WO2007125108A1

WO2007125108A1 - A method and system for controlling an industrial process including automatically displaying information generated in response to a query in an industrial installation

Info

Publication number: WO2007125108A1
Application number: PCT/EP2007/054186
Authority: WO
Inventors: Magnus Larsson; Pierre ÖBERG; Morgan Johansson; Juan Stamm'ler Jaliff
Original assignee: Abb Research Ltd
Priority date: 2006-04-27
Filing date: 2007-04-27
Publication date: 2007-11-08

Abstract

A method for controlling an industrial process or equipment by means of displaying information generated automatically in response to a query concerning an apparatus, device or equipment in an industrial process or facility. Device or process-related information and other data for each said equipment, plant or process are stored and associated with said control system. The method comprises receiving as input into a system a said query expressed in part in a natural language, determining a data category to which said query is related, retrieving one or more documents dependent on the data category, ranking and marking the retrieved documents dependent on a confidence rating of an information source, and displaying one or more documents in an order or format dependent on the ranking given to each of the one or more documents. A system and a computer program are also described.

Description

A method and system for controlling an industrial process including automatically displaying information generated in response to a query in an industrial installation

TECHNICAL FIELD

This invention relates to an improved method and system for controlling and industrial process and for providing maintenance and/or control data for an equipment, plant or a process by means of accessing data and stored information, processing, generating and displaying information relevant to the device or process . In particular the invention provides an improved method for automatically retrieving, processing and displaying information in response to a query input in a natural language concerning a process, an equipment, plant or process and controlling said such process or equipment.

TECHNICAL BACKGROUND

Monitoring and control of plant and equipment in industrial facilities is largely automated and computerised. A traditional automation system is one in which equipment or part of the equipment is described in a database. In its turn, this database is frequently modeled as a multitude of tables, and relationships between different fields and tables in the database. Process monitoring and control software operates using, in part, data stored in a flat file or relational or relational distributed databases .

Although much technical information about various devices, equipment, process sections etc exists, it is most often distributed among different computerised systems each having different categories of data and/or different methods of storing and retrieving technical data. Examples of such existing and often unconnected systems are computerized maintenance systems, process control systems, power management systems, energy management systems, and systems for process simulations and optimization. Examples of information to be found in such systems are maintenance histories, production histories, configuration set- points, production flows, compressor efficiency curves, references to load characteristics of drive solution, energy consumption logs, ratings for electric motors, information about material costs and energy costs, and so on.

An installation may typically include equipment from different suppliers and from different industries. It is a complex and difficult task to retrieve information from all of the existing and often unconnected systems so as to give a reliable and extensive picture even on a historical basis. It is even more difficult to retrieve and or access such information when an alarm or other unplanned event is reported.

However the prior art does not provide access to comprehensive information about the plant, device or process section in a timely and simple way that supports fast or on-line operations. Searching for information can often be difficult. Many people over the last couple of years have learnt to "google" for information. This approach allows a user to pose a question in the form of keywords and receive a list of documents that are possibly relevant. The process of indexing is done automatically, i.e. to find which documents contain the word or phrase "XXX NNN" . The resultant list is usually ordered by a ranking process. Google and other search mechanisms may also use Boolean searching approaches to find documents containing certain words and, for example, not containing certain other words. Traditional search engine technology for information on the Internet classifies information into a number of static categories which are pre-defined by a group of people. An example of this are Yahoo's categories for web sites. The process of assigning a web site to a category is done manually.

The quality of the search results provided by traditional Internet technology for technical documentation is, however somewhat limited, despite the existence of some manual and perhaps arbitrary classifications of information categories. The industrially applied task of finding technical information held by an industrial control system and associated data sources remains difficult and time-consuming.

One way to tackle this challenge is to create a type of indexing that would mimic the way that paper systems have been arranged. Information may manually classified into a series of files mimicking the folders, binders and similar that are traditionally used and maintained. However, this solution is time consuming, expensive, and may be heavily operator dependent on an operator or engineer or on a particular expert.

SUMMARY OF THE INVENTION

The aim of the present invention is to provide an improved method and system for automatically displaying information generated in response to a query in an industrial installation. This and other aims are obtained by a method and a system characterised by the attached independent claims . Advantageous embodiments are described in sub-claims to the above independent claims.

According to a first aspect of the invention a method is disclosed for controlling an industrial process in an industrial installation or facility comprising one or more control systems for monitoring and control, wherein certain device or process- related information and other data for each said equipment, plant or process are stored and associated with said control system, said method comprising automatically generating and displaying information in response to a query concerning an apparatus, device or equipment, and wherein the method comprises: receiving as input into a computer implemented system a said query expressed in part in a natural language, determining a data category to which said query is related based on a part of the natural language comprised in said query, retrieving one or more data items or documents dependent on the data category that may be matched to said query, ranking and marking the retrieved one or more data items or documents dependent on a confidence rating/interval of an information source, and displaying on at least one display means one or more data items or documents in an order ranked dependent on the ranking given to each of the one or more documents.

According to an embodiment of the invention a method for controlling an industrial process in an industrial installation or facility comprising one or more control systems for monitoring and control, wherein certain device or process-related information and other data for each said equipment, plant or process are stored and associated with said control system, is disclosed which additionally comprises the steps of determining for each said query which of the retrieved one or more data items or documents are examined by an operator entering said query.

According to another embodiment of the invention a method for controlling an industrial process in an industrial installation or facility comprising one or more control systems for monitoring and control, wherein certain device or process-related information and other data for each said equipment, plant or process are stored and associated with said control system, is disclosed which additionally comprises the steps of examining and recording the document source for each of the retrieved one or more documents for each said query which were examined by the operator entering said query.

According to an embodiment of the invention an improved computer- based method adapted for controlling an industrial process in an industrial installation or facility comprising one or more control systems for monitoring and control is disclosed, wherein certain device or process-related information and other data for each said equipment, plant or process are stored and associated with said control system, said method comprising automatically generating and displaying information in response to a query concerning an apparatus, device or equipment, wherein said method comprises the steps of inputting a query expressed in a natural language, resolving the query into parts for processing using a query engine, providing, by means of a natural language query engine an answer to the original question.

