anomalies which are related to warning message to the data set characterizing the standard types of damage.
- the data set concerning the standard types of damage can be extended by further data sets, which are related to a previously unknown anomaly that can cause damage to the pumping device.
- FIG 1 a structure which is equipped with several sensors, whereat a sensor
- IG 2 the monitoring of measured values according to the prior art, whereat different anomalies can be detected by means of analysing at least one measured value and whereat the anomalies, which are related to the at least one measured value are assessed by means of a computing device;
- FIG 3 the computing device which is comprising an artificial intelligence, whereat the computing device determines a difference which is related to the anomalies;
- FIG 4 the computing device, which is comprising the artificial intelligence, whereat the computing device is retrained due to a false alarm detection.
- FIG 1 shows a structure 1 , whereat several sensors 2 are attached to the structure 1.
- Each of the sensors 2 communicates actual measured data 9 to a sensor measurement 3.
- the actual measured data 9 are related to physical variables concerning the structure 1 and the sensor measurement 3 corresponds to a measurement data acquisition system.
- the sensor measurement 3 communicates the actual measured data 9 to both a computing device 10 and feature extraction 4.
- the feature extraction 4 comprises known states of the structure 1, whereat the states correspond to the structure's behaviour under the influence of environmental conditions, as well as a risk evaluation, which determines the risk for the structure 1 according to the actual measured data 9. Both the data set with the known behaviour from the feature extraction 4 and the actual measured data 9 are communicated to the computing device 10.
- the computing device 10 comprises a committee 5, a data set concerning normal behaviour 6 of the structure 1 , a data set concerning abnormal behaviour 7 of the structure 1 , and an artificial intelligence 15.
- the committee 5 evaluates the state of the structure 1 by means of the data set concerning the normal behaviour 6 and a data set concerning the abnormal behaviour 7 of the structure 1.
- the data set concerning the normal behaviour 6 corresponds to measurement data from the sensors 2, which are related to a normal state of the structure 1 , whereat the measurement data have been recorded in the past.
- the committee 5 determines a confidence value 8, whereat the confidence value 8 is used to evaluate the state of the structure 1.
- the monitoring of a pumping device is schematically illustrated in FIG 2.
- the computing device 10 determines, whether an anomaly exists, or not.
- the computing device 10 comprises a trend diagnostics 14, a pattern 16, a data set concerning the actual measured data 9, a nominal data set 1 1 and a normal state data set 12.
- the trend data diagnostics 14 determines performance trend changes based on the actual measured data 9. In other words the trend diagnostics 14 is capable to determine whether a trend anomaly 23 has occurred.
- a trend anomaly is an anomaly which is derived by comparing the time course of the actual measured data 9 with the time course of normal state data sets 12.
- the output data of the trend diagnostics 14 indicates exceptional behaviour, e.g. significant pressure change of an electric submersible pump, or an operating point of the pump which can lead to a pump failure. If a trend anomaly 23 is detected by the trend diagnostics 14 a warning message 13 related to the trend anomaly 23 is communicated to the operating personnel.
- the operating personnel which operates a SCAD A system 17 (Supervisory Control And Data Acquisition) can decide, whether any action would help prevent or reduce the chance of a shutdown in the near future, thus maintaining production or e.g. ensuring proper operation of the ESP. Both the output data and the warning message 13 which is related to the trend anomaly 23 are also communicated to the SCAD A system.
- the SCADA system 17 is capable of providing monitoring data to the operating personnel. In other words by means of the SCADA system 17, the operating personell can supervise the pumping device.
- the SCADA system 17 also comprises an interface by means of which the operating personell can communicate with the computing device 10.
- the normal state data set 12 comprises already known and normal conditions of the pumping device.
- the nominal data set 11 comprises nominal values of the actual measured data 9.
- a threshold is determined by the computing device 10, which defines a maximum and a minimum value of the normal state data set 12.
