System for proposing solutions to machine failure

Abstract

A system for proposing solutions to machine failure comprises a processor and a memory, which stores a data acquisition unit for acquiring necessary data from a machine and a data analysis unit which serves to identify an association between measurable variable values and problem data, verification method data, and problem-solving data using artificial intelligence, which is trained with machine-related variable data in real-time which are in the forms of general log and trip log, along with the processing of character row data used in machine problem notification. The results obtained are provided in the forms of problem data, verification method data, and problem-solving data, which are sent to a user interface unit to notify and display other necessary data for further diagnosis by an operator.

Description

[0001] SYSTEM FOR PROPOSING SOLUTIONS TO MACHINE FAILURE

[0002] TECHNICAL FIELD

[0003] Petroleum engineering and petroleum exploration and production processes, particularly a system for proposing solutions to machine failure

[0004] BACKGROUND OF THE INVENTION

[0005] An oil drilling station can be considered an industrial building that combines machines with multiple advanced technologies from various fields. The operation of each machine directly impacts the performance, safety, and deterioration of other machines within the station. As a result, each machine is required to be monitored closely by strictly trained personnel to be able to detect even the slightest abnormal operation, interpret it as an operational malfunction, and find appropriate solutions to the problem to allow the machines to operate efficiently at all times and to prevent potential damage to the machines within the entire system.

[0006] Considering the number of machines within a drilling station and the volume of operational variable data of various machines which must be monitored by the operator and the increasing current and historical data which must be continuously analyzed, the operator's verification duties have become more complex and stressful. This leads to the development of an invention for assessing and notifying machine malfunctions in advance to replace or enhance the operational efficiency of the operator to enable them to analyze, propose, and seek methods for preventing machine malfunction in an effective manner. This can be done by creating conditions based directly on verified variables from the machines or by applying artificial intelligence trained to find the association between these variables and the problem data, allowing for flexible detection of errors.

[0007] The said invention can truly reduce the burden on the operator and significantly reduce the malfunction. However, as the invention may malfunction in certain situations, the operator’s fatigue reduces the ability to perform preventive actions or diagnose problems, or the inability to prevent or avoid certain problems, there is a slight possibility that machines within the system may experience operational problems. As a result, it may be necessary to shut down the entire system to address the problem.

[0008] The opportunity cost incurred each time the entire system must be shut down to investigate the cause and perform maintenance on the machines is a significant amount of money. The operator is thus required to investigate, search for, and resolve the problem as quickly as possible to reduce the opportunity cost incurred. What helps the operator to work more conveniently is that each machine collects the trip log and sends it to a user interface to allow the operator to verify all the data, diagnose the cause, and resolve the problem accurately.

[0009] However, due to a wide variety of problems that arise in the industrial areas, rapid and accurate diagnosis can be difficult. Additionally, due to the differences in judgment of individual operators when considering the causes and their relevance, the obtained approaches for investigating and solving the problem may vary and not be standardized, resulting in a delay in communication for problem-solving.

[0010] Therefore, an invention that can reduce the time required to shut down the entire system for diagnosis and problem-solving is highly essential in the industry. Upon reviewing the patent database, some patents with similar objectives were found.

[0011] AU 2020102819 A4 describes a system for diagnosing faults within an offshore station by creating a reasoning model, which is acquired by collecting data on hydraulic operation and electrical power usage in various machines to obtain a model for detecting abnormalities within the system.

[0012] CN 114977492 A describes a power supply monitoring and diagnosing system of a control system in an offshore station. The system operates in conjunction with multiple sensors installed on various machines and areas to randomly inspect power distribution from each source and select alarms suitable for certain situations.

[0013] Additionally, a large number of patents describe systems for notifying the user of operational failures of different systems. However, these patents propose problem identification both in advance and in real time, as well as proposing solutions by taking into account only the variables measured within the machines. As a result, the investigation and problem-solving approaches obtained are straightforward, requiring operators to consider the outcomes from these inventions in connection with the notification data generated by the machines using their own judgment. This leads to unstandardized problem- solving and prolonged discussions which make the system downtime longer than necessary.