According to an embodiment of the invention an improved computer- based method adapted for controlling an industrial process in an industrial installation or facility comprising one or more control systems for monitoring and control is disclosed, wherein certain device or process-related information and other data for each said equipment, plant or process are stored and associated with said control system, said method further comprising automatically generating and displaying information in response to a query concerning an apparatus, device or equipment, and wherein by resolving, using a natural language query engine means, sub- questions that can be used to determine the answer to a natural language query, processing each sub-question, and collecting information from the industrial control system or other data source via the industrial control system, compiling and presenting an answer to the original question. According to an embodiment of the invention an improved computer- based method adapted for controlling an industrial process in an industrial installation or facility comprising one or more control systems for monitoring and control is disclosed, wherein certain device or process-related information and other data for each said equipment, plant or process are stored and associated with said control system, said method further comprising by identifying any domains of the industrial control system or linked to said system in which a data source comprising an answer to each sub-question can be found, and processing each sub-question, and collecting information from the industrial control system.

According to another embodiment of the invention a method for controlling an industrial process in an industrial installation or facility comprising one or more control systems for monitoring and control is disclosed, wherein certain device or process-related information and other data for each said equipment, plant or process are stored and associated with said control system, is disclosed which additionally comprises the steps of calculating a confidence interval/rating for a document source for the category or categories identified in said query dependent on which of the retrieved one or more documents were examined by the operator.

According to an embodiment of the invention a method for controlling an industrial process is disclosed in which the data category comprises any from the group of: current process data, short term history, long term history, batch history, audit log, quality history, maintenance history, alarms and events.

According to an embodiment of the invention a method for controlling an industrial process is disclosed in which the data category or data category and confidence rating is arranged available for processing a query input by more than one user and/or in more than one installation. In another aspect of the invention a method is disclosed for displaying information generated automatically in response to a query concerning an apparatus, device or equipment in an industrial process or facility comprising one or more control systems for monitoring and control, wherein certain device or process-related information and other data for each said equipment, plant or process are stored and associated with said control system, the method further comprising receiving as input into a computer implemented system a said query expressed in part in a natural language, determining a data category to which said query is related based on a part of the natural language comprised in said query, retrieving one or more data items or documents dependent on the data category that may be matched to said query, ranking and marking the retrieved one or more data items or documents dependent on a confidence rating/interval of an information source, and displaying on at least one display means one or more data items or documents in an order ranked dependent on the ranking given to each of the one or more documents .

According to another embodiment of the invention a method for controlling an industrial process is disclosed which additionally comprises the step of creating a new category and recording a flat confidence rating for the new data category.

According to another embodiment of the invention a method for controlling an industrial process is disclosed which additionally comprises the step of assigning the confidence interval rating of a document source per category.

According to another embodiment of the invention a method for controlling an industrial process is disclosed which additionally comprises the steps of receiving as input a ranking from the operator of the retrieved one or more data items or documents. According to another, further embodiment of the invention a method for controlling an industrial process is disclosed which additionally comprises the step of inferring a technical context of said query based on information from any of the group of: a log-in; a log-in related profile; recently opened displays or view; workstation location; or by means of operator input of information about a context .

According to another, further embodiment of the invention a method for controlling an industrial process is disclosed which additionally comprises the step of inferring a technical context such as any from the group of: normal operation product A, product B, product C; or by generating a prompt and requesting the operator to confirm context information or input information about a context.

According to another embodiment of the invention a method for controlling an industrial process is disclosed which additionally comprises the step of requesting the operator to confirm information representing a technical context such as any from the group of: abnormal operation product A, product B, product C; process troubleshooting, process configuration I, II, III; process quality troubleshooting product A, B, C; plant maintenance general .

According to another, further embodiment of the invention a method for controlling an industrial process is disclosed which additionally comprises the steps of inferring information about a technical role of the operator who had input said query in the industrial installation from any from the group of: a log-in; a log-in related profile; log-in privileges, log-in user authorities; recently opened displays or view; workstation location. According to another, further embodiment of the invention a method for controlling an industrial process is disclosed wherein information about a technical context or a technical role is arranged available for processing a query input by more than one user and/or in more than one installation.

According to another, further embodiment of the invention a method for controlling an industrial process is disclosed which additionally comprises the step of determining a data category by recognising a pattern of operator manipulations or selections on a graphical user interface preferably by from comparing one or more current operator actions on the graphical user interface with stored patterns of operator manipulations or selections related to one or more data categories of interest.

According to another aspect of the invention a system is disclosed for controlling an industrial process in an industrial installation or facility comprising one or more control systems for monitoring and control, wherein certain device or process- related information and other data for each said equipment, plant or process are stored and associated with said control system, said system further comprising at least one computer or computer terminal with computer input means connected to said control system, and at least one display means connected to said control system, further comprising an HMI for receiving a query input by a user, a search engine for receiving the query input in a natural language, and a display means arranged to display one or more data items or documents in an order dependent on a ranking of the data items or documents so found.

According to an embodiment of the invention a system is disclosed for controlling an industrial process in an industrial installation or facility comprising one or more control systems for monitoring and control, wherein certain device or process- related information and other data for each said equipment, plant or process are stored and associated with said control system, said system further comprising a natural language query engine comprising means for resolving the natural language query into parts for processing using the query engine, and means for providing, by means of a natural language query engine an answer to the original question.

According to another embodiment of the invention a system for controlling an industrial process in an industrial installation or facility is disclosed, said system further comprising computer program means to record an active user action to retrieve or open one of more of the automatically displayed data items or documents .

According to another embodiment of the invention a system for controlling an industrial process in an industrial installation or facility is disclosed, said system further comprising computer program means to record one or more actions input by the user for search purposes and associate those actions with the user by means of the identification that the user is logged-in with.

According to another embodiment of the invention a system for controlling an industrial process in an industrial installation or facility is disclosed, said system further comprising computer program means to record one or more actions input by the user for each of the retrieved one or more data items or documents for each said query which were examined by the operator entering said query.

According to another embodiment of the invention a system is disclosed for controlling an industrial process in an industrial installation or facility, said system further comprising computer program means for calculating a confidence interval/rating for a document source for the category or categories identified in said query dependent on which of the retrieved one or more data items or documents were examined by the operator.

According to another embodiment of the invention a system is disclosed for controlling an industrial process in an industrial installation or facility further comprising computer program means for creating a new category, assigning a confidence interval rating of a document source per category, recording a flat confidence rating for a new data category, and/or updating the confidence interval rating of a data source per data category.

