CN115688581A - Oil gas gathering and transportation station equipment parameter early warning method, system, electronic equipment and medium - Google Patents
Oil gas gathering and transportation station equipment parameter early warning method, system, electronic equipment and medium Download PDFInfo
- Publication number
- CN115688581A CN115688581A CN202211359817.9A CN202211359817A CN115688581A CN 115688581 A CN115688581 A CN 115688581A CN 202211359817 A CN202211359817 A CN 202211359817A CN 115688581 A CN115688581 A CN 115688581A
- Authority
- CN
- China
- Prior art keywords
- equipment
- parameter
- parameters
- oil
- lstm
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention relates to an oil and gas gathering and transportation station equipment parameter early warning method, a system, electronic equipment and a medium, and belongs to the field of oil and gas gathering and transportation station safety early warning. The method comprises the steps of calculating the correlation between each equipment operation parameter and other equipment operation parameters in a process data set by utilizing a spearman rank correlation coefficient, and establishing a related equipment parameter set; pre-training the LSTM model according to a relevant equipment parameter set, and determining a network hyper-parameter by using a genetic algorithm; storing the relevant equipment parameter set and the network hyper-parameter into a general database; and in the on-line prediction stage, relevant equipment parameter sets and network hyper-parameters of the operation parameters of the equipment to be predicted are obtained from the general database and are substituted into the LSTM model, the LSTM pre-training model is trained by adopting historical data, a predicted value is obtained through the trained LSTM prediction model, and safety early warning is carried out according to the comparison result of the predicted value and the corresponding threshold value. The method can improve the accuracy of safety early warning in the oil gas gathering and transportation process and the universality of the early warning method.
Description
Technical Field
The invention relates to the technical field of safety early warning of oil and gas gathering and transportation stations, in particular to a method, a system, electronic equipment and a medium for early warning of equipment parameters of an oil and gas gathering and transportation station.
Background
Along with the rapid development of Chinese economy, the urbanization process is gradually accelerated, the resource demand is gradually increased, and the oil gas is taken as an important traditional energy source and has great significance to national economy. The transfer and gathering of oil and gas are one of the core requirements of an oil and gas gathering and transferring combined station (oil and gas gathering and transferring station for short), and there are many risk factors that continuously affect the working efficiency and safety of the oil and gas gathering and transferring combined station. As a key part of an oil-gas gathering and transportation combined station, the safety of an oil field gathering and transportation system directly concerns the safety and efficiency of oil-gas gathering and transportation, wherein the work including crude oil dehydration, oil-gas separation and the like is mainly carried out, pressure vessels and thermal equipment in the system are centralized, and once safety accidents such as fire disasters occur under the influence of high temperature, high pressure, environmental factors and the like, the whole oil-gas gathering and transportation combined station has immeasurable loss.
With the development of industrial process information automation, parameters in the oil and gas gathering and transportation process are more and more. How to ensure the safe and stable operation of the gathering and transportation process by utilizing the working condition information brought by the operation variables of the gathering and transportation process becomes an urgent problem to be solved. Therefore, for the quality and safety of oil and gas gathering and transportation, risk accident prejudgment is reasonably carried out, and the safety monitoring early warning and emergency management work of the oil and gas gathering and transportation system are enhanced, so that the method has great significance for production practice. In order to realize the safety prediction of oil and gas gathering and transportation and the identification of abnormal working conditions, an oil and gas gathering and transportation safety prediction method needs to be researched.
In recent years, the gathering and transportation process safety early warning technology has also been developed to a certain extent in China. Aiming at the safety problem of the gathering and transportation process, a safety evaluation mode is mostly adopted at present, domestic research on the gathering and transportation system focuses on risk safety management and pipeline corrosion analysis of the system, the safety evaluation is to achieve the purpose of system safety, the risks existing in the process are judged, calculated and analyzed through a series of safety specifications and evaluation methods, and protective measures are adopted according to the types and the characteristics of the risks, so that the occurrence of accidents is avoided or the possibility of the occurrence of the accidents is reduced. Most of the methods are based on the sensing system or artificial experience analysis and operation management, and the supervision and investigation on various hidden dangers are enhanced in the aspects of safety responsibility, equipment management and maintenance, fire control management level and the like. However, the safety early warning method is greatly influenced by safety evaluation rules and human experience, cannot ensure the accuracy of a prediction result, and cannot be applied to the safety early warning of all equipment of the oil and gas gathering and transportation station.
