WO2018224221A1

WO2018224221A1 - System, method and device for operation and maintenance of a wind farm

Info

Publication number: WO2018224221A1
Application number: PCT/EP2018/061514
Authority: WO
Inventors: Nagaraja Kuppagadde Sheshadri Rao
Original assignee: Siemens Wind Power A/S
Priority date: 2017-06-08
Filing date: 2018-05-04
Publication date: 2018-12-13

Abstract

System, method and device for operation and maintenance of a wind farm. The device is a virtual assistant device (100) for assisting operation and maintenance of a wind park comprising wind turbines, comprising: a communication unit (130) to receive a query associated with the operation and maintenance of the wind park; a memory device (102) configured to store data comprising historical and real¬ time operation and maintenance data of the wind park; a processor (110) communicatively coupled to the memory device (102); and an output unit comprising a display (140) and a speaker to output the at least one response. The processor (110) is configured to analyze the query (210) to generate a look up table of the query and responses based on the historical operation and maintenance data and to generate a response (220) to be displayed based on the responses and the real-time operation and maintenance data.

Description

Beschreibung / Description

System, Method and Device for operation and maintenance of a Wind Farm

Typically, a wind farm is commissioned and deployed in remote regions both onshore and offshore. Wind turbines in the wind farm generally have a lifetime of 20 years and need to be continuously serviced and maintained for optimum power pro- duction during their lifetime. To ensure efficient service and maintenance of the wind farm service engineers trained and specialized on servicing and operating wind turbines are required . The servicing of wind turbines is a complicated task in view of the interaction of several hardware parts and associated software. Therefore, such operations demand higher domain knowledge and a service engineer may not have all the re^¬ quired knowledge to perform support tasks without external help. Further, over a period of time upgrades to existing or new software and hardware are installed requiring service en^¬ gineer to constantly upgrade their know-how of the new sys^¬ tems . Currently multiple specialist service engineers are involved in servicing the wind farm. A service personnel communicates with domain specialists, who support the service remotely to accomplish the tasks. In certain situations, this approach maybe time consuming as the operations may require interac- tion between multiple domain specialists.

Therefore, it is an object of the present invention to pro^¬ vide a method, device and system of the aforementioned kind that are efficient in assisting operation and maintenance of a wind farm.

The method, device and system according to the present inven^¬ tion achieve this object by analyzing the at least one query to generate a look up table of the at least one query and one or more responses based on the historical and real-time oper^¬ ation and maintenance data. Further, by generating at least one response to be displayed based on the one or more re- sponses.

The present aspect of the invention relates to a device and method for assisting operation and maintenance of a wind park comprising a plurality of wind turbines. The device is a vir- tual assistant device which is capable of providing assis^¬ tance in the operation and maintenance of the wind farm via the Internet. The virtual assistant device includes a commu^¬ nication unit to receive one or more queries associated with the operation and maintenance of the wind park. The virtual assistant includes a memory device that is configured to store data comprising historical and real-time operation and maintenance data of the wind park. Further, the virtual as^¬ sistant device includes a processor communicatively coupled to the memory device. The processor is configured to analyze the queries to generate a look up table of the queries and one or more responses based on the historical operation and maintenance data. The processor also generates a response to be displayed based on the one or more responses and the real^¬ time operation and maintenance data. The virtual assistant device also includes an output unit having a display to dis^¬ play the response and a speaker to play the response.

As used in the description "the memory device" can be configured as cloud based database implemented in a cloud computing environment, where computing resources are delivered as a service over a network. As used herein, "cloud computing environment" refers to a processing environment comprising con^¬ figurable computing physical and logical resources, for exam^¬ ple, networks, servers, storage, applications, services, etc., and data distributed over the network, for example, the internet. The cloud computing environment provides on-demand network access to a shared pool of the configurable computing physical and logical resources. The communication network 130 is, for example, a wired network, a wireless network, a com^¬ munication network, or a network formed from any combination of these networks.