According to another embodiment of the invention a system for controlling an industrial process in an industrial installation or facility is disclosed, said system further comprising computer program means and input means for receiving as input a ranking from the operator of the retrieved one or more data items or documents, and/or generating a prompt and presenting it to the operator requesting the operator to make a ranking of the retrieved one or more data items or documents, and/or a step of receiving, as an input, information about a technical context of said query.

According to another embodiment of the invention a system for controlling an industrial process in an industrial installation or facility is disclosed, said system further comprising computer program means and graphical user interface means for inferring a technical context such as any from the group of : normal operation product A, product B, product C; or interface means for generating a prompt and requesting the operator to confirm context information or input information about a context . According to another embodiment of the invention a system for controlling an industrial process in an industrial installation or facility is disclosed, said system further comprising computer program means for inferring a technical context such as any from the group of: abnormal operation product A, product B, product C; process troubleshooting, process configuration I, II, III; process quality troubleshooting product A, B, C; plant maintenance general, or means for requesting the operator to input information representing a technical context.

According to another embodiment of the invention a system for controlling an industrial process in an industrial installation or facility is disclosed, said system further comprising computer program means for determining from a plurality of operator navigations or manipulations or selections carried out on a graphical user interface a likely data category for a current query of interest.

In another aspect of the invention a system is disclosed for displaying information generated automatically in response to a query concerning an apparatus, device or equipment in an industrial process or facility comprising one or more control systems for monitoring and control, wherein certain device or process-related information and other data for each said equipment, plant or process are stored and associated with said control system, comprising at least one computer or computer terminal with computer input means connected to said control system, and at least one display means connected to said control system, further comprising an HMI for receiving a query input by a user, a search engine such as a natural language search engine, for receiving the query input in a natural language, and a display means arranged to display one or more data items or documents in an order dependent on a ranking of the data items or documents so found. The present invention provides as software and/or hardware comprising one or more methods for controlling an industrial process which software automatically displays qualified information in response to complex queries posed by a user towards an industrial control system. The question may be posed in a natural language. The invention provides a means for quicker trouble shooting and for simplified access to technical data and information that today is hard to find.

The invented method and system displays qualified information automatically in response to a query. The information found is ranked automatically according to one or more different methods. In one aspect the invented method and system automatically classifies and indexes information sources for information about an industrial process, including its operational parameters. The data categories used may be either given by the user and/or may be inferred by means in the system based on queries posed by an operator. The system may handle multiple sources in the same data category by assigning an internal confidence ranking to the various sources. In this may a control action may be carried out based on the qualified information so retrieved.

A system described in an application number US 60/795,171 entitled "An improved computer-based method of information retrieval" filed April 27, 2006, and hereby incorporated in this document in full by means of this reference, describes a method and system which can interpret queries posed in an unstructured way and in a natural language. The system includes a natural language query engine comprising means for resolving the natural language query into parts for processing using the query engine. However, another an additional challenge for handling queries in a computerised system is that the system needs to know in which sources of data store which kinds of information that may be found. In a web example, geography data may be fetched from a predefined set of sources, weather information from a second set, and so on. However in a process automation system, applying this technique would be very manpower-intensive, if the classification were to be done manually and/or under supervision of a domain expert. Furthermore, many conventional methods do not assign different confidence values to different sources.

In an aspect of the invention a natural language query engine such as the START engine developed by MIT is used in part to "understand and translate" questions posed in a natural language. The natural language system including a natural language query engine also comprises a lexicon function and a function for acquiring domain knowledge. The system is configured for one or more specific domains in an area of an industrial application by "exhausting" as many user scenarios as possible and thus making sure that the lexicon is as complete as possible and also that the domain knowledge is complete. In this description, the terms domain and domain knowledge refers to domains comprising one of more different applications of monitoring and/or control systems or processes, such as a control system of an oil and gas installation, a control system for an installation in the pulp and paper industry, a control system of an installation for electrical power generation, or transmission or distribution, etc. The natural language query engine, a START engine in this case, will translate the questions into a "symbolic query language" . These queries expressed in a symbolic query language may be handled by one or more additional applications or middleware that gathers data from the various data sources (for example many or diverse data types which may, for example, comprise OPC data, as-logged high resolution data, low resolution recordings of data, data from the aspect directory, Windows data, data from a maintenance History, production History, batch information, vibration analyses, sound clips, image data, IR image data, CCTV (closed circuit TV) images, in short any or all types of data found in or derived from data sensed, calculated or collected in or for an industrial installation comprised in or associated with the industrial control system.

An advantage of the improved method is that users of an industrial control system implementing the improved method may carry out trouble shooting more quickly and gain access to technical information that is hard to find using existing industrial methods and systems of the Prior Art. In particular a user does not require special training in search techniques, Boolean searching or similar in order to use the improved method and locate the information required. The improved method allows a user to pose one or more questions using natural language and rapidly get back the "right answer" from a process automation system. The improvement provides a mechanism of question/answer that greatly simplifies the search for technical information. The questions are, as stated above, formulated in a natural language, which is in strong contrast to the query languages, keyword searches or other techniques of the Prior Art such as SQL and google.

An additional advantage of the invention is that if an exact answer can not be determined then the query system may also make assumptions and present related information that the user can benefit from. For instance, if no answer can be found to the question; show me a detailed graphical display of motorl23, then a result could be a general display that shows the whole process section to which motorl23 belongs.

One or more aspects of the invention may be applied to many different kinds of industrial installations and to the processes and equipment included in them. In a process industry such as for example from the pulp and paper industry, the chemical industry, or in a metal processing foundry or a rolling mill. Furthermore the invention may be applied in almost any industrial type of installation including energy installations such as oil and gas extraction or processing or electrical generation and distribution installations.

A major advantage of the present invention is that information to respond to an alarm or other event may be obtained quickly so that the alarm or event may be handled in a timely way. The information retrieved may also provide instructions to one or more users for controlling a process, and/or configuring, troubleshooting or repairing etc an item of equipment.

In another aspect of the invention, the method and system of the present invention learns from the replies selected by its users which data sources may be regarded as having more confidence. It can start from a flat confidence distribution across sources, e.g.

-list for the user all possible sources of any info in the system.

As the user chooses which sources are more suitable for the query posed, the system:

-creates a category or selects an existing category, and -assigns a confidence ranking to the data sources which is consistent with the user's selection/s. The system of classification in categories may in addition also take into account a role of the user such as e.g. logged in as a process operator, systems engineer, maintenance engineer, etc., and/or a context of the query such as e.g. safety, planning of future plant operation, equipment diagnostics, etc.

