WO2020173581A1

WO2020173581A1 - System, device and method of monitoring condition of a technical installation

Info

Publication number: WO2020173581A1
Application number: PCT/EP2019/073335
Authority: WO
Inventors: Harshad PINGLE; Nilkanth DESHMUKH; Atul JAWALE
Original assignee: Siemens AG; Siemens Corp
Current assignee: Siemens AG; Siemens Corp
Priority date: 2019-02-27
Filing date: 2019-09-02
Publication date: 2020-09-03
Anticipated expiration: 2021-08-27

Abstract

A system, apparatus and method for monitoring condition of a technical installation (180, 400) is disclosed. The method comprising receiving condition data associated with the technical installation (180, 400) by a computing platform (120, 410, 460); analysing the condition data associated with each asset in the group of assets (180A-180B, 482-486) using pattern-based rules, wherein each of the pattern-based rules correspond to a predetermined pattern of anomalies likely to occur across the group of assets (180A-180B, 482-486) with respect to time; detecting a pattern of anomalies associated with the group of assets (180A-180B, 482-486) in the technical installation (180, 400) based on the analysed condition data; and generating a notification to indicate presence of the pattern of anomalies in the group of assets (180A-180B, 482-486).

Description

SYSTEM, DEVICE AND METHOD OF MONITORING CONDITION OF A TECHNICAL INSTALLATION

The present invention relates to monitoring condition of a technical installation.

Technical installations may include multiple assets, opera tions and processes to produce a resultant product. Improving productivity for individual assets may be difficult if it suffers unplanned downtime. Achieving productivity improve ment may be increasingly difficult in a group of heterogene ous assets.

The assets are susceptible to minor or significant anomalies in the operating parameters. Such anomalies may indicate a failure in the asset or the technical installation. General ly, the reasons for the failure may be known mainly from do main expertise.

The known techniques to monitor the condition of the tech nical installation involve identification of anomalies in a single asset. The techniques do not solve the problem of anomalies in multiple heterogenous assets in the technical installation. Therefore, the condition of the technical in stallation may not be accurately monitored.

Therefore, it is the object of the present invention to moni tor the condition of a technical installation and the assets. The object can be achieved by detecting one or more anomalies in a group of assets in the technical installation. The anom alies are detected based on pattern of anomalies across the group of assets that occur within a predefined time.

The object of the present invention is achieved by a method for monitoring condition of a technical installation includ ing multiple heterogenous assets defined as one or more groups of assets. As used herein "group of assets" refer to assets that are operationally dependent on each other. For example, the assets may be a power supply, a motor and a pump. The pump output is operationally dependent on the motor and the power supply. Such operational dependencies can be seen between assets in the technical installation. In an em bodiment the same asset may belong to more than one group of assets. For example, the power supply for the motor and pump can be also be input to a voltage stabilizer.

The method includes receiving condition data associated with the technical installation on a computing platform. The condition data is a measure of operating parameters associated with the assets and the technical installation. The condition data can be represented as data points. For example, the sensing and monitoring devices may include thermal imaging devices, vibration sensors, current and voltage sensors, etc. The term "operation parameter" refers to one or more characteristics of the assets and the technical installation. The operation parameters are used to define performance of the assets and the technical.

The method includes analysing the condition data based on the group of assets. Accordingly, the condition data associated with each asset in the group of assets is analysed using pat tern-based rules. As used herein "the pattern-based rules" refer to defined conditional-dependencies associated with the technical installation. The pattern-based rules govern the operation of the technical installation when applied. In an embodiment, the pattern-based rules correspond to a predeter mined pattern of anomalies likely to occur across the group of assets with respect to time. The predetermined pattern of anomalies may be autonomously learnt from prestored condition data of the technical installation based on known pattern recognition algorithms.

The method may include generating at least one pattern-based rule associated with the condition of the technical installa tion in real-time. The pattern-based rule corresponds to a pattern of anomalies which occur across the group of assets in real-time. In an embodiment, the pattern of anomalies is predicted in real-time based on the predetermined pattern of anomalies .

