KR20230102431A - Oil gas plant equipment failure prediction and diagnosis system based on artificial intelligence - Google Patents

Abstract

The artificial intelligence-based oil and gas plant facility failure prediction and diagnosis system according to the embodiment performs failure prediction and diagnosis of plant equipment through soundness determination including the presence or absence of defects, uniformity, and appropriateness of mechanical properties of the plant equipment. Through the embodiment, an AI-based health prediction model for key facilities in a demonstration plant is implemented, and the AI prediction model is optimized through virtualization operation and linkage with a plant maintenance simulation system. In addition, simulation for failure prediction and diagnosis and actual plant operation data are established and verified to improve system efficiency and diagnosis accuracy.

Description

AI-based oil and gas plant facility failure prediction and diagnosis system {OIL GAS PLANT EQUIPMENT FAILURE PREDICTION AND DIAGNOSIS SYSTEM BASED ON ARTIFICIAL INTELLIGENCE}

This disclosure relates to a system for predicting and diagnosing failures of oil and gas plant facilities. Specifically, through the application of virtualization and artificial intelligence technology, the operation and maintenance of gas oil plants are advanced, and a diagnostic evaluation model for process risk is established and dynamic risk assessment is performed. It is about an artificial intelligence-based oil and gas plant facility failure prediction and diagnosis system that provides a model.

Unless otherwise indicated herein, material described in this section is not prior art to the claims in this application, and inclusion in this section is not an admission that it is prior art.

Recently, in the oil and gas plant field, project orders are increasing due to the increase in international oil prices and energy demand, and the market size is increasing in the Middle East and around the world. In addition, cutting-edge technologies such as AI (Artificial Intelligence) and ICT are being introduced into plant operation to improve oil and gas plant operation efficiency and safety. In particular, AI-based data analysis prediction model development and virtualization platform technology development, AI-based process efficiency, risk diagnosis evaluation technology and process facility monitoring portal development, AI-based facility safety and health prediction diagnostic evaluation technology development, and AI-based virtualization operation Research on technologies related to platform demonstration technology development is being actively conducted.

In the case of data analysis prediction model development and virtualization platform technology based on artificial intelligence, there is a case of building a sensor-based virtual plant in previous research. However, it was usable to the extent of diagnosing the current state, and it did not reach the level of being used for prognostic diagnosis by combining AI-based prediction models and simulation models using big data.

In addition, AI-based process efficiency, risk diagnosis evaluation technology, and process facility monitoring portal technology exist for commercial products using 3D model weight reduction and simplification technology for virtualization, but the range of support for all of them is limited. In particular, in the case of AI-based facility safety and soundness predictive diagnostic evaluation technology development, despite the efforts of individual domestic plant companies to develop smart solutions, since most technologies are at the original technology stage, conservatism in plant operation is continuously maintained through demonstration. need to break through

On the other hand, dynamic risk assessment technology is a new technology even overseas, and even in areas where dynamic risk assessment should be dealt with, risk assessment for changes during the life cycle / operator's activity (maintenance work) tracking Two types of are presented. Complex data generated from various structures is collected, but methods for processing and analyzing the information are limited so that users can utilize the information efficiently. In addition, the accuracy of diagnosis and prediction and the level of satisfaction received by users have not been empirically evaluated and known, and it is difficult to know in advance how artificial intelligence's automatic diagnosis and prediction methods are performed in various scenario situations, so a simulator is needed. In addition, there is currently no quantified dataset that can measure and compare the stability of plant facilities, and there is a lack of datasets related to prognostic diagnosis.

In Korea, smart plants are recently applied mainly to the petrochemical industry, but service solutions are using products from advanced foreign companies, and the stability and reliability of smart core equipment are relatively low. In order to secure technology, sharing of real-time operation data and application of process equipment are required, but it is difficult to accumulate experience due to business secrets and stable plant operation. In addition, while the size of the domestic engineering market is small, commercialization of related R&D achievements has not been achieved due to intensified dependence on major base solutions such as foreign design and analysis that preoccupied the market.

1. Korean Patent Registration No. 10-1199290 (2012.11.02) 2. Korean Patent Registration No. 10-2145323 (2020.08.11)

The artificial intelligence-based oil and gas plant facility failure prediction and diagnosis system according to the embodiment performs failure prediction and diagnosis of plant equipment through soundness determination including the presence or absence of defects, uniformity, and appropriateness of mechanical properties of the plant equipment.

Through the embodiment, an AI-based health prediction model for key facilities in a demonstration plant is implemented, and the AI prediction model is optimized through virtualization operation and linkage with a plant maintenance simulation system. In addition, simulation for failure prediction and diagnosis and actual plant operation data are established and verified to improve system efficiency and diagnosis accuracy.