An advantage of the improved method and system is that users of an industrial control system implementing the improved method may carry out trouble shooting more quickly and gain easy access to information that is hard to find using existing industrial methods and systems of the Prior Art. It is a major advantage that for example, a maintenance action required to respond to a new alarm or other event may be handled in a more timely way. The information retrieved and/or developed during collaboration provides instruction to the user for monitoring, troubleshooting, configuring or repairing the equipment.

Another advantage of the invention is that it allows a user to pose questions using natural language and get the "right answer" back from a process automation system. The improvement provides a "question - answer" mechanism that simplifies the search for information. The questions are formulated in natural language in contrast to the query languages of the Prior Art such as SQL and google. Another advantage of the invention is that it allows a user to pose questions by making selections in a graphic user interface (GUI) or other human machine interface (HMI) and get the "right answer" back from a process automation system. The improvement provides a "question - answer" mechanism that simplifies the search for information. Another advantage of the invention is that it allows a user to pose questions and get a qualified answer or a "right answer" relatively quickly which leads to a reduction in operator stress.

Another further advantage of the invention is that if an exact answer can not be determined then the query system will make assumptions and present related information that the user can benefit from. For instance, if no answer can be found to the question -show me a detailed graphical display of motorl23, then a result could be a general display that shows the whole process section to which this motor belongs . In another embodiment the invention provides the advantage that the development and qualification of data sources within the categories and classification that are constructed provides a display tool that: - shares knowledge across operators and plants

- expands easily with new sources of information in a plant, e.g. a new database. The method and system are facilitated by one or more computer programs comprised in the control system.

BRIEF DESCRIPTION OF THE DRAWINGS

Embodiments of the invention will now be described, by way of example only, with particular reference to the accompanying drawings in which:

FIGURE 1 shows a schematic diagram of a GUI for inputting a query according to an embodiment of the invention;

FIGURE 2 is a flowchart for a method according to an embodiment of the invention;

FIGURE 3 is another flowchart for a method according to another embodiment of the invention; FIGURE 4 shows a flowchart for a method for finding documents by category of data source and displaying them, according to an embodiment of the invention;

FIGURE 5 is a flowchart for a method for ranking documents found, according to an embodiment of the invention; FIGURE 6 is a flowchart for another method in which ranking is updated by automatically recording user actions, according to another embodiment of the invention;

FIGURE 7 is a diagram of data categories for an industrial installation, and Figure 8 is a diagram comprising contexts for an industrial installation, according to an embodiment of the invention; FIGURE 9 is a flowchart for a method in which documents may also be ranked according to additional inputs for a context, a role, a plant location, according to another embodiment of the invention;

FIGURE 10 is a schematic diagram for a display of a display means presenting ranked documents found by a method according to an embodiment of the invention;

FIGURE 11 is a schematic diagram for one or more graphic user interfaces with graphical representations of objects controlled by an industrial control system according to another embodiment of the invention.

DETAILED DESCRIPTION OF THE EMBODIMENTS

The invented system has features important for an operator or user concerned with controlling an industrial process in an industrial plant. The operator or other user may retrieve information from the multiple sources of information and data associated with an industrial plant. As mentioned above, the method in a first aspect comprises use of a natural language query engine comprising means for resolving the natural language query into parts for processing. Thus when a query is input via an input means connected to a computer, computing device or computer terminal, the natural language query engine, or NLQE, breaks the input down into parts for processing.

Figure 1 shows an exemplary interface according to an embodiment of the invention. A screen for input 4 is displayed. The input screen comprises a box in the form of a text field 2 into which a question 1 may be typed, copied or otherwise input. The screen may have instructions 5 informing the user how to input a question or use the natural language query (NLQ) function. The screen may include links to more information on use of the NLQ function. When a question has been input the user starts the search function by clicking on a Search button 3, pressing "Enter" or by clicking any- similar GUI activation means.

In one embodiment of the invention, the method uses a version of the START natural language query engine developed by MIT to "understand and translate" what may be a relatively complex question posed in a natural language. The natural language system including a natural language query engine may also comprise a lexicon function and a function for acquiring domain knowledge. The system has preferably been configured to identify a plurality of specific knowledge domains by testing or "exhausting" as many user scenarios as possible and thus making sure that the lexicon compiled is as complete as possible and also that the domain knowledge has been processed. The START engine will translate the question when input into a form of symbolic query language. The queries may be handled by another application such as a middleware that gathers data from various data sources (e.g. OPC data, data from the aspect directory, Windows data, data from maintenance History, production History, or other and diverse data types as described above) comprised in or associated with the industrial control system.

The answer to the following simple examples of questions, (each question being written here enclosed in double quotation marks eg "question"), which would form part of a technical task of controlling, configuring or maintaining a technical process may be generated by means of sub-queries as follows:

"Show me all motors that are running above 1000 rpm"

- This question is answered by displaying a list of all motors that meet the stated criteria.

"List operator actions the last 5 hours on PID loops in Areal Building5 that have a deviation greater than 2%" sub queries: "List PID loops in Areal and Buildingδ that have a deviation greater than 2%" from the list of objects new subqueries are compiled:

"List operator actions the last 5 hours on PIDl" "List operator actions the last 5 hours on PID2"

"List operator actions the last 5 hours on PID3", and so on.

"Which of all valves in Blending that are not closed have had an alarm the last 24 hours?" sub queries: "Which of all valves in Blending are not closed"

From the list of objects new sub-queries are compiled: "List alarms for VALVEl the last 24 hours" "List alarms for VALVE2 the last 24 hours"

"List alarms for VALVE3 the last 24 hours" "List alarms for VALVE4 the last 24 hours", and so on.

An important aspect is that information about the context of a previous question is kept and used to determine which device eg which PIDCON to query. The information may be classified dependent in some way on the operational context of the previous question, or to a domain and/or task and/or to a user ID.