The method may include generating the at least one pattern- based rule by detecting one or more anomalies at each asset of the group of assets based on the condition data associated with group of assets. The anomalies are detected by comparing the condition data and the prestored condition data associat ed with the group of assets. Further, the anomalies are de tected when the anomalies occur within the predefined time period. For example, reduction in the pump output occurs at time tl. If fluctuation in power is detected within tl-x (where x is the predefined time period) then the fluctuation in power and the reduction in pump output are detected as anomalies .

The method may include analysing the anomalies detected across group of assets. In an embodiment, the anomalies are analysed by identifying a first anomaly associated with a first asset the group of assets by comparing the condition data and the prestored condition data associated with first asset. Further, the method includes identifying a subsequent anomaly associated with a second asset in the group of assets based on the condition data and the prestored condition data associated with second asset. The subsequent anomaly is iden tified based on the occurrence of the subsequent anomaly within the predefined time period from occurrence of the first anomaly.

In an embodiment, the first anomaly is identified while the subsequent anomaly may not be identifiable. The method may then include determining a subsequent anomaly sequence across the group of assets within the predefined time period based on the first anomaly and an asset-relationship between the group of assets. As used herein, "asset-relationship" is a representation of operational dependencies between each asset of the group of assets. The asset-relationship may be learnt autonomously from technical documentation associated with the technical installation.

In another embodiment, the first anomaly may not be identifi able. Therefore, the method may then include determining the subsequent anomaly sequence across the group of assets within the predefined time period based on the asset-relationship. The asset-relationship advantageously enables identification of the pattern of anomalies across the group of assets. Therefore, the present invention addresses the object of de tecting the pattern of anomalies in heterogenous assets in the technical installation.

The method includes detecting the pattern of anomalies asso ciated with the group of assets in the technical installation based on the analysed condition data. In an embodiment, the method may include identifying the pattern of anomalies de tected across the group of assets within the predefined time period based on the analysis of the anomalies in real-time. Further, the method may include generating the at least one pattern-based rule in real-time based on the identified pat tern of anomalies.

The method may include predicting a failure in the technical installation based on the detected the pattern of anomalies. Further, the method may include validating the predicted failure in the technical installation through one of inspec tion and simulation of the predicted failure. For example, the simulation of the predicted failure may be performed on a dynamic model of the technical installation. As used herein dynamic model refers to a real-time representation of the technical installation and the assets. The dynamic model rep resents evolving condition of the technical installation that describe the performance of the group of assets and the tech nical installation.

The method may include determining accuracy of the predicted failure based on the validation and modifying the pattern- based rules based on the accuracy of the predicted failure. Further, the method may include predicting a down-time for the technical installation based on the predicted failure in the technical installation.

The method includes generating a notification to indicate presence of the pattern of anomalies in the group of assets. The method may include generating the notification indicating the possibility of the predicted failure of the technical in stallation .

During the implementation of this method, the pattern of anomalies are identified across multiple assets. Further, the method is not limited to identifying one anomaly per asset. Accordingly, the method advantageously identifies a pattern of anomalies when the number of anomalies exceeds the number of assets. For example, during implementation, three anoma lies can be identified from a group of two assets. Conse quently, the pattern of anomalies associated with the three anomalies and the two assets is identified.

The object of the present invention is also achieved by an apparatus for a technical installation including at least one group of assets. The apparatus comprises one or more pro cessing units and a memory unit communicative coupled to the one or more processing units. The memory unit comprises an condition monitoring module stored in the form of machine- readable instructions executable by the one or more pro cessing units. The condition monitoring module is configured to perform method steps described above. The execution of the asset module can also be performed using co-processors such as Graphical Processing Unit (GPU) , Field Programmable Gate Array (FPGA) or Neural Processing/Compute Engines.

According to an embodiment of the present invention, the ap paratus can be an edge computing device. As used herein, "edge computing" refers to computing environment that is ca pable of being performed on an edge device (e.g., connected to the sensors unit in an industrial setup on one end and to a remote server (s) such as for computing server (s) or cloud computing server (s) on the other end), which may be a compact computing device. A network of the edge computing devices can also be used to implement the apparatus . Such a network of edge computing devices is referred to as a fog network.

Additionally, the object of the present invention is achieved by a system comprising a cloud computing platform that in cludes a condition monitoring module configured to perform one or more method steps described above.

The object of the present invention is achieved by a comput er-program product having machine-readable instructions stored therein, which when executed by a processor unit, cause the processor unit to perform a method as described above .