An artificial intelligence-based oil and gas plant facility failure prediction and diagnosis system according to an embodiment includes a simulation system construction module for designing and verifying a plant core facility hardware-based simulation system; An evaluation module that uses a simulation system to evaluate the safety and soundness of oil and gas plant facilities; An integrated control module that performs integrated plant control based on artificial intelligence; A facility health judgment module that predicts and diagnoses plant facility health based on artificial intelligence; and a data construction module for constructing data sets related to plant core equipment and external loads; includes

As described above, the artificial intelligence-based oil and gas plant facility failure prediction and diagnosis system establishes a safe soundness inspection system centered on constant monitoring and operation management, constantly exposing related personnel to risks due to the rapid growth of plant maintenance and inspection fields. reduce the possibility In addition, through the implementation of diagnostic modules and platform software, it enables standardization of maintenance and health evaluation platforms.

In addition, through examples, it is possible to demonstrate plant operation and secure performance, and to seek reduction of government budget as well as technology performance verification and experience accumulation. In addition, it secures prognostic diagnosis technology using domestic technology rather than prognostic diagnosis dependent on foreign products, and enables application of the latest technology to the plant market and efficient plant operation support through real-time control monitoring and introduction of AI. In addition, through examples, it is possible to minimize damage to human life, environment, and property in various accident scenarios, and secure technological competitiveness in the plant industry, such as expanding overseas orders, by securing reliability in the field of plant safety. In addition, by securing the source technology for prognosis, it is possible to increase the overall plant and various factories' high added value and create new businesses.

In addition, through examples, health predictive management (Prognostics and Health Management), which early detects system defects by considering both safety and cost in system operation, diagnoses the type and severity of defects that have occurred, and predicts the time of failure in advance. , PHM) to reduce the possibility that equipment failures can lead to large-scale accidents. In addition, by breaking away from the traditional periodic maintenance or preventive maintenance method and realizing condition-based maintenance that performs maintenance only at the necessary time, it is possible to create many economic benefits.

The effects of the present invention are not limited to the above effects, and should be understood to include all effects that can be inferred from the detailed description of the present invention or the configuration of the invention described in the claims.

1 is a diagram showing core technologies for operation and maintenance according to the entire life cycle of a plant
2 is a diagram showing a data processing configuration of an artificial intelligence-based oil and gas plant facility failure prediction and diagnosis system according to an embodiment.
3 is a diagram for explaining a process of implementing an artificial intelligence-based oil and gas plant facility failure prediction and diagnosis system according to an embodiment.
4 is a diagram showing an example of an artificial neural network data sheet for data acquisition;

Advantages and features of the present invention, and methods for achieving them, will become clear with reference to the embodiments described below in detail in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various different forms, and only the present embodiments make the disclosure of the present invention complete, and common knowledge in the art to which the present invention belongs It is provided to fully inform the holder of the scope of the invention, and the present invention is only defined by the scope of the claims. Like reference numerals designate like elements throughout the specification.

In describing the embodiments of the present invention, if it is determined that a detailed description of a known function or configuration may unnecessarily obscure the subject matter of the present invention, the detailed description will be omitted. In addition, terms to be described below are terms defined in consideration of functions in the embodiments of the present invention, which may vary according to the intention or custom of a user or operator. Therefore, the definition should be made based on the contents throughout this specification.

1 is a diagram showing core technologies for operation and maintenance according to the entire life cycle of a plant.

The artificial intelligence-based oil and gas plant facility failure prediction and diagnosis system according to the embodiment provides a plant safety, soundness diagnosis function and predictive diagnosis function, thereby improving the health of the plant facility, including the presence or absence of defects, uniformity, and appropriateness of mechanical properties. Through judgment, failure prediction and diagnosis of plant facilities are performed. In addition, through examples, an AI-based health prediction model for core facilities of a demonstration plant is implemented, and the AI prediction model is optimized through virtualization operation and linkage with a plant maintenance simulation system. In addition, simulation for failure prediction and diagnosis and actual plant operation data are established and verified to improve system efficiency and diagnosis accuracy.

2 is a diagram showing a data processing configuration of an artificial intelligence-based oil and gas plant facility failure prediction and diagnosis system according to an embodiment.

Referring to FIG. 2, the artificial intelligence-based oil and gas plant facility failure prediction and diagnosis system according to the embodiment includes a simulation system construction module 100, an evaluation module 200, an integrated control module 300, a facility soundness judgment module ( 400) and a data building module 500. The term 'module' used in this specification should be interpreted as including software, hardware, or a combination thereof, depending on the context in which the term is used. For example, the software may be machine language, firmware, embedded code, and application software. As another example, the hardware may be a circuit, processor, computer, integrated circuit, integrated circuit core, sensor, micro-electro-mechanical system (MEMS), passive device, or combination thereof.