- the pattern 16 is identified by a diagnostic analysis 26, which brings together the actual measured data 9, a proprietary data model and expert knowledge to contribute to the diagnosis of problems.
- the diagnostic analysis 26 applies "fuzzy logic" to identify patterns 16 in the actual measured data 9.
- the operating personnel is notified of potential problems along with a measure of how closely the actual measured data 9 matches the pattern 16 related to a problem respectively an anomaly.
- the pattern 16 comprises a segment of time series by means of which a certain deviation of the normal state data set 12 is determined.
- a time-series-threshold which is a limit value by which the time-series of the actual measured data 9 may differ from the time-series of the normal state data set 12
- a time series anomaly 24 is detected, whereat both the warning message 13 and a data set, which is related to the time series anomaly 24 is communicated to the SCADA system 17.
- FIG 3 shows the contents of FIG 2 in substantial parts and therefore only the differences are described.
- the computing device 10 is extended by an artificial intelligence 15. Based on a comparison between the normal state data set 12 and the actual measured data 9 the artificial intelligence 15 determines a difference 18.
- the difference 18 is independent of both the trends diagnostics 14 and the pattern 16, which was derived from diagnostics analysis 26.
- the artificial intelligence 15 is capable to determine a difference anomaly 25, whereat the difference anomaly 25 comprises anomalies, which are unknown to both the trend diagnostics 14 and the diagnostics analysis 26.
- FIG 3 illustrates a situation, where both the trend diagnostic 14 and the diagnostic analysis 26 determine a normal behaviour 6 whereas the artificial intelligence 15 determines a difference anomaly 25 and communicates both the data set concerning the different anomaly 25 and the warning message 13 to the operating personnel which operates the SCADA system 17.
- the artificial intelligence 15 increases the reliability of the monitoring, since the artificial intelligence 15 is capable of detecting new kinds of anomalies, which are unknown to both the trend diagnostic 14 and the diagnostics analysis 26.
- FIG 4 shows the contents of FIG 2 respectively FIG 3 in substantial parts and therefore only the difference is described.
- FIG 4 describes two different states, wherein on the one hand the artificial intelligence 15 has successfully detected an anomaly and on the other hand the anomaly which was detected by the artificial intelligence 15 is a false alarm 21, which is communicated to the SCADA system 17.
- the operating personnel initiates a data set extension 22 by means of the SCADA system, whereat both the trend diagnostic 14 and the diagnostic analysis 26, is extended by a data set, which is related to a previously unknown anomaly, that was detected by the artificial intelligence 15.
- the anomaly, which was detected by the artificial intelligence 15 is declared as a false alarm 21 by the operating personnel, a retraining 19 of the artificial intelligence 15 is initiated by the operating personnel to prevent the artificial intelligence 15 from causing the same or a similar false 21 alarm again in the future.
- the diagnostics analysis 26, which comprises the pattern 16 initiates a diagnostic retraining 20, by means of which it is ensured, that the artificial intelligence 15 is provided with a data set concerning the pattern 16 of the false anomaly detection which caused the false alarm 21.
- the artificial intelligence 15 can be both retrained by the operating personnel and by the diagnostics analysis26.
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention relates to a method for monitoring a pumping device for a fluid medium by means of a computing device (10), comprising the following steps: capturing sets of actual measured data (9) of a pumping device characterizing an actual state; comparing the respective sets of actual measured data (9) to respective, stored nominal data sets (11) and, if applicable, capturing a deviation of at least one set of actual measured data (9) from a nominal data set (11); and in case a deviation has been detected: associating a deviation with at least one standard type of damage. The computing device (10) determines a difference (18) of at least one set of the actual measured data (9) from the normal state data sets (12) captured during the operation of the pumping device.
Description
Method for monitoring a pumping device
Technical Field
The invention relates to a method for monitoring a pumping device for a fluid medium by means of a computing device, comprising the following steps:
- capturing actual measured datasets of the pumping
device characterizing an actual state;
- comparing the respective actual measured datasets to
respective, stored nominal data sets and, if
applicable, capturing a deviation of at least one
actual measured data set from a nominal data set; and
- in case a deviation has been detected:
associating the deviation with at least one standard
type of damage.