The above-mentioned and other features of the invention will now be addressed with reference to the accompanying drawings of the present invention. The illustrated embodiments are intended to illustrate, but not limit the invention.

The present invention is further described hereinafter with reference to illustrated embodiments shown in the accompanying drawings, in which:

FIG 1 illustrates a block diagram of a system to monitor condition of a technical installation, according to an embodiment of the present invention;

FIG 2 is a schematic representation illustrating opera tion of the system in FIG 1, according to an embod iment of the present invention;

FIG 3A-3C illustrates operation of the system in FIG 1 in re- lation to different scenarios in the technical in stallation; FIG 4 illustrates a block diagram of an apparatus to mon itor condition of a technical installation, accord ing to an embodiment of the present invention; and

FIG 5 is a flowchart of a method of monitoring the condi tion of a technical installation.

Hereinafter, embodiments for carrying out the present inven tion are described in detail. The various embodiments are de scribed with reference to the drawings, wherein like refer ence numerals are used to refer to like elements throughout. In the following description, for purpose of explanation, nu merous specific details are set forth in order to provide a thorough understanding of one or more embodiments. It may be evident that such embodiments may be practiced without these specific details.

FIG 1 illustrates a block diagram of a system 100 for condi tion monitoring of a technical installation 180, according to an embodiment of the present invention. The technical instal lation 180 includes multiple assets 182-188. The technical installation 180 includes sensing and monitoring devices (not shown in FIG 1) capable of generating condition data of the technical installation 180. The condition data is communicat ed to a cloud computing platform via a network interface 150.

The condition data is a measure of operating parameters associated with the assets 182-188 and the technical installation 180. The condition data can be represented as data points. For example, the sensing and monitoring devices may include thermal imaging devices, vibration sensors, current and voltage sensors, etc. The term "operation parame ter" refers to one or more characteristics of the assets 182- 188 and the technical installation 180. The operation parame ters are used to define performance of the assets 182-188 and the technical 180. The system 100 includes the cloud computing platform 120 with a communication unit 122, a processing unit 124, a memory 130 and a database 160. The database 160 is configured to store dynamic model 162 of the technical installation 180. As used herein "dynamic model" refers to a real-time representation of the technical installation 180 and the assets 182-188. The dynamic model represents evolving condition of the technical installation that describe the performance of the assets 182- 188 and the technical installation 180.

The memory 130 includes a condition monitoring module 135. The condition monitoring module 135 includes pattern module 132, a rule generation module 134, a rule processing module 136 and a notification module 138.

The pattern module 132 includes an asset-relationship genera tor 142 and an anomaly detection module 144. The asset- relationship generator 142 is configured to generate an as set-relationship between the assets 182-188. In an embodi ment, the asset-relationship generator 142 is configured to generate the asset-relationship based on the dynamic model 162. The asset-relationship defines conditional-dependencies of the operating parameters associated with the assets 182- 188. In an embodiment, the asset-relationship is used to identify asset groups such as group of assets 180A and 180B.

For example, assets 182-186 be associated to a process and are referred to as the group of assets 180A. Similarly, the assets 187, 188 may be related to another process and can be long to the group of assets 180B. Accordingly, the asset- relationship between assets 182-186 is different from the as set-relationship with the assets 187, 188.

The anomaly detection module 144 detects anomalies between the condition data and prestored condition data associated the group of assets 180A and 180B. The prestored condition data may be determined based on a threshold. In an embodi ment, the prestored condition data is predicted for each time instant based on the dynamic model 162 of the technical in stallation 180.

Further, the anomaly detection module 144 analyses the anoma lies detected across group of assets 180A-180B to identify a pattern of anomalies that occur across each group within a predefined time period. For example, in the group of assets 180A, asset 182 is a power supply, asset 184 is a motor and asset 186 is a pump. The anomaly in the pump 186 output oc curs at time tl and fluctuation in power occurs tl-x (where x is predefined time period) . The anomaly in the pump 186 out put can be mapped to fluctuation in the power supply 182. This mapping is enabled by the asset-relationship between the pump 186 and the power supply 182.