The user may interact with an industrial control system to control, configure or maintain etc a part of a controlled process or apparatus in the following way (see Fig 1) :

1. Activate the search page 4 or other interface to be able to enter a query for the system,

2. The user enters a question, question 1, using a natural language 3. The natural language query engine resolves the question into sub-questions that can be used to determine the answer, dependent on domain knowledge of the industrial control system or systems,

4. The query system determines in which data source the answer to each sub-query can be found

5. Each sub-query is processed and any information requested is collected from the industrial control system

6. All data received for each sub-query are processed and compiled into an answer of the original posted question 7. The answer is presented to the user in a suitable form that represents the answer, which may be dependent on an operational context, such as of: maintenance, troubleshooting, operations, configuration, and which may be in a form appropriate to a context and/or a task, for example generated in the form of a printout, a diagram, a display of a limited amount of text, display of any amount of scrollable text, document with active embedded links, a video clip, an audio clip, and so on. The answer may also for instance comprise a list of named objects with properties, and/or information in the form of text, text and diagrams, a report or a graphical display containing the information requested.

The question 1 is resolved into parts or sub-questions for processing using a query engine which classifies one or more parts of the question in terms of what type of information each part of the question represents. Data items in all the known data sources stored by or accessible to the Industrial Control System (ICS) may be classified by context, task or in some other way according to each type of information found in the item.

Figure 2 shows a flowchart for a method of providing an answer to a question expressed in a natural language. Figure 2 begins :

22 Input a question expressed in a natural language, 24 resolve, using a natural language query engine means, the question into one or more sub- queries,

28a process each sub query,

28b collect information from the Industrial Control System (ICS) or a data source via the ICS,

32 compile answers to each sub query and present the answer to the original question.

Optionally in an alternative method as indicated by the dotted line surrounding the process stage, a control signal may be generated:

34 provide a control signal based on the information provided by the answer to directly or indirectly control an apparatus, device, equipment or process.

Figure 3 shows a flowchart for a method of providing an answer to a question expressed in a natural language according to another embodiment . Figure 3 begins :

20 activate an interface with which to input a question, 22 input a question expressed in a natural language,

24' resolve the question into one or more parts or sub- questions that can be used to determine the answer using the natural language query engine,

25 apply domain knowledge of the industrial control system (ICS) to identify one or more relevant data sources

26 determine using the natural language query engine in which data source the answer to each sub query may be found,

28 process each sub- query and collect any information requested from the industrial control system (ICS) or from other data sources via the ICS 30 process and compile all data received for each sub query into an answer to the originally posed question,

32 present the answer to the user in a suitable form that represents the answer and corresponds in part to the operational context and/or task of the user, eg in a form such a small-format visual display, an SMS or other display of a limited amount of text, a display of any amount of scrollable text, a document with active embedded links, a video clip, an audio clip and so on.

The information may be arranged as, for example, a printout, a wiring schema, configuration chart, a diagram, process flow diagram, an animated computer simulation or model of a process or device, text, pictures, video etc. The information may be presented in computerized combinations of information, such as a virtual reality image superimposed on part of a real world view. The information may be generated and presented dependent on for example task, user, and the display/presentation capabilities of the user equipment. Information in such formats may also be comprised for example as a list of objects, a table of objects with properties, text information, a value, a parameter, a graphical display or diagram, or a composite image comprising any of the above, corresponding to the information content required and the context of one or more of the information types identified.

The main steps taken by the system when processing the query may be: a) the system determines a data category from the natural language in a query, e.g. "equipment diagnostics", and creates it as new if not already present b) when a category is first created, the system responds to the query with the whole matching indexing, i.e. searching in all sources for the type of information which is asked for in the query (this may be similar to a search for text on the world wide web looking at all indexed sites) c) the system records which documents from each data source the user chooses to "see", and classifies these data sources for the category identified in a) with an equal value of confidence level for all sources d) after the user is done, and only if this feature is enabled, the system presents to the user all sources which he/she "saw", and asks the user to rank the documents found according to confidence for the data category of interest e) based on the reply, it will modify the confidence values to the data sources for the given category

Fig 7 shows a list for an exemplary set of data categories within an industrial installation, which may comprise the following categories : current process, short term history, long term history, batch history, audit log, quality history, maintenance history, alarms and events .

In step (c) above, the system records which documents from each data source the user chose to see or examine. This may be determined by the system by recording which document links or files were opened for this category. Thus a confidence rating per category for a given data sources is established and, over time modified and updated based on an automatic registration of user response. The confidence rating may also be further modified depending on response from a user to a request made by the system to the user to rate the source, as described in step (d) .

In response to a query input by a user in a natural language via an input means of a computer or computer terminal, the system automatically displays all sources found for the type of information which is asked for in the query, ranked according to a confidence rating. The automatic display is provided to any display means that may be connected in some way to a computer or the system. Thus the user when inputting a query may not be required to describe a category but still receives a display of a ranked list of sources. The ranking in response to a query is indicated in the display of information automatically. In its simplest form, the display presents documents from data sources wherein a document with the highest confidence rating for a given category is, for example, displayed first at the top of a list; and the source with the lowest or a flat confidence rating is for example displayed last or towards the end of the list. The list is usually a list of document titles. Entries on the list may include a category of interest. Entries on the list may include a value for confidence rating, expressed for example as a percentage.

The format of the content automatically displayed by the display means may be varied depending on the specification of the display means, and/or on the data capacity of the data communication to the device in which the display means is comprised. Thus, for example, a "text only" version of the ranked list may be presented on a display that is limited to text displays only. If a particular format or protocol such as of a WAP device or mobile phone based device is detected, the display format is selected or optimised by the system for display on a small screen.

Alternatively, or as well, the display screen specification may be found by examining the user log-in identification information. Such log in information may contain or be linked to information about the display capabilities of the user's workstation or portable computing device., and/or the communications capacity of the wired or wireless connection. The display content may alternatively be optimised to show one or more of documents from the highest ranked sources be presented in another or in a different visual format. For example, one or more of the highest ranked sources may also be displayed in part in an additional window, box or other graphic element, providing an extract or preview of the document. Thus the system may automatically display for the user a preview of some of the text or graphics for a high or highest ranked source. The additional information provided by the "preview" box, or similar display function, is displayed to the user without any further action required from the user.