Background Art
Presently usual methods for monitoring pumping devices, such as electric submersible pumps (ESP), can be divided in four categories. The basic output of most of the monitoring packages provides alarms based on hitting a specified high or low point of interest, or respectively based on exceeding or falling below a certain threshold. Alarms can be set to protect the system from operating overloaded, underloaded, or e.g. running too hot. Another aspect is to ensure the protection from excessively high or low pressures, such as e.g. well-head pressures, which can be critical for ESPs, if they are used for oil production from an oil well. As one of the four categories "alarming" from critical conditions corresponds to a reactive reporting, so the operating personell can only react on critical conditions, since the alarm was issued, when the problem has already occured.
Another example for reactive reporting is caused by the so-called "downtime". In this case the operating personnel is notified of what has already happened, which now requires action to
improve the situation. In other words downtime has already occurred and requires the reaction of the operating personnel.
The so-called "trend diagnostics exception" reporting belongs to the third category. Based on performance trend changes a software provides the recognition of an exceptional status. The identified exceptions are transmitted to the operating personnel, whereat a decision may be taken, if any action would help prevent or reduce the chance of a shutdown or poor operating conditions of the ESP in the near future. Thus the maintenance of the production can be ensured, as well as the proper operation of the pump system. The program highlights wells having significant change in downhole pressure, power draw, temperature change, system loading, power quality, pump design operation outside of suggested range, pump discharge change, system lift efficiency and other defined exception criteria.
The last category of reporting strategies is referred to as a diagnostics analysis. Diagnostics analysis brings together measured data, proprietary data model and expert knowledge to contribute to the diagnosis of problems in an ESP application. It applies "fuzzy logic" to identify patterns in measured data. The operating personnel is notified of potential problems along with the measure of how closely the measured data matches the pattern for a problem.
Summary of Invention
The presented solutions according to the prior art bear several disadvantages, whereat in the case of alarming or downtime which are related to a reactive reporting, the operating personnel is informed about the malfunction e.g. damage of the pumping device if it is imminent or has already occurred. In other words, the prohibition of a system failure is impossible, since the reporting is reactive. Furthermore both the "trend diagnostics exceptions" and the "diagnostics analysis" can only predict an imminent failure, if the present exception criteria respectively the corresponding pattern of the exception criteria is known to the program.
Therefore, it is the object of the present invention to provide a particularly effective and at the same time particularly reliable method of the initially mentioned kind.
This object is solved by a method having the features of claim 1. Advantageous configurations with convenient developments of the invention are specified in the dependent claims.
In the method according to the invention, a difference of at least one of the actual measured data sets from the normal state datasets captured during the operation of the pumping device is determined.
Deriving a difference between actual measured data sets and normal state data sets provides a very fast possibility to detect anomalies related to the pumping device. Especially in comparison with "trend diagnostics exceptions" and "diagnostics analysis" the effort for determining a deviation between actual measured data sets and normal state data sets is very small, since unlike these reporting systems the result, thus the difference, is available very quickly, whereas the comparison of the actual measured data sets with defined exception criteria, respectively the identification of patterns in the measured data is very time
consuming.
In an advantageous configuration of the invention the established difference is used for the detection of types of damage differing from the standard types of damage. In contrast to the solutions according to the prior art by means of the established difference new respectively unknown types of damages can be detected. In other words due to the usage of the established difference also those damages can be detected, which have not been previously defined. Therefore the method is particularly reliable since also unforeseen events can be detected.
It has proven further advantageous, if in case the difference of the data of the nominal data set differs by a pre-determined threshold value, a warning message is issued.
The threshold value can be set to both small values, where high reliability is ensured, whereat the warning message is communicated when small changes of measured data have to be supervised exactly, and it can be set to high values, if the tolerance of the measured data is e.g.
of low relevance for the maintenance of the operation and thus the warning message is communicated infrequently.