Furthermore, the anomaly detection module 144 identifies the pattern of anomalies irrespective of sequence of occurrence of the anomalies. The anomaly detection module is configured to determine the anomalies that occur within the predefined time period. The operation of the anomaly detection module 144 is elaborated in Figures 3A-3C.

The rule generation module 134 is configured to generate the pattern-based rules based on the pattern of anomalies. The pattern-based rules when applied define conditional- dependencies associated with the technical installation 100 that govern the operation of the technical installation 100. In an embodiment, the pattern-based rules correspond to a predetermined pattern of anomalies likely to occur across the group of assets 180A, 180B within the predefined time period. The rule generation module 134 autonomously learns the pat tern-based rules from predetermined pattern of anomalies based on known rule generation algorithms such as associative rule generation algorithm.

The rule processing module 136 includes failure prediction module 146 and an accuracy module 148. The failure prediction module 146 is configured to predict the failure in the tech- incal installation 180 based on the pattern of anomalies and/or the pattern-based rules. The failure prediction module 146 monitors the condition of the assets 182-188 in real-time based on the pattern-based rules from the rule generation module 134. When the pattern-based rules are satisfied, the group of assets 180A associated with the pattern is deter mined. Further, one or more anomalies are detected in the group of assets 180A. For example, fluctuation in the power supply 182 is determined as the anomaly.

The failure prediction module 146 is further configured to estimate an asset remaining life of the group of assets based on the predicted failure. Also, predict a down-time for the technical installation based on the asset remaining life.

The accuracy module 148 determines accuracy of the predicted failure by simulating the predicted failure on the dynamic model 162 of the technical installation 180. In another em bodiment, the accuracy is determined through on-premise in spection. The pattern-based rules are modified based on the accuracy .

The notification module 138 is configured to generate a noti fication associated with the one or more anomalies and the failure predicted. The system 100 is communicatively coupled to a user device 110. For example, the cloud computing plat form 120 is communicatively coupled to the user device 110 via a communication unit 112 and the network interface 150. The user device 110 includes a processor 114, a memory 116 and a display 118. The user device 110 receives the notifica tion from the notification module and generates alerts on the display 118. The user device 110 is further configured to re ceive and display the asset remaining life and the predicted down-time of the technical installation 180. In an embodi ment, the user device 110 is configured to predict the fail ure and determine the asset remaining life and the down-time of the technical installation 180 upon receipt of the pattern of anomalies. FIG 2 is a schematic representation illustrating operation of the system 100 in FIG 1, according to an embodiment of the present invention. The operation of the system 100 is per formed in 4 main steps 210-240.

At step 210, the condition data 202 is received in time se ries by the pattern module 132. Further, at step 210 the pat tern module 132 is configured to identify the pattern of anomalies for the condition data 202. Further, the pattern of anomalies is updated based on feedback 250 from the technical installation 180.

At step 220, the rule generation module 134 is configured to generate the pattern-based rules based on the pattern of anomalies. In an embodiment, at step 220 pre-existing rules are updated based on the pattern of anomalies. For example, the pre-existing rules are defined by a technical expert for the technical installation. In another example, the pre existing rules are defined based on pattern of anomalies identified from similar technical installations performing similar operations as that of technical installation 180.

At step 230, the rule processing module 136 detects one or more anomalies in the condition data 202 associated with the group of assets 180A. Further, at step 230 the failure in the technical installation 180 is predicted, specifically the failure in the assets 182-186 in the group of assets 180A is predicted. Additionally, at step 230 the down-time of the technical installation 180 is predicted based on the predict ed failure.

At step 240, the notification module 138 generates a notifi cation to a user device to alert a user regarding the pre dicted failure and estimated down-time.

FIG 3A-3C illustrates operation of the system 100 in FIG 1 in relation to different scenarios in a technical installation. The technical installation includes a group of assets (asset 310, asset 320 and asset 330) . The different scenarios in FIG 3A, 3B and 3C represent different time instants at which anomalies are detected and analysed. Further, the different scenarios also represent situations when a first anomaly and a subsequent anomaly are known or unknown.

FIG 3A represents a scenario in which the first anomaly in asset 310 is known and the subsequent anomaly in asset 320 and asset 330 is known. Further, the first anomaly and the subsequent anomalies in assets 320 and 330 are known to occur within predefined time periods of 302 and 304, respectively.