The simulation system building module 100 builds a simulation system for designing and verifying a plant core facility hardware-based simulation system. In an embodiment, the simulation system construction module 100 may design a core facility hardware-based simulation system operation scenario and verify the simulation system through the scenario.

The evaluation module 200 utilizes a simulation system to evaluate the safety and soundness of oil and gas plant facilities. In the embodiment, the evaluation module 200 performs soundness determination including the presence or absence of defects, uniformity, and appropriateness of mechanical properties of plant equipment through a simulation system. In the embodiment, the evaluation module 200 may use the vibration diagnosis data of plant equipment for safety, soundness evaluation and safety precision diagnosis. In the embodiment, vibration diagnosis data is analyzed to evaluate plant safety and soundness. In the embodiment, vibration diagnosis data is analyzed through time domain analysis, frequency domain analysis, and component analysis, and safety and soundness evaluation results are calculated according to the analysis results.

In the embodiment, the time domain analysis of the evaluation module 200 includes a process of analyzing the amplitude or temporal variation of the amplitude, the impulse and symmetry of the waveform, etc. using a time waveform, a probability density function of the amplitude, and the like. In the embodiment, the frequency domain analysis of the evaluation module 200 includes a process of identifying which frequency components (spectrum) are included in the vibration using FFT (Fast Fourier Transform) or the like. Spatial domain analysis uses the Lissajous figure and the like to figure out what kind of trajectory the center of rotation axis (vibration) moves in space.

In addition, in the embodiment, the evaluation module 200 may perform safety and soundness evaluation using a method using a primary effect parameter representing the performance or function of a facility and a secondary effect parameter generated by the operation of the facility. The primary effect parameter is a state parameter observed when the equipment performs its original purpose, and in the embodiment, the state of the equipment is monitored using the primary effect parameter. In an embodiment, performance monitoring and behavior monitoring may be performed.

Secondary effect parameters are additional state parameters such as vibration, sound, temperature, etc., which change according to the operation of the facility. In the embodiment, parameter data is collected through a sensor for parameter diagnosis, and the collected parameter data is analyzed to enable early detection of plant anomalies and accurate identification of causes or occurrence locations of anomalies. In the embodiment, the primary effect parameter is mainly used for simple diagnosis, and the secondary effect parameter can be used for both simple diagnosis and precise diagnosis.

The integrated control module 300 performs integrated plant control based on artificial intelligence. In an embodiment, the integrated control module 300 monitors plant facility data in conjunction with a virtualization operation and maintenance platform, and responds to changing monitoring data characteristics through an adaptive machine learning model. In addition, the integrated control module 300 is an integrated control module; silver

You can analyze the control and simulation environment using IoT.

In the embodiment, the integrated control module 300 considers both risk and cost in the operation of the system, early detects system faults, and diagnoses the type and severity of faults that have occurred. (diagnosis) and performs prognostics and health management (PHM) that predicts the time of failure in advance. Condition Based Maintenance (CBM), which performs maintenance only when necessary, breaking away from the traditional periodic or preventive maintenance (PM) method, while reducing the possibility of large-scale accidents through predictive health management ) is realized. The integrated control module 300 according to the embodiment processes data obtained through sensors installed in the system (Signal processing), extracts characteristic factors to represent the soundness of the current system from it (Feature extraction), and It performs fault diagnosis to identify the cause of failure and its severity, and failure prognosis to predict the time of future failure. The purpose of signal processing of the integrated control module 300 according to the embodiment is to remove other information while leaving information necessary for fault diagnosis for a signal mixed with various other signals and noise. To this end, in the embodiment, a process of discrete signal separation to remove periodic signals other than the faulty signal and signal enhancement to enhance the faulty signal having a minute size are performed. In the embodiment, features are extracted from data subjected to specific signal separation and fault signal enhancement processing so that defect-related information can be obtained more easily. In addition, in the embodiment, defect-related information may be extracted through a signal decomposition process in which an original signal is decomposed into various components without going through the above two steps.