By means of a warning message the operating personnel can be informed about a present anomaly particularly fast. Several alarms are useful to warn both visually and audibly when the threshold is reached. This will ensure more substantially, that any faults or damages are observed.
It is further advantageous, if in case the warning message is a false alarm the actual measured data set is assigned to the normal state data sets by means of an interface.
By means of an interface, the operating personnel can respond particularly effective to alarms. If the present alarm is a false alarm, the interface enables the operating personnel to classify the actual measured data set, which is related to the false alarm, as a normal state data set. In other words, the operating personnel can overrule the computing device by means of the interface whereat the actual measured data set, which were incorrectly assigned to the false alarm firstly, can be stored to the normal state data sets by the operating personnel.
It is particularly advantageous, if in case the warning message concerns at least one of the types of damage, the data characterizing the at least one type of damage by means of the interface are assigned to a data set characterizing the standard types of damage.
By means of the interface, the operating personnel also has the possibility to add new types of damages to the data set characterizing the standard types of damage. In other words, the operating personnel can add data sets of previously unknown anomalies which are related to warning message to the data set characterizing the standard types of damage. Hence the data set concerning the standard types of damage can be extended by further data sets, which are related to a previously unknown anomaly that can cause damage to the pumping device.
Furthermore, it is advantageous, if for determining the difference of the data sets
characterizing the types of damage are used, as well as the normal state data sets.
Due to the use of multiple data sets that refer to both types of damage, as well as the normal data sets, the difference is determined with particularly great confidence level. Issuing a warning message is therefore accordingly initiated only, if an anomaly is detected with a certain confidence.
It has proven particularly advantageous, if the normal state data sets and the difference are determined by an artificial intelligence.
Since artificial intelligences do not simulate the interaction of complex physical relationships, but rather interpret the correlations between input and output variables, the calculation results are obtained very quickly. Furthermore, systems of artificial intelligence can interpret and represent linear and non-linear relations which can be derived from present data sets, and can learn from both the input data provided by the operating personell and other monitoring systems which can be used parallel. If a system of artificial intelligence is used in addition to other monitoring systems, a redundancy is established, which ensures a higher reliability.
The features and feature combinations mentioned above in the description as well as the features and feature combinations mentioned below in the description of figures and/or shown in the figure alone are usable not only in the respectively specified combination, but also in other combinations or alone without departing from the scope of the invention.
Further advantages, features and details of the invention are apparent from the claims, the following description of preferred embodiments as well as from the drawings.
Brief Description of Drawings
The drawings show in:
FIG 1 a structure which is equipped with several sensors, whereat a sensor
measurement is used to detect, whether a normal behaviour or an abnormal behaviour of the structure exists, which is according to the prior art;
IG 2 the monitoring of measured values according to the prior art, whereat different anomalies can be detected by means of analysing at least one measured value and whereat the anomalies, which are related to the at least one measured value are assessed by means of a computing device;
FIG 3 the computing device which is comprising an artificial intelligence, whereat the computing device determines a difference which is related to the anomalies; and
FIG 4 the computing device, which is comprising the artificial intelligence, whereat the computing device is retrained due to a false alarm detection.
Description of Embodiments
According to the prior art, FIG 1 shows a structure 1 , whereat several sensors 2 are attached to the structure 1. Each of the sensors 2 communicates actual measured data 9 to a sensor measurement 3. The actual measured data 9 are related to physical variables concerning the structure 1 and the sensor measurement 3 corresponds to a measurement data acquisition system. The sensor measurement 3 communicates the actual measured data 9 to both a computing device 10 and feature extraction 4. The feature extraction 4 comprises known states of the structure 1, whereat the states correspond to the structure's behaviour under the influence of environmental conditions, as well as a risk evaluation, which determines the risk for the structure 1 according to the actual measured data 9. Both the data set with the known behaviour from the feature extraction 4 and the actual measured data 9 are communicated to the computing device 10. The computing device 10 comprises a committee 5, a data set concerning normal behaviour 6 of the structure 1 , a data set concerning abnormal behaviour 7 of the structure 1 , and an artificial intelligence 15. The committee 5 evaluates the state of the structure 1 by means of the data set concerning the normal behaviour 6 and a data set concerning the abnormal behaviour 7 of the structure 1. The data set concerning the normal behaviour 6 corresponds to measurement data from the sensors 2, which are related to a normal state of the structure 1 , whereat the measurement data have been recorded in the past. By means of the data set of the normal behaviour 6 and the abnormal behaviour 7 the
committee 5 determines a confidence value 8, whereat the confidence value 8 is used to evaluate the state of the structure 1.