Thus the user is presented with one or more source documents which are automatically displayed as ranked in the order of the confidence rating for that data source for that category which, over time, shall be the same as the relevance of the source to the subject of the question. Figure 4 shows a flowchart for a method, in which the following actions are represented:

110 Receive query input, in the form of a natural language 112 Identify a data category, from the query 114 Search a data category for query

116 Retrieve documents, within the category that match the query

120 Rank the documents according to confidence rating for the category 124 Display the documents dependent on rank order

Figure 5 shows a flowchart for a method that checks if sources in a data category are ranked, and if not, creates a basic ranking, in which the following actions in addition to blocks 110, 112, 114, 116 are represented:

117 Are the sources in this data category ranked? 119 If No, the document source is given a flat ranking to begin with;

120 If Yes, the documents are ranked according to the ranking of the data sources for the given category 124 Documents displayed dependent on the rank order

In one aspect of the invention each given data source is further ranked by automatically tracking which of the displayed documents the user examines. After block 124, in which the documents are displayed in a way dependent on rank order, an application program or part thereof records if the user opens any document. The information, about which of the data source documents found for that category are also examined by the user, is then used as an additional input to the automatic ranking process. The information about which documents were examined is used to update or otherwise refine the confidence rating for the given data source for the given category.

Figure 6 shows a flowchart for a method as described under section (c) above, a method that records if any and if so which of the documents displayed are examined by a user, and as in (d) where ranking is updated according to such user actions, in which the following actions in addition to the blocks 110, 112, 114, 116, 117, (119), 120 are represented: 124 Documents displayed dependent on the rank order

126 Record which documents are examined, record which documents are opened, or which links are clicked on, for example

128 Update ranking of the data source of the examined document for this category. Figure 10 is a schematic illustration for a display. The figure shows a display means 106. Displayed on the display means is a display X which comprises a list 103 of found documents arranged ranked in an order 103b (here indicated with numbers 1-10 as an example) and displayed by means of at least as a title of some kind 101. Each document is indicated 103a to have a ranking, such as document 10 in the list, 1 in the list. Each item in the list, each document, is represented such that the title or other graphic object is an executable item, such as a HTML-type link to a document, or a self executing, or self opening file. In this example a graphic element 107, a preview box, is schematically shown displaying additional content details for a high ranked document, ranked number 1 on the display of documents found.

Furthermore, in another embodiment, the system: f) may use the procedures above and share categories and confidence values across more than one user and more than one plant g) the system may also reassign the category (identified for a question from its natural language) by taking into consideration a "context" given by the user, e.g. "safety", "planning of future plant operations" h) the system can share categories and confidence values as in f) but across groups of users sharing a particular "role", e.g. "operator", "systems engineer", "maintenance engineer", "production manager"

The system and methods may provide an additional means for classifying and qualifying technical information by processing queries according to an indicated context. An exemplary set of contexts within an industrial installation, as shown in the example in Fig 6, may comprise the following functional or operational contexts:

-normal operation product A, product B, product C -abnormal operation product A, product B, product C -process troubleshooting, process configuration I, II, III -process quality troubleshooting, product A, B, C -plant maintenance general .

Figure 9 shows a flowchart for a method as described under sections (g, h) above, a method in which other qualifiers "context" and/or "role" and/or "plant location" may also be input, in which the following actions are represented:

110 Receive query input, in the form of a natural language Ilia Receive context input, manual or other input

111b Receive role input, derived from any of log-in, location in plant, current alarms in plant location, recorded user activity history, manual input or other input

111c Receive plant, section or installation input, input which defines physical location in terms of a process, process section or equipment, (which input may be provided by position measurement, estimates of position measurement, and or stored data on position in plant for devices or processes) ,

112 Identify a data category, from the query 114 Search a data category for query

116 Retrieve documents, within the category that match the query

120' Rank the documents according to confidence rating for the category and for any of context, user role, plant section

124 Documents displayed dependent on the rank order

Queries to the system, and replies from it, may be provided via computer implemented functions, for example such as:

- web services

- programmatic API ' s - standard database query language interfaces, e.g. JDBC In another embodiment a data category may be calculated based on graphical software entities or objects on a graphical user interface. An operator may examine a view in the control system of a part of the process. The relative position of graphic representations of objects to each other on the screen may be assessed and recognized by a pattern recognition application. This may be based on graphical representation parts and relationships between the real world entities they represent; and/or may be based in part on proximity to one graphic object or point to another graphic object; and/or may be based in part on a graphical path leading to the present display or document on the display; such as that it may be based on a succession of computer input control actions, for example a pattern of Left and /or Right mouse clicks, Function key selections and the screens or views so navigated.

For example, Figure 11 shows a simplified display for one or more graphical user interfaces with graphical representations of control objects in a control system displaying information related to different functional facets of a pump. A unique real world control valve 90 may be represented by a software entity 91 on a graphical user interface CS running on a display 106'. The software entity 91 may comprise many facets of the control valve. In this schematic example only four displays of different GUIs of facets in which the control valve has a role are shown. A first GUI, representing a Control Diagram 93 represents a part of a control diagram of a process section of which the control valve may be a part; a Process Diagram 94, showing process details for a process section including the control valve, with real time values and valve positions and, for example, showing location and connections of the control valve in the process scheme; a display of a Faceplate 95, which may display one or more set points, values etc. of the control valve; a Maintenance history 96 for the control valve in which maintenance events and information may be found all the way back to commissioning.

When an operator navigates to various GUIs such as 93-96 the keystrokes, mouse-clicks, graphic objects clicked on, graphic objects in the vicinity of another or a particular graphic object, are recorded and associated. When an operator selects a certain graphic object eg the pump, and for example, right-clicks and selects "Input Query the screen actions up to that point may be compared with previously recorded screen actions. A calculated or estimated data category, eg "Alarms and Events" may be selected as a match or near match for the current keystrokes or mouse clicks. The estimated data category is then selected as a default first data category to be searched when the present Query has been input .

Thus in reference to Figure 11 an operator may operate a computer mouse in the usual way and move a on-screen pointer 92 and select a graphic object 91 representing a process object 90, a powered control valve. The operator may select Al to view the Control Diagram 93 or to view A2 a Process diagram 94. The operator may then select or activate A3 an object of interest on the Process Diagram 94 and see a view of a Faceplate for that process object (the control valve in this case) . A Faceplate 95 is displayed, and the operator may then activate A4 a selection made available because of the Faceplate context to retrieve one or more records to examine for example from Maintenance History 96. A series of actions such as the above exemplary pattern of actions made by the operator is recorded and may be compared with other recorded patterns to find the most likely data category for Queries that occur after, or in connection with, this pattern of selections, object or process views, and graphic control objects selected or manipulated.