According to the embodiment, among the decomposed components, there is data by which a defect signal is clearly identified, and thus defect characteristics can be extracted through the corresponding data. In addition, in the embodiment, linear prediction, self-adaptive noise cancellation (SANC), and time synchronous averaging (TSA) processes are performed to remove signals unrelated to the faulty signal. is available. Linear prediction is a method of estimating a future output value by linearly modeling a deterministic part of a signal using past data. Self-adaptive noise cancellation is a filter that uses a primary signal and a reference signal, and a signal that optimizes the reference signal to be similar to a specific signal inherent in the primary signal. processing method. In the embodiment, the primary input is an original signal in which noise and signals generated by facilities are mixed, and the reference input is a signal generated only by plant facilities. In the embodiment, filter parameters are calculated to obtain an error (e) obtained by removing noise signals in the basic input through an optimization process, so that only desired plant signals can be selected. Time Synchronous Averaging is a method used to remove noise inherent in signals and leave only signals with periodic characteristics. In the embodiment, the vibration signal of the plant equipment is resampled according to the rotation period of the currently rotating shaft, and then the average value thereof is calculated. Through this, random noise components irrelevant to the fault signal are removed, and only signal components having periodic characteristics according to rotation remain. To this end, the embodiment may generally use a tachometer that generates a pulse signal at every rotation to measure the rotation period of the shaft or an optical encoder that measures the angle of the shaft.

The equipment soundness judgment module 400 predicts and diagnoses the health of plant equipment based on artificial intelligence. In an embodiment, soundness may include defects, uniformity, appropriateness of mechanical properties, and the like of plant equipment. In the embodiment, the facility soundness determination module 400 may detect abnormality by measuring signal data such as vibration or noise, and various signal processing such as frequency analysis performs data preprocessing. In an embodiment, the facility health determination module 400 may determine an abnormal state of the facility by analyzing vibration data. In the embodiment, a method of monitoring the presence or absence of an abnormal state by measuring a Mahalanobis distance based on steady state data is used. In the embodiment, the Mahalanobis distance is measured for a vector composed of statistical indicators extracted from the signal data for each section, and it is determined whether or not the corresponding section is in an abnormal state. The Mahalanobis distance is a distance scale often used in cluster analysis. When there is a correlation between variables in multivariate data, the distance is measured through data standardization using a covariance matrix. In addition, in the embodiment, the Mahalanobis distance classifier constructs a Mahalanobis space in advance from training data, measures a distance to the space for arbitrary data, and determines whether or not it is the same class as the training data according to the distance. In addition, in the embodiment, the performance of detecting abnormal state of plant equipment can be improved through pre-processing of signal data. In an embodiment, the preprocessing process may include standardization, Hamming window, cepstrum analysis, and statistical analysis.

In the embodiment, in the preprocessing step, a Hamming window is applied to eliminate leakage errors generated in the signal data collection process, and a cepstrum is applied to separate formants, which are original signals of signal data. ) to perform the analysis. Hamming window and cepstrum analysis are performed to remove discontinuous quantities, that is, leakage, generated in the process of arbitrarily sectioning continuous data. When the Hamming window according to the embodiment is applied, when the length of the extracted signal data is exactly an integer multiple of the signal period, an accurate frequency domain spectrum can be obtained, but otherwise, an error occurs due to discontinuity at both ends. In the embodiment, in order to remove this, a more accurate spectrum can be obtained from the FFT result by applying a window function that converts the signal data to a value close to 0 from the center to both ends. In the embodiment, a Hamming window is applied through Equation 1 in order to minimize such a data leakage effect. In Equation 1, N represents the length of the window.

Equation 1

Cepstrum analysis is a method of performing inverse FFT on a logarithmic scale of a signal expressed in a frequency domain by applying fast Fourier transformation (FFT). On the time axis, the response to a general shock signal appears as a signal in a complex form by convolution of the input signal and the transfer function. At this time, if cepstrum analysis is applied, the waveform of the input signal and the transfer function that implements the response can be separated. can In an embodiment, by applying a cepstrum, a logarithmic function formed by multiplying two functions is separated into a sum, and thus the magnitude and phase of a signal converted to a frequency domain can be separated.

The data construction module 500 builds a data set related to plant core equipment and external loads. In the embodiment, the data construction module 500 responds to changing data characteristics through an adaptive machine learning model, builds a system integration data set after determining the final artificial intelligence model, and sets the data set for application to oil/gas plant demand companies to build In addition, in the embodiment, the data building module 500 compares software values and dataset data through monitoring data measurement in oil and gas plant facilities, and builds basic data for analyzing the measured data and existing data.

3 is a diagram for explaining a process of implementing an artificial intelligence-based oil and gas plant facility failure prediction and diagnosis system according to an embodiment.