The monitoring of a pumping device according to the prior art is schematically illustrated in FIG 2. Based on a data set corresponding to actual measured data 9 of at least one sensor 2, which determines at least one measurement value concerning the pumping device, the computing device 10 determines, whether an anomaly exists, or not. The computing device 10 comprises a trend diagnostics 14, a pattern 16, a data set concerning the actual measured data 9, a nominal data set 1 1 and a normal state data set 12. The trend data diagnostics 14 determines performance trend changes based on the actual measured data 9. In other words the trend diagnostics 14 is capable to determine whether a trend anomaly 23 has occurred. A trend anomaly is an anomaly which is derived by comparing the time course of the actual measured data 9 with the time course of normal state data sets 12. The output data of the trend diagnostics 14 indicates exceptional behaviour, e.g. significant pressure change of an electric submersible pump, or an operating point of the pump which can lead to a pump failure. If a trend anomaly 23 is detected by the trend diagnostics 14 a warning message 13 related to the trend anomaly 23 is communicated to the operating personnel. The operating personnel, which operates a SCAD A system 17 (Supervisory Control And Data Acquisition) can decide, whether any action would help prevent or reduce the chance of a shutdown in the near future, thus maintaining production or e.g. ensuring proper operation of the ESP. Both the output data and the warning message 13 which is related to the trend anomaly 23 are also communicated to the SCAD A system. The SCADA system 17 is capable of providing monitoring data to the operating personnel. In other words by means of the SCADA system 17, the operating personell can supervise the pumping device. The SCADA system 17 also comprises an interface by means of which the operating personell can communicate with the computing device 10.
The normal state data set 12 comprises already known and normal conditions of the pumping device. The nominal data set 11 comprises nominal values of the actual measured data 9. By means of the nominal data set 1 1 a threshold is determined by the computing device 10, which defines a maximum and a minimum value of the normal state data set 12. The pattern 16 is identified by a diagnostic analysis 26, which brings together the actual measured data 9, a
proprietary data model and expert knowledge to contribute to the diagnosis of problems. The diagnostic analysis 26 applies "fuzzy logic" to identify patterns 16 in the actual measured data 9. The operating personnel is notified of potential problems along with a measure of how closely the actual measured data 9 matches the pattern 16 related to a problem respectively an anomaly. Furthermore, the pattern 16 comprises a segment of time series by means of which a certain deviation of the normal state data set 12 is determined. If the modulus of the deviation outnumbers a time-series-threshold, which is a limit value by which the time-series of the actual measured data 9 may differ from the time-series of the normal state data set 12, a time series anomaly 24 is detected, whereat both the warning message 13 and a data set, which is related to the time series anomaly 24 is communicated to the SCADA system 17.