Disclosure of Invention
The invention aims to provide a method, a system, electronic equipment and a medium for early warning of equipment parameters of an oil and gas gathering and transportation station, which can improve the accuracy of safety early warning in the oil and gas gathering and transportation process and the universality of an early warning method.
In order to achieve the purpose, the invention provides the following scheme:
on one hand, the invention provides an oil and gas gathering and transportation station equipment parameter early warning method, which comprises the following steps:
acquiring historical data of equipment operation parameters of all equipment in the oil and gas gathering and transportation station, and preprocessing the historical data to form a process data set; all the equipment in the oil-gas gathering and transportation station comprises a three-phase separator, an external transportation pump, an oil storage tank and a heating furnace; the equipment operation parameters comprise three-phase separator pressure, three-phase separator liquid level, external conveying pump pressure, oil storage tank liquid level and heating furnace temperature;
for each equipment operation parameter in the process data set, calculating the correlation between each equipment operation parameter and all other equipment operation parameters by using the spearman rank correlation coefficient, and establishing a relevant equipment parameter set of each equipment operation parameter;
pre-training the LSTM model according to the relevant equipment parameter set of each equipment operation parameter, selecting and optimizing the LSTM model hyper-parameters by using a genetic algorithm, and determining network hyper-parameters;
storing the relevant equipment parameter set of each equipment operation parameter and the corresponding network hyper-parameter into a database to obtain a general database;
in the online prediction stage, a relevant device parameter set of the device operation parameters to be predicted and a corresponding network hyper-parameter are obtained from the general database and substituted into the LSTM model to obtain an LSTM pre-training model;
training the LSTM pre-training model by adopting the historical data of the operation parameters of the equipment to be predicted and the operation parameters of the equipment in the relevant equipment parameter set, and obtaining the LSTM prediction model of the operation parameters of the equipment to be predicted after the training is finished;
and obtaining a predicted value of the operation parameter of the equipment to be predicted by adopting the LSTM prediction model, and carrying out safety early warning according to a comparison result of the predicted value and the corresponding equipment operation parameter threshold value.
Optionally, the obtaining and preprocessing device operation parameter historical data of all devices in the oil and gas gathering and transportation station to form a process data set specifically includes:
acquiring historical data of equipment operation parameters of all equipment in the oil and gas gathering and transportation station, and sequencing the historical data according to a time sequence to obtain a time sequence data set of the operation parameters of all the equipment;
and identifying and processing abnormal values in the time sequence data set by adopting a Lauda criterion method and a Showville method to obtain a process data set of all equipment operation parameters.
Optionally, for each device operation parameter in the process data set, calculating a correlation between each device operation parameter and all other device operation parameters by using a spearman rank correlation coefficient, and establishing a related device parameter set for each device operation parameter, specifically including:
for each equipment operation parameter in the process data set, calculating the correlation between each equipment operation parameter and all other equipment operation parameters by using a spearman rank correlation coefficient, and determining that the two equipment operation parameters are closely related when the absolute value of the spearman rank correlation coefficient is greater than 0.8;
and finding out other equipment operation parameters closely related to each equipment operation parameter, and establishing a related equipment parameter set of each equipment operation parameter.
Optionally, the pre-training of the LSTM model according to the relevant device parameter set of each device operating parameter, the selection and optimization of the LSTM model hyper-parameters by using a genetic algorithm, and the determination of the network hyper-parameters specifically include:
taking a relevant equipment parameter set of each equipment operation parameter as an input variable of the LSTM model, taking each equipment operation parameter as an output variable, and training the LSTM model;
in the LSTM model training process, selecting and optimizing the LSTM model hyper-parameters by using a genetic algorithm to determine network hyper-parameters; the network hyper-parameters comprise step frequency, the number of network nodes, an activation function, an optimizer and a learning rate.
Optionally, before the obtaining the predicted value of the to-be-predicted device operating parameter by using the LSTM prediction model and performing early warning according to the comparison result between the predicted value and the corresponding device operating parameter threshold, the method further includes:
calculating T corresponding to all equipment operating parameters by adopting a principal component analysis method 2 And taking the statistical threshold as an equipment operation parameter threshold.