As used herein the term "operation and maintenance" includes any activity relating to operating the wind park including service operation data, monitoring data, wind turbine speci^¬ fication data, and/or cost data generated from inputs from client systems. Further, maintenance includes activities per^¬ formed to prevent failures, preventive maintenance e.g. using scheduled intervals, or it can be performed correcting a failure event, corrective maintenance. The term "historical and real-time operation and maintenance data" includes all information pertaining to the wind farm such as operating procedures of the wind park including restriction and limit- ing operations, standard operating procedures, high risk op^¬ erating procedure, and identity and access control infor^¬ mation of a servicing entity. Additionally, it includes in^¬ formation regarding meteorological conditions and parameters and performance parameters of the wind farm.

According to an embodiment, the queries and responses can be in the form of a text, an audio stream, a video stream and an image. The text can also include a phrase, a word or a code such as an alarm code or a service ticket. For example, the processor is configured to read the alarm code to generate the look up table of the alarm code and the probable respons^¬ es based on the associated operating procedure such as stand^¬ ard operating procedure or the risk operating procedure. For example, if the alarm code relates to ice detected on wind blade one or more responses, based on the standard operating procedure for ice detection on wind blade, are mapped to the alarm code. Similarly, in case of the service ticket the processor uses a service ticket identifier to map the responses to the service ticket. The processor then determines the response to be dis^¬ played based on the real-time operation and maintenance data. For example, a service ticket is raised by a service person- nel having access only to the wind turbine tower and not to the wind blade. Considering that the service ticket is with regard to wind blade de-icing, the processor generates the response indicating that the activity related to de-icing of wind turbine blade cannot be performed by the service person- nel.

According to a preferred embodiment, the analyzer of the pro^¬ cessor includes an audio recognition unit to analyze an audio stream to determine the query. To analyze an image, the pro- cessor includes an image recognition unit to determine the query. The processor further includes a natural language unit to analyze a text to determine the query. In case of a video stream, the video stream is divided into audio, visual and textual components and the audio recognition unit, the image recognition unit and the language processing unit analyzes the components to determine the query. For example, the lan^¬ guage processing unit analyzes the collection of texts from past service records using various language processing tech^¬ niques, such as stemming and/or natural language processing, to identify significant features in the texts.

According to another embodiment, the processor is configured to automatically generate the query based on comparison of the historical operation and maintenance data and the real- time operation and maintenance data. The query is analyzed and a look up table of the query and the one or more respons^¬ es is generated based on the real-time operation and mainte^¬ nance data. The processor then generates the response based on the one or more responses as an alert. For example, the communication unit receives input regarding change in meteorological parameters in real-time. This input is automatically read and compared with the historical operation and mainte^¬ nance data. Based on the comparison, the query is generated based on the real-time operation and maintenance data of the wind farm. Accordingly, the query is mapped to one or more responses, such as shutting down of one or more wind turbine or reducing the wind blade rotation speed. The processor then determines the most suitable response based on real-time pa- rameters . This is advantageous when coupled with a condition monitoring system associated with the wind farm. The condi^¬ tion monitoring system monitors the wind farm and wind turbines during the operation of the wind farm. The monitoring data is analyzed at by the virtual assistant device to iden- tify anomalies in the operation of the wind farm. These anom^¬ alies are used as the query and the response is automatically generated as the alert.

According to yet another embodiment, the processor is config- ured to initiate a communication session with a remote expert via the communication unit, to generate the response. This is advantageous when information regarding the wind farm is not fully available. In such cases, the processor initiates com^¬ munication with the remote expert to obtain the response to the query. The response from the remote expert is updated in the historical operation and maintenance data for reference in the future . Another aspect of the present invention relates to a system for virtually assisting operation and maintenance of a plu^¬ rality of wind parks. The system includes a database config^¬ ured to store data comprising historical and real-time opera^¬ tion and maintenance data of each of the plurality of wind parks. The system also includes a server communicatively cou^¬ pled to the database by means of a network interface. The server comprises one or more virtual assistant devices for assisting operation and maintenance of the plurality of wind parks .

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 in^¬ tended to illustrate, but not limit the invention.