[0014] The system for proposing solutions to machine failure according to the present invention focuses on verification of variables related to common problems within machines, such as problems within a fuel system, power transmission system, or lubrication system, together with the processing of notification messages in the form of character row or image, which is one of the display formats commonly transmitted by machines along with the trip log. The system then processes the data related to the problem the machine is experiencing by taking into account the training data input in advance by a system operator and subsequently notifying the operator along with the problem data, recommendations for investigation, and solutions to the problems. The user can access other relevant data, such as maintenance data, user manual, or historical variable data, for further self-diagnosis.

[0015] Such principle of operation provides outcomes that are similar to human decision-making by using work experience, judgment, and data from machines as much as possible. Therefore, the user, who is the operator, receives recommendations for operations that correspond with the standard of a highly experienced operator. As a result, the operator can carry out investigation and solve the problems quickly and in a standardized manner, while communicating effectively with other operators, thereby significantly reducing the system downtime.

[0016] SUMMARY OF THE INVENTION

[0017] A system for proposing solutions to machine failure comprises a processor equipped with artificial intelligence which is trained to learn how to associate various variables with the problem data. The system operates in conjunction with a variable measurement device to determine which problem data the verified variables in the general log and the trip log resemble before issuing a notification and proposing approaches for investigation and problem-solving.

[0018] The invention aims to reduce the investigation burden of system operators, allowing them to focus on monitoring other variables that cannot be calculated or measured by a calculator. Additionally, it assists less experienced operators in performing initial investigation and problem- solving on their own.

[0019] BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Fig. 1 shows an aspect of the functional units within the system for proposing solutions to machine failure.

[0021] Fig. 2 shows an aspect of the data on the verification method and the resolution of problems related to a machine lubrication system, which are displayed to the user.

[0022] DETAILED DESCRIPTION

[0023] The details of the invention will now be described to provide a better understanding, in conjunction with the accompanying drawings. The same reference numbers in the drawings are used to denote identical or similar components throughout the detailed description of the invention.

[0024] Any aspects illustrated herein shall encompass applications to other aspects of the present invention, unless otherwise specified.

[0025] Technical or engineering terms used herein are defined as understood by a person of ordinary skill in the art, unless otherwise specified. The terms “consist of,” “comprise,” “contain,” and “include” are open-ended verbs. For example, “consisting of,” “comprising,” “containing,” “having,” and “including” one or more components are not intended to be limited to the one or more components mentioned but also encompass components that are not specified.

[0026] The system for proposing solutions to machine failure as presented herein is merely an illustration of one construction and embodiment. The system for proposing solutions to machine failure is an invention intended to search for and investigate the problems that occurred within machines, as well as to propose additional problem investigation approaches or problem-solving approaches to operators. It is thus possible to reduce the management burden of system operators, establish operation standard, and help reduce the time required for malfunction investigation and problem-solving.

[0027] According to Fig. 1 , an aspect of the functional units in the system for proposing solutions to machine failure is shown. The system comprises a processor and a memory, which stores at least one of the specification data of all machines in the system to enable a user to easily search for the reference data, the historical maintenance data of each machine, and the data related to the past problems for each machine. Additionally, said memory further stores at least the following functional units:

[0028] A data acquisition unit 100 operating in conjunction with a variable measurement device 110 at various positions to monitor the variable values used to investigate machine malfunction, which can be varied in each machine or each industrial area, such as the machine cycle data, the pressure and temperature of a lubrication oil supply system, the temperature of a lubrication oil tank, the resistance data of a material debris detection sensor, the combustion chamber temperature, the fuel supply rate, or the machine operation cycle. The data acquisition unit 100 receives all the aforementioned data in the form of general log and trip log, which comprises at least verified variable data and character row data. The data are then recorded in the memory and continuously collected as historical data throughout the operation. The variable measurement device 110 is a starting device within a machine with the property and function of verifying the aforementioned variables in order to display the real-time data and the trip log to the user. That is, the variable measurement device 110 measures said variables at predetermined time points, e.g., every five seconds or ten seconds, and report the results to the user as a general log. When the variable measurement device 110 detects a problem within a machine based on its predetermined design, it automatically increases the frequency of measurements. For example, it may change from measuring the variables every five seconds to every thirty milliseconds and continue measuring at that frequency until the machine stops operating in order to generate more detailed variable data than the general log. The device then displays the data to the user along with appropriate character row data in the form of trip log to help the user to quickly identify initial error data that can lead to faster problem diagnosis. The data acquisition unit 100 serves to collect all the data in the memory for future retrieval.

[0029] Said trip log can further comprise at least one of notification time data or problem notification data which can be in the form of image, chart, picture, or audio data which describes the problem that occurred, the alarm name, the problem identity number, or any other data used for communicating with the user to inform them of the occurrence of the problem within the machine, which have been predetermined by the machine manufacturer for each type of error or problem, which may be identical or different for each type of machines.

[0030] The data analysis unit 200 compares the data in the memory, which were collected by the data acquisition unit 100 in the form of at least one of general log, trip log, problem notification data, with the data predetermined by a user in order to identify the association between the occurrence of said data and the problems currently experienced by the machine within the system before summarizing the problems experienced by the machine. Such verification can be performed by verifying only one type of data, comparing different types of data with each other or with defined criteria, or considering more than one type of data collectively in other manners. Additionally, it can be at least one of real-time data consideration or historical data verification within a specified period of time predetermined by the user.

[0031] The data analysis unit 200 is in the form of artificial intelligence trained by receiving the problem data, which at least comprises the type of problem comprising problems related to a fuel supply system and a lubrication system, a name of problem, an urgency of problem categorized as 'monitoring,' 'danger,' and 'requiring immediate resolution', the location within the machine where the problem occurs, and other specific problem data. These data are associated to both general log and trip log input in advance by the user for the training and comprise at least one of the machine cycle data, the pressure and temperature of a lubrication oil supply system, the temperature of a lubrication oil tank, the resistance data of a material debris detection sensor, the combustion chamber temperature, the fuel supply rate, or the machine operation cycle. Such association represents correlations between the aforementioned variable values over a period of time or the data at the time the problem occurs and the prepared problem data. For example, changes in the machine cycle and the pressure of the lubrication oil supply system over the past two months, a detailed trip log of the machine that failed in the past three minutes, and the problem identity data defined by the manufacturer present in the problem notification data sent at the time the problem occurs are associated with the problem of lubricant filter clogging. The lubricant tank temperature data are associated with the resistance data of the material debris detection sensor over a two-week period. The detailed machine pressure data from the past three minutes and the character row data displayed to notify the user to replace the material debris trapping device, which were sent within the past two weeks, are associated with the problem data related to the damage of the material debris trapping device. As a result, the data analysis unit 200 can select problem data that is associated with the data stored in the memory, which was collected by the data acquisition unit 100, in a manner similar to the consideration process of a human operator. This process is based on a verification of the general historical variables together with the detailed trip log data of the machine and the character row data or various notification data received from the machine, using judgment and experience to perform the problem data analysis and to obtain additional diagnostic approaches or resolution.

[0032] In addition to the association between the problem data and various data used in problem diagnosis, the problem data is also associated with the verification method data, which serves as a guideline for verifying other variable data that are not present in the verification by the data acquisition unit 100. This is to provide further suggestions for problem investigation, allowing the user to confirm the evaluation results of the data analysis unit 200, or to perform detailed diagnosis of problems that the data analysis unit 200 has not been specifically trained on, leading to less reliability in the verification. The problem data is further associated with the problemsolving data to provide the user with appropriate problem-solving approaches, either for independent problem- solving or for recommending the assignment of the problem to the appropriate expert. Such recommendation may be immediate problem-solving at that moment or future maintenance approaches, as shown in Figure 2 which illustrates an aspect of the verification method and problem-solving data displayed to the user. According to an example shown herein, a problem inspection plan related to the machine's lubrication system is proposed, along with problem- solving approaches for when the operator can identify the cause of the problem based on said data.