Methods of the invention may be supervised, controlled, implemented or otherwise carried out under the control of one or more computer programs. One or more microprocessors (or processors or computers) comprise a central processing unit CPU connected to or comprised in one or more of the above computers or computer systems, which processors or computers perform the steps of the methods according to one or more aspects of the invention, as described for example in reference to the Figures such as for example to Figures 2-6, 8, 9. It is to be understood that the computer programs for carrying out methods according to the invention may also be run on one or more general purpose industrial microprocessors or PLCs or computers instead of one or more specially adapted computers or processors.

The computer program comprises computer program code elements or software code portions that make the computer or processor perform the methods using equations, algorithms, data, stored values, calculations, synchronisations and the like for the methods previously described. A part of the program may be stored in a processor as above, but also in a ROM, RAM, PROM, EPROM or EEPROM chip or similar memory means . The or some of the programs in part or in whole may also be stored in a memory storage device of an industrial control system. The or some of the programs in part or in whole may also be stored locally (or centrally) on, or in, other suitable computer readable medium such as a magnetic disk, CD-ROM or DVD disk, hard disk, magneto- optical memory storage means, in volatile memory, in flash memory, as firmware, or stored on a data server. Other known and suitable media, including removable memory media such as a memory stick or USB memory stick and other removable flash memories, hard drives etc. may also be used. The program may also in part be supplied or updated from a data network, including a public network such as the Internet.

It is also noted that while the above describes exemplifying embodiments of the invention, there are several variations and modifications which may be made to the disclosed solution without departing from the scope of the present invention as defined in the appended claims.

Claims

1. A method for controlling an industrial process in an industrial installation or facility comprising one or more control systems for monitoring and control, wherein certain device or process- related information and other data for each said equipment, plant or process are stored and associated with said control system, said method comprising automatically generating and displaying information in response to a query concerning an apparatus, device or equipment, characterised by: -receiving as input (22, 110, 2) into a computer implemented system a said query expressed in part in a natural language,

-determining a data category (112, 26) to which said query is related based on a part of the natural language comprised in said query, -retrieving (116, 28b) one or more documents dependent on the data category that may be matched to said query,

-ranking (120, 120') and marking the retrieved one or more documents dependent on a confidence rating/interval of an information source, and -displaying (124) on at least one display means (106) one or more documents in an order ranked dependent on the ranking given to each of the one or more documents .

2. A method according to claim 1, further comprising the steps of determining for each said query which of the retrieved one or more documents are examined (126) by an operator entering said query.

3. A method according to claim 2, further comprising the steps of examining and recording the document source for each of the retrieved one or more documents for each said query which were examined by the operator entering said query.

4. A method according to claim 3, further comprising the steps of calculating (120, 120', 119) a confidence interval/rating for a document source for the category or categories identified in said query dependent on which of the retrieved one or more documents were examined by the operator.

5. A method according to claim 1, further comprising the steps of: receiving as input a question (1) expressed in a natural language, resolving the natural language query into parts (24) for processing using a query engine, automatically providing, by means of a natural language query engine, an answer to the original question.

6. A method according to claim 5, further comprising the steps of: resolving, using a natural language query engine means, sub- questions that can be used to determine the answer to the natural language question (1) , processing each sub-question, collecting information from the industrial control system, and compiling and presenting an answer to the original question.

7. A method according to claim 5, further comprising the steps of: identifying one or more domains (26) of the industrial control system in which a data source for the answer to each sub-question may be found, and processing each sub-question, and collecting information from one or more data sources stored in the one or more domains of the industrial control system.

8. A method according to claim 5, further comprising the steps of: resolving the question into parts (24, 24') for processing using a query engine by classifying one or more parts in terms of the type of information each part represents .

9. A method according to claim 5, further comprising the steps of: classifying one or more parts of the question into type of information each part represents dependent on a label or descriptor for a type of information which is stored in a lexicon or other part of the natural language query engine.

10. A method according to claim 1 or 9, wherein the data category comprises any from the group of: current process data, short term history, long term history, batch history, audit log, quality history, maintenance history, alarms and events.

11. A method according to claim 1 or 9, wherein the data category or data category and confidence rating is arranged available for processing a query input by more than one user and/or in more than one installation.

12. A method according to claim 1, further comprising creating a new category and recording a flat confidence rating (119) for the new data category.

13. A method according to claim 12, further comprising the step of assigning (120) the confidence interval rating of a document source per category.

14. A method according to claim 13, further comprising the step of updating (128) the confidence interval rating of a data source per data category.

15. A method according to claim 1 or 3 , further comprising the step of receiving as input (Ilia, 111b, 111c) ranking from the operator of the retrieved one or more documents.

16. A method according to claim 15, further comprising the step of generating a prompt and presenting it to the operator requesting the operator to make a ranking of the retrieved one or more documents .

17. A method according to any previous claim, further comprising the steps of receiving as input one or more selections from a graphic user interface, dividing up the one or more selections into parts and determining the category comprised in said documentation query by comparing the parts of the one or more selections with a classification list and determining a category for said query.

18. A method according to claim 1 or 17, further comprising the step of receiving, as an input, information about a technical context (Ilia) of said query.

19. A method according to claim 18, further comprising the step of generating a prompt and presenting it to the operator requesting the operator to input information representing a technical context such as any from the group of: normal operation product A, product B, product C.

20. A method according to claim 15 or 19, comprising the step of requesting the operator to input information representing a technical context such as any from the group of: abnormal operation product A, product B, product C; process troubleshooting, process configuration I, II, III; process quality troubleshooting product A, B, C; plant maintenance general.

21. A method according to claim 1, further comprising the step of receiving, as an input, information about a technical role (111b) in the industrial installation of the operator who had input said query.

22. A method according to claim 21, further comprising the step wherein information about a technical context (Ilia) or a technical role (111b) is arranged available for processing a query input by more than one user and/or in more than one installation.

23. A method according to any of claims 1-22, further comprising receiving as input one or more operator manipulations or selections on a graphic user interface, dividing up the one or more selections into parts and determining a data category comprised in said documentation query by comparing the parts of the one or more selections with a classification list and determining a data category for said query.

24. A method according to claim 1 or 23, further comprising one or more actions of determining a data category by recognising a pattern of operator manipulations or selections on a graphical user interface preferably by from comparing one or more current operator actions on the graphical user interface with stored patterns of operator manipulations or selections related to one or more data categories of interest.