On the other hand, the invention also provides an oil and gas gathering and transportation station equipment parameter early warning system, which comprises:
the process data set construction module is used for acquiring and preprocessing the historical data of the equipment operation parameters of all equipment in the oil and gas gathering and transportation station to form a process data set; all the equipment in the oil-gas gathering and transportation station comprises a three-phase separator, an external transportation pump, an oil storage tank and a heating furnace; the equipment operation parameters comprise three-phase separator pressure, three-phase separator liquid level, external conveying pump pressure, oil storage tank liquid level and heating furnace temperature;
a related device parameter set construction module, configured to calculate, for each device operating parameter in the process data set, a correlation between each device operating parameter and all other device operating parameters using a spearman rank correlation coefficient, and establish a related device parameter set for each device operating parameter;
the LSTM model pre-training module is used for pre-training the LSTM model according to the relevant equipment parameter set of each equipment operation parameter, selecting and optimizing the LSTM model hyper-parameters by using a genetic algorithm, and determining network hyper-parameters;
the general data storage module is used for storing the relevant equipment parameter set of each equipment operation parameter and the corresponding network hyper-parameter into a database to obtain a general database;
the pre-training model building module is used for acquiring a relevant equipment parameter set of the operation parameters of the equipment to be predicted and corresponding network hyper-parameters from the general database in an online prediction stage, and substituting the relevant equipment parameter set and the corresponding network hyper-parameters into the LSTM model to obtain the LSTM pre-training model;
the LSTM model training module is used for training the LSTM pre-training model by adopting historical data of the operation parameters of the equipment to be predicted and the operation parameters of the equipment in the relevant equipment parameter set, and obtaining the LSTM prediction model of the operation parameters of the equipment to be predicted after training is finished;
and the equipment parameter prediction early warning module is used for obtaining a predicted value of the operation parameter of the equipment to be predicted by adopting the LSTM prediction model and carrying out safety early warning according to a comparison result of the predicted value and a corresponding equipment operation parameter threshold value.
In another aspect, the present invention further provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method for warning the device parameter of the oil and gas gathering and transportation station when executing the computer program.
In another aspect, the present invention further provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed, implements the method for hydrocarbon gathering station equipment parameter early warning.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
the invention provides a parameter early warning method, a parameter early warning system, electronic equipment and a medium for oil and gas gathering and transportation station equipment, wherein the method comprises the following steps: acquiring historical data of equipment operation parameters of all equipment in the oil and gas gathering and transportation station, and preprocessing the historical data to form a process data set; for each equipment operation parameter in the process data set, calculating the correlation between each equipment operation parameter and all other equipment operation parameters by using a spearman rank correlation coefficient, and establishing a relevant equipment parameter set of each equipment operation parameter; pre-training the LSTM model according to the relevant equipment parameter set of each equipment operation parameter, selecting and optimizing the LSTM model hyper-parameters by using a genetic algorithm, and determining network hyper-parameters; storing the relevant equipment parameter set of each equipment operation parameter and the corresponding network hyper-parameter into a database to obtain a general database; in the online prediction stage, a relevant device parameter set of the device operation parameters to be predicted and a corresponding network hyper-parameter are obtained from the general database and are substituted into the LSTM model to obtain an LSTM pre-training model; training the LSTM pre-training model by adopting historical data of the operation parameters of the equipment to be predicted and the operation parameters of the equipment in the relevant equipment parameter set, and obtaining the LSTM prediction model of the operation parameters of the equipment to be predicted after training; and obtaining a predicted value of the operation parameter of the equipment to be predicted by adopting the LSTM prediction model, and carrying out safety early warning according to a comparison result of the predicted value and the corresponding equipment operation parameter threshold value. The method can improve the accuracy of safety early warning in the oil gas gathering and transportation process and the universality of the early warning method.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required in the embodiments will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flow chart of an oil and gas gathering station equipment parameter early warning method provided by an embodiment of the invention;
FIG. 2 is a route diagram of an oil and gas gathering and transportation station equipment parameter early warning technology provided by the embodiment of the invention;
FIG. 3 is a schematic diagram of a basic structure of an LSTM model provided in an embodiment of the present invention;
fig. 4 is a schematic diagram of an LSTM model building and training process provided by the embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention aims to provide a method, a system, electronic equipment and a medium for early warning of equipment parameters of an oil and gas gathering and transportation station, which can improve the accuracy of safety early warning in the oil and gas gathering and transportation process and the universality of an early warning method.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
Fig. 1 is a flowchart of an oil and gas gathering and transportation station equipment parameter early warning method provided by an embodiment of the invention, and fig. 2 is a route diagram of an oil and gas gathering and transportation station equipment parameter early warning technology provided by an embodiment of the invention. Referring to fig. 1 and 2, the method for early warning the equipment parameter of the oil and gas gathering and transportation station comprises the following steps:
step 101: and acquiring historical data of equipment operation parameters of all equipment in the oil and gas gathering and transportation station, and preprocessing the historical data to form a process data set.