[0033] The operator can prepare the trip log, which further comprises the notification time data or various types of problem notification data associated with the problem data, the verification method data, and problem- solving data, in a manner similar to training using the trip log that contains only verified variable data and character row data associated with the problem data. Such training allows the data analysis unit 200 to expand its operational scope for diagnosing and proposing solutions for a greater variety of machines.

[0034] In one aspect, such verification method and problem- solving data are associated with at least one of the historical maintenance data, the past problem data, and other problem data by creating conditions for the associated data and the data volume such that the data and the data volume are sufficiently determined to exhibit an association with the specified verification method and problem- solving data. In some cases, problems occurring at different points within the same period of time or in the past may require alternative or additional problem-solving methods other than the current problem-solving methods. In other words, the problems may be related to each other, act as a cause of another problem, or lead to a new problem-solving approach.

[0035] The processed problem data exhibiting relevance to the data stored in the memory collected by the data acquisition unit 100 is transmitted to the user interface unit 300 to display the problem data along with the verification and problem- solving methods. The processed problem data may further comprise a confidence value of the results obtained by calculating the similarity of the data recorded by the data acquisition unit 100 in comparison with the data that the data analysis unit 200 has been pre-trained on. That is, if both data are very similar, the confidence value of the result will increase. On the other hand, if the most similar data differs from the data that the data analysis unit 200 was pre-trained on, the confidence value of the result will decrease.

[0036] The user interface unit 300 serves to receive the problem data from the data analysis unit 200 and display such problem data, and possibly the obtained confidence value of the result, to the user through the user interface 310. Additionally, the user interface unit 300 further operates by sending a notification to a notification device 320 as soon as it receives the problem data. The notification may be sent by making a notification sound or activating lights of different colors. Such notification is referenced from the urgency of problem data within the problem data to help the operator to easily categorize the urgency of problem. For example, in the case where the received problem data indicates that the problem requires immediate resolution, the user interface unit 300 will send a signal to the notification device 320 to activate a red light, for example.

[0037] In another aspect, the user interface unit 300 further operates by selecting the machine specification data, the past maintenance data, and the past problem data, which is related to the problem data received from the data analysis unit 200, from memory and displaying them to the user as a reference for further problem diagnosis or problem-solving.

[0038] In practical application, the user inputs the problem data associated with the problem notification data, general log, and trip log of the machine requiring diagnosis and problemsolving before or at the time of the occurrence of the problem or both, along with the verification method and problem-solving data to facilitate problem- solving. This allows artificial intelligence within the data analysis unit 200 to learn the provided data and identify the association among various data. In terms of the association that occurred, the user can enhance the correlation by determining the data and the data volume to exhibit an association with the defined problem data mentioned above.

[0039] Upon completion of the learning process, the user can send an instruction to the data acquisition unit 100 to start receiving data from the variable measurement device 110 and record it in the memory to allow the data analysis unit 200 to perform real-time data verification. When any machine encounters a problem that cannot be prevented or avoided, the variable measurement device 110 automatically adjusts its operation within a determined period of time according to the settings predetermined by the manufacturer or operator in order to increase the frequency of variable verification. The data acquisition unit 100 then sends such verification data to the memory to be recorded, along with problem notification data in the form of trip log together with at least the character row data used to notify the user.

[0040] As soon as the trip log is recorded in the memory, the data analysis unit 200 starts verifying all the data recorded in the memory, including both the data recorded in advance and the data recorded by the data acquisition unit 100, to identify the association between the various data with the associated problem data, as well as to identify the association with the verification method data and resolve problems as trained. The user interface unit 300 serves to display the identified association to the user, along with appropriate notification and the confidence value of the processed results in percentage via the user interface 310 and the notification device 320.