25. A system for an industrial process in an industrial installation or facility comprising one or more control systems for monitoring and control, wherein certain device or process- related information and other data for each said equipment, plant or process are stored and associated with said control system, said method comprising automatically generating and displaying information in response to a query concerning an apparatus, device or equipment, comprising at least one computer or computer terminal with computer input means connected to said control system, and at least one display means connected to said control system, characterised by an HMI (106) for receiving a query input by a user, a search engine for receiving the query input in a natural language, and a display means arranged to display one or more data items or documents in an order dependent on a ranking of the data items or documents so found.

26. A system according to claim 25, further comprising computer program means to record an active user action to retrieve or open 126) one of more of the automatically displayed documents.

27. A system according to claim 25, further comprising computer program means to record one or more actions input by the user for search purposes and associate those actions with the user by means of the identification that the user is logged-in with.

28. A system according to claim 25, further comprising computer program means to record one or more actions input by the user for each of the retrieved one or more documents for each said query which were examined by the operator entering said query.

29. A system according to claim 25, wherein the system further comprises: a natural language query engine comprising means for resolving the natural language query into parts for processing using the query engine, and means for providing, by means of a natural language query engine an answer to the original question (D .

30. A system according to claim 25 or 29, further comprising means for resolving, using a natural language query engine means, the question into one or more sub-questions or sub-queries that can be used to determine the answer to a natural language query, processing each sub-question, and collecting information from the industrial control system, compiling and presenting an answer to the original query.

31. A system according to claim 29, further comprising means for resolving, using a natural language query engine means, the questions into parts or sub-questions dependent on a label or descriptor for a type of information which is stored in a lexicon or other part of the natural language query engine.

32. A system according to claim 29, further comprising means for matching the context of the question to one or more predetermined operational contexts, such as any from the group of, and not limited to: maintenance, troubleshooting, operations, configuration .

33. A system according to claim 32, wherein the means for matching the operational context further comprises means for generating a question for the user to answer in order to provide information which may be used to match one or more predetermined operational contexts for the question.

34. A system according to claim 25, further comprising computer program means for handling a plurality of data categories, wherein the data category comprises any from the group of: current process data, short term history, long term history, batch history, audit log, quality history, maintenance history, alarms and events.

35. A system according to claim 25 or 34, wherein the system is arranged with means or computer program means for processing a query input such that one or more data categories or data category and confidence ratings are arranged available by more than one user and/or in more than one installation.

36. A system according to claim 25, further comprising computer program means for creating a new category, assigning a confidence interval rating of a document source per category, recording a flat confidence rating for a new data category, and/or updating the confidence interval rating of a data source per data category.

37. A system according to claim 25, further comprising computer program means and input means for receiving as input a ranking from the operator of the retrieved one or πjore data items or documents, and/or generating a prompt and presenting it to the operator requesting the operator to make a ranking of the retrieved one or more data items or documents, and/or a step of receiving, as an input, information about a technical context of said query.

38. A system according to claim 25, further comprising computer program means and graphical user interface means for inferring a technical context such as any from the group of: normal operation product A, product B, product C; or interface means for generating a prompt and requesting the operator to confirm context information or input information about a context.

39. A system according to claim 25, further comprising computer program means for inferring a technical context such as any from the group of: abnormal operation product A, product B, product C; process troubleshooting, process configuration I, II, III; process quality troubleshooting product A, B, C; plant maintenance general, or means for requesting the operator to input information representing a technical context.

40. A system according to any of claims 25-39, comprising computer program means and input means for receiving as input one or more selections from a graphic user interface, dividing up the one or more selections into parts and determining the category comprised in said documentation query by comparing the parts of the one or more selections with a classification list and determining a category for said query.

41. A system according to any of claims 25-38, further comprising computer program means for determining from a plurality of operator navigations or manipulations or selections carried out on a graphical user interface a likely data category for a current query of interest.

42. A system according to claim 25, further comprising at least one memory storage device for storing one or more computer programs for carrying a method of controlling an industrial process in an industrial installation or facility comprising one or more control systems for monitoring and control, wherein certain device or process-related information and other data for each said equipment, plant or process are stored and associated with said control system, said method comprising automatically generating and displaying information in response to a query concerning an apparatus, device or equipment, wherein the method further comprises:

-receiving as input into a computer implemented system a said query expressed in part in a natural language,

-determining a data category to which said query is related based on a part of the natural language comprised in said query, -retrieving one or more documents dependent on the data category that may be matched to said query,

-ranking and marking the retrieved one or more documents dependent on a confidence rating/interval of an information source, and -displaying on at least one display means (6) one or more documents in an order ranked dependent on the ranking given to each of the one or more documents.

43. A system according to claim 25, further comprising at least one computer readable memory storage medium containing one or more of the programs .

44. A computer program for controlling an industrial process in an industrial installation or facility comprising one or more control systems for monitoring and control, wherein certain device or process-related information and other data for each said equipment, plant or process are stored and associated with said control system, said method comprising automatically generating and displaying information in response to a query concerning an apparatus, device or equipment, comprising computer code means and/or software code portions which when run on a computer or processor will make said computer or processor perform the steps of a method according any of claims 1-24.

45. A computer program comprising a computer program according to claim 44 comprised in one or more computer readable media.

46. A method for displaying information generated automatically in response to a query concerning an apparatus, device or equipment in an industrial process or facility comprising one or more control systems for monitoring and control, wherein certain device or process-related information and other data for each said equipment, plant or process are stored and associated with said control system, wherein the method comprises:

-receiving as input into a computer implemented system a said query expressed in part in a natural language, -determining a data category to which said query is related based on a part of the natural language comprised in said query,

-retrieving one or more data items or documents dependent on the data category that may be matched to said query, -ranking and marking the retrieved one or more data items or documents dependent on a confidence rating/interval of an information source, and

-displaying on at least one display means (6) one or more data items or documents in an order ranked dependent on the ranking given to each of the one or more documents.

47. A system for displaying information generated automatically in response to a query concerning an apparatus, device or equipment in an industrial process or facility comprising one or more control systems for monitoring and control, wherein certain device or process-related information and other data for each said equipment, plant or process are stored and associated with said control system, comprising at least one computer or computer terminal with computer input means connected to said control system, and at least one display means connected to said control system, further comprising an HMI (106) for receiving a query input by a user, a search engine for receiving the query input in a natural language, and a display means arranged to display one or more data items or documents in an order dependent on a ranking of the data items or documents so found.