As shown in FIG 3A, anomaly in asset 310 is followed by anom aly in asset 320 within predefined time period 302 from the occurrence of anomaly in asset 310. In addition, the anomaly in asset 320 is followed by anomaly in asset 330 within pre defined time period 304. The pattern module 132 in system 100 detects anomalies based on the above scenario. The rule gen eration module 134 generates pattern-based rules associated with the pattern of anomalies. For example, the rule generat ed will be:- if anomaly in asset 310 is followed by anomaly in asset 320 and asset 330 within time periods of 302 and 304 between each of the anomalies, then an anomaly is detected.

FIG 3B represents a scenario in which the first anomaly in asset 310 is known and the subsequent anomaly is not known. Therefore, FIG 3B represents alternative situations 305 and 315. In situation 305, the anomaly detection in asset 320 is followed by anomaly detection in asset 330. In situation 315, the anomaly detection in asset 330 is followed by anomaly de tection in asset 320.

For the scenario in FIG 3B, the pattern module 132 determines total time between the anomalies in asset 320 and asset 330. If the total time is within a predefined time period 306, then the anomaly is detected. FIG 3C represents a scenario in which the first anomaly in asset 310 and subsequent anomaly in assets 320, 330 are not known. Further, FIG 3C represents a possibility that anomaly in asset 330 may be detected prior to detection of anomaly in asset 310. Accordingly, the pattern module 132 identifies whether there is a likelihood occurrence of the anomalies in the assets 310 and 330 within a predefined time period 308. The likelihood occurrence of the anomalies is used to define the pattern of anomalies and thereby the pattern-based rules.

To detect the scenarios in FIG 3A-3C, the pattern module 132 is configured to identify the first anomaly associated with the group of assets. Further, the subsequent anomaly is iden tified within the predefined time period from occurrence of the first anomaly. To identify subsequent anomaly, the pat tern module 132 is configured to determine a subsequent anom aly sequence within the predefined time period based on the first anomaly and the anomaly-relationship. In FIG 3C, the anomaly-relationship defines the likelihood of anomaly within the time period 308. The asset-relationship advantageously enables identification of the pattern of anomalies across the group of assets.

FIG 4 illustrates a block diagram of an apparatus 410 to mon itor condition of a technical installation 400, according to an embodiment of the present invention. The technical instal lation 400 includes groups of assets 482-486, each including assets. For example, the group of assets 482 includes a power supply 402 communicatively coupled to a motor 404 and a pump 406. Each of the assets are associated with sensing and moni toring devices capable of measuring operating parameters of the assets.

The apparatus 410 is an edge device and includes an operating system 412, a memory 414 and application runtime 416. The edge device 410 also includes a graphical user interface 418. The operating system 412 is an embedded real-time operating system (OS) such as the Linux™ operating system. The edge operating system 412 enables communication with the sensing and monitoring devices, the assets and with an IoT cloud platform 460. The edge operating system 412 also allows run ning one or more software applications rule processing module 422 and notification module 424 deployed in the edge device 410. The application runtime 416 is a layer on which the one or more software applications 422, 424 are installed and exe cuted in real-time. The edge device 410 communicates with the cloud platform 460 via a network interface 450. The cloud platform 460 is configured to execute modules 462 and 464.

During operation, the edge device 410 receives the condition data of the assets in the technical installation 400. The condition data is transmitted to the cloud platform 460. The cloud platform 460 includes a pattern module 462 configured to identify a pattern of anomalies for the condition data that is received.

For example, the condition data indicates reduction in output of the pump 406. The pattern module 460 determines the rela tionship of the pump 406 to the assets in the technical in stallation 400 using an asset-relationship. The asset- relationship indicates association of the pump 406 to the mo tor 404 and the power supply 402. The pattern module 460 is configured to identify the pattern of anomalies. The pattern of anomalies is identified by determining whether there the anomalies in the operation of the power supply 402 and the motor 404 occur within a predefined time period.

The cloud platform 460 also includes a rule generator module 464 to generate pattern-based rules associated with a prede termined pattern of anomalies. Considering the previous exam ple of irregular pump output is prestored as a pattern of anomalies. The irregular pump output is mapped to a shaft misalignment in the motor 404 within a predefined time period tl, the pattern-based rule is accordingly defined. If the ir regular pump output is mapped to irregular power supply with- in a predefined time period t2, the pattern-based rule is ac cordingly defined.