The present invention is further described hereinafter with reference to illustrated embodiments shown in the accompany^¬ ing drawings, in which:

FIG 1 is a block diagram of a virtual assistant device according to the present invention;

FIG 2 is a graphical representation of a display of the virtual assistant device of FIG 1 ;

FIG 3 is a block diagram illustrating a system for assisting operation and maintenance of a wind farm according to the present invention;

FIG 4 is a process flowchart illustrating a method of virtually assisting operation and maintenance of a wind park according to the present invention; and FIG 5 is a process flowchart illustrating a method of virtually assisting operation and maintenance of a wind park, according to another embodiment of the present invention.

Various embodiments are described with reference to the draw^¬ ings, wherein like reference numerals are used to refer to like elements throughout. Further, numerous specific details are set forth in order to provide thorough understanding of one or more embodiments of the present invention. These exam^¬ ples must not be considered to limit the invention to config^¬ urations disclosed in the figures. It may be evident that such embodiments may be practiced without these specific de- tails.

FIG 1 is a block diagram of a virtual assistant device 100 according to the present invention. The virtual assistant de^¬ vice 100 is installed on and accessible by a user device, for example, a personal computing device, a workstation, a client device, a network enabled computing device, any other suita^¬ ble computing equipment, and combinations of multiple pieces of computing equipment. The virtual assistant device 100 in^¬ cludes a memory device 102 and a processor 110 communicative- ly coupled to the memory device 102.

The memory device 102 is, for example, a structured query language (SQL) data store or a not only SQL (NoSQL) data store. In an embodiment of the memory device 102 according to the present invention, the memory device 102 can also be a location on a file system directly accessible by the virtual assistant device 100. In a preferred embodiment according to the present invention, the virtual assistant device 100 is downloadable and usable on the user device. In another embodiment according to the present invention, the virtual assistant device 100 is con- figured as a web based platform, for example, a website host^¬ ed on a server or a network of servers. In another embodiment according to the present invention, the virtual assistant de^¬ vice 100 is implemented in the cloud computing environment. The virtual assistant device 100 is developed, for example, using Google App engine cloud infrastructure of Google Inc., Amazon Web Services® of Amazon Technologies, Inc. In an em^¬ bodiment, the virtual assistant device 100 is configured as a cloud computing based platform implemented as a service for analyzing data.

The memory device 102 is configured to store data comprising historical 106 and real-time 104 operation and maintenance data of the wind park. The processor 110 is configured to ex^¬ ecute the defined computer program instructions in units 112, 114, 116 etc. As illustrated in FIG 1, the virtual assistant device 100 comprises a communication unit 130 to receive one or more queries associated with the operation and maintenance of the wind park and a display 140. The processor 110 executes instructions and the execution of instructions is not limited to any specific combination of hardware circuitry and software instructions. To execute in^¬ structions the processor 110 disclosed herein can also in^¬ clude memory, input channels, and/or output channels. In the embodiments described herein, memory may include, without limitation, a computer-readable volatile medium, such as a random access memory (RAM), and/or a computer-readable nonvolatile medium, such as flash memory. Alternatively, a flop^¬ py disk, a compact disc-read only memory (CD-ROM) , a magneto- optical disk (MOD), and/or a digital versatile disc (DVD) may also be used. Also, in the embodiments described herein, in^¬ put channels may include, without limitation, sensors and/or computer peripherals associated with an operator interface, such as a mouse and a keyboard.

The processor 110 includes an analyzer 118 and a response generator 120. The analyzer 118 includes an audio recognition unit 112, an image recognition unit 114 and a language pro^¬ cessor unit 116. Therefore, the analyzer 118 is capable of analyzing the query independent of the format in which it is received by the communication unit 130. The analyzer 120 ana^¬ lyzes the queries to generate a look up table of the queries and one or more responses based on the historical operation and maintenance data. The processor also generates a response to be displayed based on the one or more responses and the real-time operation and maintenance data. For example, when the query is a service ticket or an alarm code the analyzer is configured to read the service ticket or the alarm code as the query and generate the look up table of the service tick^¬ et and the one or more responses based on a service ticket identifier/ the alarm code and the historical operation and maintenance data. In another embodiment, the analyzer 118 automatically gener^¬ ates the query based on the comparison of the historical op^¬ eration and maintenance data and the real-time operation and maintenance data. The query is then analyzed to generate a look up table of the query and one or more responses based on the real-time operation and maintenance data.