FIG 3 shows the contents of FIG 2 in substantial parts and therefore only the differences are described. In FIG 3 the computing device 10 is extended by an artificial intelligence 15. Based on a comparison between the normal state data set 12 and the actual measured data 9 the artificial intelligence 15 determines a difference 18. The difference 18 is independent of both the trends diagnostics 14 and the pattern 16, which was derived from diagnostics analysis 26. Hence by means of the difference 18 the artificial intelligence 15 is capable to determine a difference anomaly 25, whereat the difference anomaly 25 comprises anomalies, which are unknown to both the trend diagnostics 14 and the diagnostics analysis 26. In contrast to FIG 2, FIG 3 illustrates a situation, where both the trend diagnostic 14 and the diagnostic analysis 26 determine a normal behaviour 6 whereas the artificial intelligence 15 determines a difference anomaly 25 and communicates both the data set concerning the different anomaly 25 and the warning message 13 to the operating personnel which operates the SCADA system 17. Hence the artificial intelligence 15 increases the reliability of the monitoring, since the artificial intelligence 15 is capable of detecting new kinds of anomalies, which are unknown to both the trend diagnostic 14 and the diagnostics analysis 26.
FIG 4 shows the contents of FIG 2 respectively FIG 3 in substantial parts and therefore only the difference is described.
FIG 4 describes two different states, wherein on the one hand the artificial intelligence 15 has successfully detected an anomaly and on the other hand the anomaly which was detected by
the artificial intelligence 15 is a false alarm 21, which is communicated to the SCADA system 17. In case of a correct anomaly detection by the artificial intelligence 15 the operating personnel initiates a data set extension 22 by means of the SCADA system, whereat both the trend diagnostic 14 and the diagnostic analysis 26, is extended by a data set, which is related to a previously unknown anomaly, that was detected by the artificial intelligence 15. Otherwise if the anomaly, which was detected by the artificial intelligence 15 is declared as a false alarm 21 by the operating personnel, a retraining 19 of the artificial intelligence 15 is initiated by the operating personnel to prevent the artificial intelligence 15 from causing the same or a similar false 21 alarm again in the future. Additionally the diagnostics analysis 26, which comprises the pattern 16, initiates a diagnostic retraining 20, by means of which it is ensured, that the artificial intelligence 15 is provided with a data set concerning the pattern 16 of the false anomaly detection which caused the false alarm 21. In other words the artificial intelligence 15 can be both retrained by the operating personnel and by the diagnostics analysis26.
Claims
1. A method for monitoring a pumping device for a fluid medium by means of a computing device (10), comprising the following steps:
- capturing actual measured data sets (9) of the pumping device characterizing an actual state;
- comparing the respective actual measured data sets (9) to respective, stored nominal data sets (11) and, if applicable, capturing a deviation of at least one actual measured data set (9) from a nominal dataset (11); and
- in case a deviation has been detected: associating the deviation with at least one standard type of damage;
characterized by the step:
determining a difference (18) of at least one of the actual measured data sets (9) from the normal state data sets (12) captured during the operation of the pumping device.
2. The method according to claim 1,
characterized in that
the established difference (18) is used for the detection of types of damage differing from the standard types of damage.
3. The method according to any one of claims 1 or 2,
characterized in that
in case the difference (18) of the data of the nominal data set (1 1) differs by a pre-determined threshold value, a warning message (13) is issued.
4. The method according to claim 3,
characterized in that
in case the warning message (13) is a false alarm (21), the actual measured data set (9) is assigned to the normal state data sets (12) by means of an interface (17).
5. The method according to claim 3,
characterized in that
in case the warning message (13) concerns at least one of the types of damage, the data characterizing the at least one type of damage by means of the interface (17) are assigned to a data set characterizing the standard types of damage.
6. The method according to any one of the preceding claims,
characterized in that
for determining the difference (18) the data sets characterizing the types of damage are used, as well as the normal state data sets (12).
7. The method according to any one of the preceding claims,
characterized in that
the normal state data sets (12) and the difference (18) are determined by an artificial intelligence (15).