The pattern-based rules are received by the edge device 410 and is processed by the rule processing module 422. The rule processing module 422 identifies whether the condition data from the technical installation 400 satisfies the pattern- based rules and accordingly detects an anomaly and predicts a failure in the technical installation 400 based on the anoma ly. When the anomaly is detected, the notification module 424 generates a notification indicating the anomaly and the pre dicted failure.

The detected anomaly and the predicted failure is rendered on the graphical user interface 418 as an alert. In an embodi ment, the anomaly and the predicted failure are rendered on a digital representation of the technical installation 400 dis played on the graphical user interface 418.

FIG 5 is a flowchart of a method of monitoring the condition of a technical installation. The method begins at step 502 with receipt of condition data associated with assests in the technical installation. The condition data is received in time series.

At step 504, the condition data associated with each asset in the group of assets is analysed using pattern-based rules. The pattern-based rules correspond to a predetermined pattern of anomalies likely to occur across the group of assets with respect to time. In an embodiment, at least one pattern-based rule associated with the condition of the technical installa tion in real-time. The at least one pattern-based rule corre sponds to a pattern of anomalies which occur across the group of assets.

At step 506 a pattern of anomalies is detected, associated with the group of assets in the technical installation based on the analysed condition data. To detect the pattern of anomalies, anomaly between the condition data and a prestored condition data is determined. Further, group of assets that are associated with the detected anomaly are determined based on an asset-relationship. The anomaly within a predefined time period is used to detect the pattern of anomalies. The pattern of anomalies includes the conditional-dependencies in the anomalies and the group of assets with respect to the predefined time period between anomalies.

At step 508 pattern-based rules are generated based on the pattern of anomalies. The pattern-based rules define condi tion-dependencies in the anomalies with respect to time of occurance of the anomaly and the group of assets associated with the anomaly.

At step 510 one or more anomalies are detected in the group of assets associated based on the pattern-based rules. When the pattern-based rules are satisfied, one or more anomalies are detected in the group of assets. At step 512 a failure is predicted based on the detected one or more anomalies. The accuracy of the predicted failure is determined by validating the predicted failure. The validation may be done via inspec tion of the technical installation or by simulation of the predicted failure. The pattern-based rules may be modified based on the accuracy.

At step 514 an asset remaining life of the group of assets is estimated based on the predicted failure. At step 516 a down time is predicted for the technical installation based on ei ther on the asset remaining life or the historical condition data of the technical installation. The steps 514 and 516 can be done in parallel. In an embodiment, step 516 may be per formed independent of step 514 by predicting the down-time of the technical installation based on the historical condition data .

At step 518 a notification is generated associated with the one or more anomalies and the predicted failure. At step 520 the predicted failure, the asset remaining life and the pre dicted down-time are rendered on a display device.

The present invention can take a form of a computer program product comprising program modules accessible from computer- usable or computer-readable medium storing program code for use by or in connection with one or more computers, proces sors, or instruction execution system. For the purpose of this description, a computer-usable or computer-readable me dium can be any apparatus that can contain, store, communi cate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. The medium can be electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system (or appa ratus or device) or a propagation mediums in and of them selves as signal carriers are not included in the definition of physical computer-readable medium include a semiconductor or solid state memory, magnetic tape, a removable computer diskette, random access memory (RAM) , a read only memory (ROM) , a rigid magnetic disk and optical disk such as compact disk read-only memory (CD-ROM) , compact disk read/write, and DVD. Both processors and program code for implementing each aspect of the technology can be centralized or distributed (or a combination thereof) as known to those skilled in the art .

While the present invention has been described in detail with reference to certain embodiments, it should be appreciated that the present invention is not limited to those embodi ments. In view of the present disclosure, many modifications and variations would be present themselves, to those skilled in the art without departing from the scope of the various embodiments of the present invention, as described herein. The scope of the present invention is, therefore, indicated by the following claims rather than by the foregoing descrip tion .