In yet another embodiment, when the analyzer 118 analyzes the query and is unable to generate the look-up table of one or more responses. The processor 110 initiates communication with the communication unit 130. This is used in instances where information regarding the query or the wind park is limited and the virtual assistant device is unable to respond satisfactorily to the query.

The response generator 120 determines the most suitable re^¬ sponse out of the one or more responses based on the real^¬ time operation and maintenance data. The display 140 displays the response to the user. In the present example, the user is a service personnel of the wind farm.

FIG 2 is a graphical representation of the display 140 on a user device. A user using the user device can access the vir^¬ tual assistant device 100 via a GUI (graphic user interface) 202. The GUI 202 is, for example, an online web interface, a web based downloadable application interface, etc. As shown in FIG 2 interaction between the user and the virtual assis^¬ tant device 100 is display by means of the GUI 202. The blocks indicate chat sessions between the virtual assistant device 100 and the user.

In an embodiment, the chat sessions include an initial wel^¬ come message at block 204. For example, at block 204 the fol^¬ lowing message may be displayed "Hi, welcome to virtual wind farm assistant. Say ^xhi' if you would like to chat". At block 206 if the user says ^xhi' , then at block 208 the message may include "Kindly enter your query". At block 240 the query will be input by the user and the same is visible in block 210. For example, "How to start and enable wind farm specific supervisory control and data acquisition service". At block 220, the virtual assistant device 100 responds to the query in clock 210. For the above query, the response may be "Open SCADA.sequence.ini file in the location

C:\SCADA\ProcessManager and type the below in the SCADA.sequence.ini file. #Enable SCADAHost, Enable=true. Save file". The GUI 202 also has the option of Downloading chat logs and Uploading chat logs as shown by blocks 230 and 235, respectively. The downloading chat logs are advantageous as the downloaded chats can be used for assessing the quality of the response provided by the virtual assistant device 100. Also, by uploading chat logs, the virtual assistant device 100 can treat a previous chat log as the query to generate the response.

FIG 3 is a block diagram illustrating a system 300 for assisting operation and maintenance of wind farms 330 and 340. The wind farms 330 and 340 each include a plurality of wind turbines 332a-f and 342a-c. The wind farms 330 and 340 also include private networks 335 and 345. For the purpose of the description the term "private network" refers to a network that belongs to an entity responsible for the operation of the wind power plants (WPPs . For example, the private network is the network within the demilitarized zone (DMZ) perimeter of the WPPs and is not directly accessible from outside the DMZ perimeter. The term excludes Internet 320 and Virtual Private Network (VPN) 325.

The system 100 also includes a database 302 to historical op- eration and maintenance data of each of the wind turbines 332a-f and 342a-c. Further, the system 100 includes a server 304 comprising the virtual assistant device 100. The server 304 is communicatively coupled to the database 302 via a net^¬ work interface 315. The network interface 315 is a local or restricted communications network, especially a private net^¬ work accessible to a service providing entity. The wind farms 330 and 340 are located in a remote location while the server 304 is located on a cloud server for example, using Google App engine cloud infrastructure of Google Inc., Amazon Web Services of Amazon Technologies, Inc., the Amazon elastic compute cloud EC2^® web service of Amazon Technologies, Inc., the Google^® Cloud platform of Google Inc., the Microsoft^® Cloud platform of Microsoft Corporation, etc. The server 304 is configured as the cloud server by means of the Internet 320 and communicated with the private networks 335 and 345 via the VPN 325.

The operation of the virtual assistant device 100 of the server 304 is explained in FIG 4 and 5. In FIG 4, the method 400 begins with receiving an audio stream, a video stream, a text or an image associated with the operation and mainte^¬ nance of the wind park, at step 402. The audio stream, video stream, text and image act as input to a virtual assistant device.