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/IB2013/001277 WO2014177902A1 (en) | 2013-04-30 | 2013-04-30 | Method for monitoring a pumping device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/IB2013/001277 WO2014177902A1 (en) | 2013-04-30 | 2013-04-30 | Method for monitoring a pumping device |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2014177902A1 true WO2014177902A1 (en) | 2014-11-06 |
Family
ID=48948464
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/IB2013/001277 WO2014177902A1 (en) | 2013-04-30 | 2013-04-30 | Method for monitoring a pumping device |
Country Status (1)
Country | Link |
---|---|
WO (1) | WO2014177902A1 (en) |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1845553A1 (en) * | 2004-12-28 | 2007-10-17 | Tokyo Electron Ltd. | Semiconductor manufacturing apparatus, abnormality detection in such semiconductor manufacturing apparatus, method for specifying abnormality cause or predicting abnormality, and recording medium wherein computer program for executing such method is recorded |
US20130103353A1 (en) * | 2010-07-09 | 2013-04-25 | Gerd Klöppner | Monitoring and diagnostic system for a fluid energy machine system and fluid energy machine system |
-
2013
- 2013-04-30 WO PCT/IB2013/001277 patent/WO2014177902A1/en active Application Filing
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1845553A1 (en) * | 2004-12-28 | 2007-10-17 | Tokyo Electron Ltd. | Semiconductor manufacturing apparatus, abnormality detection in such semiconductor manufacturing apparatus, method for specifying abnormality cause or predicting abnormality, and recording medium wherein computer program for executing such method is recorded |
US20130103353A1 (en) * | 2010-07-09 | 2013-04-25 | Gerd Klöppner | Monitoring and diagnostic system for a fluid energy machine system and fluid energy machine system |
Non-Patent Citations (1)
Title |
---|
ANONYMOUS: "Real versus nominal value - Wikipedia, the free encyclopedia", 21 February 2013 (2013-02-21), XP055088247, Retrieved from the Internet <URL:http://en.wikipedia.org/wiki/Real_versus_nominal_value> [retrieved on 20131114] * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP3055747B1 (en) | Correlation and annotation of time series data sequences to extracted or existing discrete data | |
EP2905665B1 (en) | Information processing apparatus, diagnosis method, and program | |
US11078774B2 (en) | System and method for detecting, diagnosing, and correcting trips or failures of electrical submersible pumps | |
CA2695450C (en) | Fuzzy classification approach to fault pattern matching | |
KR101065767B1 (en) | Online early fault detection and diagnostic method for plant operation | |
US20150346066A1 (en) | Asset Condition Monitoring | |
WO2014207789A1 (en) | Status monitoring device | |
EP4062030B1 (en) | Well annulus pressure monitoring | |
EP3055746B1 (en) | Correlation and annotation of time series data sequences to extracted or existing discrete data | |
US11613985B2 (en) | Well alarms and event detection | |
KR102067344B1 (en) | Apparatus and Method for Detecting Abnormal Vibration Data | |
KR20210017651A (en) | Method for Fault Detection and Fault Diagnosis in Semiconductor Manufacturing Process | |
WO2017159016A1 (en) | Abnormality diagnosis system | |
KR102470112B1 (en) | Intelligent condition monitoring method and system for nuclear power plants | |
CN105466707A (en) | Method, system and computer program for fault detection in a machine | |
WO2015101856A2 (en) | A method and system for fault detection during a plunger lift operation in hydrocarbon wells | |
WO2024124672A1 (en) | Fault diagnosis method for tunnel boring device and intelligent tunnel boring device | |
WO2014177902A1 (en) | Method for monitoring a pumping device | |
CN113778044A (en) | Monitoring method and device for blower system of thermal power plant | |
Raza et al. | An integrated qualitative trend analysis approach to identify process abnormalities: a case of oil export pumps in an offshore oil and gas production facility | |
CN117520964B (en) | Motor pre-diagnosis method based on artificial intelligence | |
KR102547102B1 (en) | Digital twin-based predictive diagnostic method for inverter type pump systems | |
TWI709010B (en) | Device abnormality reasons diagnosis method and system thereof | |
JP7447758B2 (en) | Plant abnormality prediction system | |
US20220004176A1 (en) | Method and control system for detecting condition of plurality of process equipment in industrial plant |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 13745901 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 13745901 Country of ref document: EP Kind code of ref document: A1 |