Claims

1. A method for monitoring condition of a technical installa tion (180, 400) including at least one group of assets (180A- 180B, 482-486) , the method comprising:

receiving condition data associated with the technical in stallation (180, 400) by a computing platform (120, 410, 460) ;

analysing the condition data associated with each asset in the group of assets (180A-180B, 482-486) using pattern-based rules, wherein each of the pattern-based rules correspond to a predetermined pattern of anomalies likely to occur across the group of assets (180A-180B, 482-486) with respect to time ;

detecting a pattern of anomalies associated with the group of assets (180A-180B, 482-486) in the technical installation (180, 400) based on the analysed condition data; and

generating a notification to indicate presence of the pat tern of anomalies in the group of assets (180A-180B, 482- 486) .

2. The method according to claim 1, further comprising:

predicting a failure in the technical installation (180,

400) based on the detected the pattern of anomalies; and

generating the notification indicating the possibility of the predicted failure of the technical installation (180, 400) .

3. The method according to one of claim 1 and 2, further com prising :

generating at least one pattern-based rule associated with the condition of the technical installation (180, 400) in re al-time, wherein the at least one pattern-based rule corre sponds to a pattern of anomalies which occur across the group of assets (180A-180B, 482-486) .

4. The method according to claim 3, further comprising:

storing the pattern-based rules on a rules database.

5. The method according to claim 3, wherein generating the at least one pattern-based rule comprises:

detecting one or more anomalies at each asset of the group of assets (180A-180B, 482-486) based on the condition data associated with group of assets;

analysing the anomalies detected across group of assets (180A-180B, 482-486);

identifying the pattern of anomalies detected across the group of assets (180A-180B, 482-486) within a predefined time period based on the analysis of the anomalies;

generating the at least one pattern-based rule based on the identified pattern of anomalies.

6. The method according to claim 5, wherein detecting one or more anomalies at each asset of the group of assets (180A- 180B, 482-486) comprises:

detecting each of the one or more anomalies, by comparing the condition data and a prestored condition data associated with the group of assets (180A-180B, 482-486) , wherein the one or more anomalies occur within the predefined time peri od .

7. The method according to claim 5, wherein analysing the one or more anomalies comprises:

identifying a first anomaly associated with a first asset the group of assets (180A-180B, 482-486) by comparing the condition data and the prestored condition data associated with first asset; and

identifying a subsequent anomaly associated with a second asset in the group of assets (180A-180B, 482-486) based on the condition data and the prestored condition data associat ed with second asset; wherein the subsequent anomaly occurs within the predefined time period from occurrence of the first anomaly.

8. The method according to one of claim 5, further compris ing : determining a subsequent anomaly sequence across the group of assets (180A-180B, 482-486) within the predefined time pe riod based on the first anomaly and an asset-relationship be tween the group of assets (180A-180B, 482-486) , wherein the asset-relationship is a representation of operational depend encies between each asset of the group of assets (180A-180B, 482-486) .

9. The method according to one of claim 8, further compris ing :

determining the subsequent anomaly sequence across the group of assets (180A-180B, 482-486) within the predefined time period based on the asset-relationship.

10. The method according to claim 2, further comprising: validating the predicted failure in the technical instal lation (180, 400) through one of inspection and simulation of the predicted failure.

determining accuracy of the predicted failure based on the validation; and

modifying the pattern-based rules based on the accuracy of the predicted failure.

11. The method according to one of the preceding claims, further comprising:

predicting a down-time for the technical installation (180, 400) based on the predicted failure in the technical installation (180, 400) .

12. An apparatus (410) for monitoring condition of a tech nical installation (180, 400) including at least one group of assets (180A-180B, 482-486) , the apparatus comprising:

one or more processing units (412); and

a memory unit (416) communicative coupled to the one or more processing units, wherein the memory unit comprises a condition monitoring module (422, 424) stored in the form of machine-readable instructions executable by the one or more processing units, wherein the condition monitoring module (422, 424) is configured to perform one or more method steps according to claims 1 to 11.

13. A system (100) for of monitoring condition of a tech- nical installation (180, 400) including at least one group of assets (180A-180B, 482-486) , the system comprising:

a cloud computing platform (120) comprising:

a condition monitoring module (135) configured to per form one or more method steps according to claims 1 to 11.

14. A computer-program product, having machine-readable in structions stored therein, that when executed by a processor, cause the processor to perform method steps according to any of the claims 1-11.