[0041] Any improvements or changes to the invention can be clearly understood and carried out by those skilled in the art, within the scope and spirit of the present invention as indicated in the appended claims.

[0042] BEST MODE OF THE INVENTION

[0043] Best mode of the invention is as described in the detailed description of the invention.

Claims

WHAT IS CLAIMED IS:

1. A system for proposing solutions to machine failure comprises a processor and a memory, which stores at least the following functional units: a data acquisition unit (100) operating in conjunction with a variable measurement device (110) to measure machine -related variables in real time in the forms of general log and trip log and record them in the memory; a data analysis unit (200) which compares the data stored in the memory, which were collected by the data acquisition unit (100), with the data predetermined by a user to identify an association between occurrences of said data and problems currently experienced by the machine within the system before summarizing the problems experienced by the machine and sending the processed results of a problem data to a notification unit; and a user interface unit (300) which is determined to be in a ready state throughout the entire operation of the system and starts operating upon receiving the problem data from the data analysis unit (200) by displaying said problem data to the user via a user interface (310) characterized in that the trip log acquired by the data acquisition unit (100) and recorded in the memory further comprises character row data representing problems that occurred, and the data analysis unit (200) is in a form of artificial intelligence, which is trained by receiving the problem data associated with the general log verified by the variable measurement device (110), the trip log, a verification method data, and a problem-solving data input in advance by the user for the training to identify the association between the data to be input in the future by the data acquisition unit (100) and the problem data, the verification method data, and the problem-solving data recorded.

2. The system for proposing solutions to machine failure according to claim 1, wherein the user interface unit (300) further operates by displaying a confidence value of the processed results obtained from the data analysis unit (200) by calculating a similarity of the data recorded by the data acquisition unit (100) in comparison with the data that the data analysis unit (200) has been pre-trained on.

3. The system for proposing solutions to machine failure according to claim 1, wherein the trip log is in the form of associated variable verification data obtained from the operation of the variable measurement device (110), which has a higher data collection frequency than normal data collection due to an operational adjustment when a machine malfunction is detected.

4. The system for proposing solutions to machine failure according to claim 1, wherein the problem data comprises at least one of a type of problem, name of problem, urgency of problem, and location within the machine where the problem occurs.

5. The system for proposing solutions to machine failure according to claim 4, wherein the type of problem is at least one of a fuel supply system-related problem or a lubrication system-related problem.

6. The system for proposing solutions to machine failure according to claim 1 or 5, wherein the data verified by the variable measurement device (110) and the data used for training the data analysis unit (200) comprise at least one of machine cycle data, pressure and temperature of a lubrication oil supply system, temperature of a lubrication oil tank, resistance data of a material debris detection sensor, combustion chamber temperature, fuel supply rate, or machine operation cycle.

7. The system for proposing solutions to machine failure according to claim 1, wherein the verification method data or the problem- solving data or both are associated with at least one of historical maintenance data, past problem data, and other problem data by creating conditions for the associated data and the data volume such that the data and the data volume are sufficiently determined to exhibit an association with the defined verification method data and problem- solving data.

8. The system for proposing solutions to machine failure according to claim 1, wherein the user interface unit (300) further operates by selecting a machine specification data, the historical maintenance data, and the past problem data associated with the problem data received from the data analysis unit (200) from the memory and displaying them to the user as a reference in further diagnosis or problem- solving.. The system for proposing solutions to machine failure according to claim 1, wherein the trip log further comprises at least one of a notification time data or problem notification data in the form of image, chart, picture, or audio data which describes problem that occurred, alarm name, problem identity number, or any other data used for communicating with the user to inform them of the occurrence of the problem within the machine.

10. The system for proposing solutions to machine failure according to claim 9, wherein the data analysis unit (200) is further trained by associating at least one of the notification time data or the problem notification data with the problem data, the verification method data, and the problem- solving data input in advance by the user for the training in order to identify the association between data to be input in the future by the data acquisition unit (100) and the problem data recorded.

Images