At step 404, the input is analyzed to determine a query asso^¬ ciated with the operation and maintenance of the wind farm. In case of the audio stream as the input, audio recognition techniques such as Hidden Markov models, Dynamic time warping (DTW) and Neural Networks are used to enable the recognition and translation of audio into text. When the input is a text, language processing techniques like stemming, phrase analysis and natural language processing, to analyze the text and de- termine the query. For an image, image processing techniques are used along with pattern recognition techniques to determine the query. In case of video stream as input, a combina^¬ tion of audio recognition techniques, image processing and language processing techniques are used to determine the que- ry.

At step 406, the query is analyzed to generate a look up ta^¬ ble of the query and one or more responses based on a histor^¬ ical operation and maintenance data of the wind park. For ex- ample, if the query is an alarm code then a look up table is generated for all the possible responses that can be given for the alarm code based on previous occurrences of the alarm.

At step 408, it is determined whether the one or more re^¬ sponses are relevant to the query. According, the virtual as^¬ sistant device decides whether to proceed with selection of a response to be displayed or to initiate communication with a remote expert.

At step 412, there is no sufficient historical operation and maintenance data associated with the query to generate the look up table and therefore, communication with the remote is initiated. The communication is initiated by means of a transmitter in the communication unit. The remote expert re^¬ ceives an alert regarding the query. Once an expert response is received, the same is output at step 416. At step 414, a response is selected from the one or more re^¬ sponses generated at step 406. The response is based on a re^¬ al-time operation and maintenance data. The real-time opera^¬ tion and maintenance data is compared with the historical op^¬ eration and maintenance data. Based on the comparison the most suitable response is selected to be output. The response can be in the format of text, audio stream, video stream and image

At step 416, the response is output based on the format. In case of text, image and video stream the response is dis^¬ played on a display of the virtual assistant device. In case of audio stream and the video stream the response is played on a speaker. At step 418, the response from the remote expert is updated in the database for the query. Additionally, the response is reviewed and assigned a positive feedback or a negative feed^¬ back by a service personnel or a reviewer. The feedback is used by the virtual assistant device to select a suitable re^¬ sponse if the query is repeated at a later date.

FIG 5 is a process flowchart illustrating a method of virtu^¬ ally assisting operation and maintenance of a wind park, ac- cording to another embodiment of the present invention. The method 500 begins at step 502 by automatically generating the query based on a comparison of the historical operation and maintenance data with the real-time operation and maintenance data. The method 500 is particularly relevant when coupled with a condition monitoring system in the wind farm. The condition monitoring system monitors the wind farm and wind turbines during the operation of the wind farm. The monitoring data is analyzed at step 502 to identify anomalies in the op^¬ eration of the wind farm.

At step 504, the query is analyzed to generate the look up table of the query and one or more responses based on the historical operation and maintenance data. At step 506, the response to be displayed is selected from the one or more re- sponses based on the real-time operation data. At step 508, an alert is generated with the response displayed as the alert .

Reference in the specification to "one embodiment" or "an em^¬ bodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the invention. The ap^¬ pearances of the phrase "in one embodiment" in various places in the specification are not necessarily all referring to the same embodiment.

In the foregoing specification, the invention has been described with reference to specific embodiments thereof. It will, however, be evident that various modifications and changes can be made thereto without departing from the broad- er spirit and scope of the invention. The specification and drawings are, accordingly, to be regarded in an illustrative rather than a restrictive sense.

Claims

Patentanspriiche / Patent claims

1. A virtual assistant device (100) for assisting operation and maintenance of a wind park comprising a plurality of wind turbines, the virtual assistant device (100) comprising:

a communication unit (130) to receive at least one query associated with the operation and maintenance of the wind park;

a memory device (102) configured to store data comprising historical and real-time operation and maintenance data of the wind park;

a processor ( 110 ) communicatively coupled to the memory de^¬ vice (102), the processor ( 110 ) configured to:

analyze the at least one query (210) to generate a look up table of the at least one query and one or more re^¬ sponses based on the historical operation and maintenance data;

generate at least one response (220) to be displayed based on the one or more responses and the real-time oper- ation and maintenance data; and

an output unit comprising a display (140) and a speaker to output the at least one response.