Referring to FIG. 3, in the embodiment, a core facility hardware-based simulation system, a core facility hardware-based simulation system operating scenario, a facility integrated control platform and a simulation system using IoT sensors are designed. In addition, an integrated facility control platform and safety health evaluation simulation system will be implemented, and the simulation system will be advanced through the development of AI-based health prediction and diagnosis technology. To this end, sensor data is defined and collected to identify plant conditions, and an AI-based health prediction model for key facilities in demonstration plants is developed. In addition, a prognostic diagnosis inspection data set is established to build a system integrity prognostic diagnosis system based on the hardware of core facilities of the demonstration plant.

In addition, for the design of a hardware-based simulation system for core facilities of a demonstration plant, scenarios for operating a simulation system based on hardware for core facilities of a demonstration plant are developed, and an integrated control platform and simulation system for facilities using IoT sensors are designed.

4 is a diagram showing an example of an artificial neural network data sheet for data acquisition.

The artificial intelligence-based oil and gas plant facility failure prediction and diagnosis system according to the embodiment includes artificial neural networks, linear discriminant analysis, principal component analysis, and Mahalanobis distance classifier. It is possible to predict and diagnose plant abnormalities by collecting and analyzing data using various data mining analysis techniques such as distance classifier) and logistic regression.

Through the artificial intelligence-based oil and gas plant facility failure prediction and diagnosis system according to the embodiment, the existing virtualization smart plant technology is developed one step further to advance the platform construction technology that integrates all prediction and simulation models and existing distributed systems. In addition, anomaly detection through artificial intelligence can improve individual plant-customized process efficiency and provide a framework and platform that can be extended to various processes.

In addition, it is possible to secure technological competitiveness compared to advanced companies by improving the practicality of the existing plant safety training education system and securing new application technology for safety training methods for large plant facilities. In addition, it is possible to establish a technology development system leading to individual sensor-data center-plant management due to the development of plant-specific maintenance and predictive diagnosis techniques. In particular, secure prognostic diagnosis technology using domestic technology rather than prognostic diagnosis dependent on foreign products through plant facility safety and health prognosis diagnosis and evaluation technology, and apply the latest technology to the plant market and support efficient plant operation through real-time control monitoring and AI introduction this is possible In addition, through examples, it is possible to minimize damage to human life, environment, and property in various accident scenarios, and secure technological competitiveness in the plant industry, such as expanding overseas orders, by securing reliability in the field of plant safety. In addition, by securing the original technology for prognostic diagnosis through safety and soundness prognostic diagnosis technology, it is possible to create high added value and new business in the overall plant and various factories.

The disclosed content is only an example, and can be variously modified and implemented by those skilled in the art without departing from the subject matter of the claim claimed in the claims, so the protection scope of the disclosed content is limited to the specific It is not limited to the examples.

Claims (7)

In the AI-based oil and gas plant facility failure prediction and diagnosis system,
A simulation system construction module that designs and verifies a plant core facility hardware-based simulation system;
An evaluation module that executes safety and health evaluation of oil and gas plant facilities using the simulation system;
An integrated control module that performs integrated control of plant facilities based on artificial intelligence;
A facility health judgment module that predicts and diagnoses plant facility health based on artificial intelligence; and
A data building module that builds data sets related to plant core equipment and external loads; Artificial intelligence-based oil and gas plant facility failure prediction and diagnosis system including
According to claim 1, The simulation system construction module; silver
An artificial intelligence-based oil and gas plant facility failure prediction and diagnosis system, characterized by designing core facility hardware-based simulation system operation scenarios and verifying the simulation system through the scenarios.
According to claim 1, wherein the evaluation module; silver
An artificial intelligence-based oil and gas plant facility failure prediction and diagnosis system characterized by performing soundness judgments including the presence of defects, uniformity, and appropriateness of mechanical properties of plant facilities through a simulation system.
According to claim 1, The integrated control module; silver
An artificial intelligence-based oil and gas plant facility failure prediction and diagnosis system characterized by monitoring plant facility data in conjunction with a virtualization operation and maintenance platform and responding to changing monitoring data characteristics through an adaptive machine learning model. .
According to claim 4, The integrated control module; silver
An artificial intelligence-based oil and gas plant facility failure prediction and diagnosis system characterized by analyzing control and simulation environments using IoT.
The method of claim 1, wherein the data building module; silver
AI-based oil and gas that responds to changing data characteristics through an adaptive machine learning model, builds a system integration data set after determining the final artificial intelligence model, and builds a data set for application to plant demand companies Plant facility failure prediction and diagnosis system.
The method of claim 6, wherein the data building module; silver
An artificial intelligence-based oil and gas plant facility failure prediction and diagnosis system, characterized in that it compares data values and data set data monitored in oil and gas plant facilities and establishes basic data for analyzing monitored data and existing data.

Images