2. The virtual assistant device (100) of claim 1, wherein the at least one query comprises one of a text including an alarm code and a service ticket, an audio stream, a video stream and an image .

3. The virtual assistant device (100) of claim 2, wherein the processor (110) is configured to read the service ticket as the at least one query and generate the look up table of the service ticket and the one or more responses based on a ser^¬ vice ticket identifier and the historical operation and maintenance data.

4. The virtual assistant device (100) of claim 2, wherein the processor (110) is configured to read the alarm code as the at least one query to generate the look up table of the alarm code and the one or more responses based on the historical operation and maintenance data.

5. The device of claim 1, wherein the processor comprises: an audio recognition unit (112) configured to analyze an audio stream to determine the at least one query;

an image recognition unit (114) configured to analyze an image to determine the at least one query; and

a language processing unit (116) configured to analyze a text to determine the at least one query,

wherein the audio recognition unit, the image recognition unit and the language processing unit are configured to ana^¬ lyze a video stream to determine the at least one query.

6. The virtual assistant device (100) of claim 1, wherein the historical and real-time operation and maintenance data com^¬ prises operating procedures of the wind park including re^¬ striction and limiting operations, standard operating proce^¬ dures, high risk operating procedure, and identity and access control information of a servicing entity.

7. The virtual assistant device (100) of claim 6, wherein the processor (110) is configured to generate the at least one response based on the identity and the access control infor- mation of the servicing entity.

8. The virtual assistant device (100) of claim 1, wherein the one or more responses comprises one of a text, an audio stream, a video stream and an image.

9. The virtual assistant device (100) of claim 1, wherein the processor (110) is configured to:

automatically generate the at least one query based on comparison of the historical operation and maintenance data and the real-time operation and maintenance data;

analyze the at least one query to generate a look up table of the at least one query and the one or more responses based on the real-time operation and maintenance data; and generate the at least one response based on the one or more responses as an alert.

10. The virtual assistant device (100) of claim 1, wherein the processor (110) is configured to initiate a communication session with a remote expert via the communication unit (130), to generate the at least one response.

11. A system (300) for virtually assisting operation and maintenance of a plurality of wind parks (330,340) each com^¬ prising a plurality of wind turbines (332 a-f, 342 a-c) , the system comprising:

a database (302) configured to store data comprising his^¬ torical operation and maintenance data of each of the plural- ity of wind parks; and

a server (304) communicatively coupled to the database by means of a network interface (315) ,

wherein the server comprises at least one virtual assistant device (100) according to claims 1-9 for assisting operation and maintenance of the plurality of wind parks (330,340) .

12. A method of virtually assisting operation and mainte^¬ nance of a wind park comprising a plurality of wind turbines, the method comprising:

receiving at least one query associated with the operation and maintenance of the wind park;

analyzing the at least one query to generate a look up ta^¬ ble of the at least one query and one or more responses based on a historical operation and maintenance data of the wind park;

generating at least one response based on the one or more responses and a real-time operation and maintenance data; and outputting the at least one response in response to the at least one query.

13. The method of claim 12, further comprising: automatically generating the at least one query based on a comparison of the historical and real-time operation and maintenance data;

analyzing the at least one query to generate the look up table of the at least one query and the one or more responses based on the historical operation and maintenance data;

generating the at least one response based on the one or more responses based on the real-time operation data; and

displaying the at least one response as an alert.

14. The method of claim 12, further comprises:

initiating a communication session with a remote expert to generate the at least one response.

15. The method of claim 12, wherein receiving at least one query associated with the operation and maintenance of the wind park comprises:

receiving one of an audio stream, a video stream, a text and an image associated with the operation and maintenance of the wind park; and

determining the at least one query from one of the audio stream, the video stream, the text and the image associated with the operation and maintenance of the wind park.

16. The method of claim 12, comprising:

reviewing the at least one response to the at least one query; and

assigning one of a positive feedback and a negative feed^¬ back to the at least one response based on a user experience.