The method comprises the steps of collecting historical data of equipment operation parameters of all equipment from an oil-gas gathering and transportation combined station through a DCS (Distributed Control System), wherein the equipment in the oil-gas gathering and transportation station generally comprises equipment such as a three-phase separator, an outward transportation pump, an oil storage tank and a heating furnace, and the corresponding equipment operation parameters comprise parameters such as pressure of the three-phase separator, liquid level of the three-phase separator, pressure of the outward transportation pump, liquid level of the oil storage tank and temperature of the heating furnace. Specifically, the operation parameters of the equipment such as the oil inlet pressure of a # 1 water separator, the oil-water interface of a # 1 water separator, the air outlet pressure of a # 1 water separator, the oil inlet pressure of a # 2 water separator, the oil-water interface of a # 2 water separator, the oil level of a # 2 heating furnace, the water jacket temperature of a # 2 heating furnace, the inlet temperature of a # 3 heating furnace, the liquid level of a # 1 settling tank, the liquid level of a # 3 settling tank, the liquid level of a # 8 settling tank, the inlet temperature of an external feeding pump, the oil output pressure, the oil output flow, the liquid level of a # 8 light oil tank, the outlet temperature of a # 1 air compressor, the oil output temperature and the like can be further divided into the operation parameters.
The above history data is a data set of the device operation parameter values of all the devices acquired from the DCS system, and since the data to be input in the LSTM model is time series data, the history data needs to be processed and converted into structured data aligned in time series with time series. Specifically, the acquired historical data of the equipment operation parameters of all the equipment in the oil and gas gathering and transportation station are sorted according to the time sequence, and a time sequence data set of the operation parameters of all the equipment is obtained.
The data preprocessing is a process of performing change processing on historical data directly obtained from the DCS system to obtain data that can be directly used in the LSTM model, and therefore, the preprocessing process further includes: and identifying and processing abnormal values in the time sequence data set by adopting a Lauda criterion method and a Showville method to obtain a process data set of all equipment operation parameters.
Step 102: for each equipment operating parameter in the process data set, calculating the correlation between each equipment operating parameter and all other equipment operating parameters by using the spearman rank correlation coefficient, and establishing a relevant equipment parameter set of each equipment operating parameter.
For each equipment operating parameter in the process data set (which may be referred to as a target equipment operating parameter), calculating the correlation between each target equipment operating parameter and all other equipment operating parameters by using a Spearman rank correlation coefficient, finding out other equipment operating parameters closely related to each target equipment operating parameter, constructing a related equipment parameter set of each target equipment operating parameter, and determining whether each equipment operating parameter in the related equipment parameter set has a positive or negative influence on the corresponding target equipment operating parameter. The spearman rank correlation coefficient can describe the direction and degree of the change trend among variable data, the value range is from-1 to 1, positive values represent positive correlation, negative values represent negative correlation, and the larger the value is, the stronger the correlation is. If Y tends to increase as X increases, the Spireman correlation coefficient is positive; if Y tends to decrease as X increases, the Spireman correlation coefficient is negative.
In the invention, when the absolute value of the correlation coefficient of the spearman rank is greater than 0.8, the close correlation between the operation parameters of the two devices is determined; therefore, other equipment operation parameters closely related to each equipment operation parameter can be found out according to the value of the spearman rank correlation coefficient, closely related variables selected by the spearman rank correlation coefficient are used as input features of the LSTM model for training, and model prediction accuracy can be improved. The spearman rank correlation coefficient can also provide support for constructing a symbol directed graph, and the SDG symbol directed graph reveals the influence of each equipment node on adjacent variables and reveals the propagation path of the predicted abnormal working condition in the future as much as possible.
The related equipment parameter set selected by the invention is the characteristic variables of each equipment operation parameter needing to be predicted, the characteristic variables can influence the value of the predicted parameter, the value is used as the input of the LSTM neural network, and then the required variable value is predicted. The related device parameter set is constructed by putting selected characteristic variables (other device operation parameters) closely related to the target device operation parameters into the parameter set according to the Spearman rank correlation coefficient. The set of relevant device parameters thus obtained is the set of characteristic variables selected on the basis of the Spearman rank correlation coefficient.
Step 103: and pre-training the LSTM model according to the relevant equipment parameter set of each equipment operation parameter, and selecting and optimizing the super-parameters of the LSTM model by using a genetic algorithm to determine the network super-parameters.
Fig. 3 is a schematic diagram of a basic structure of an LSTM model according to an embodiment of the present invention. X in FIG. 3 t The input unit of the LSTM model is used for inputting parameter data of the LSTM model at each time t. h is t Is the output of the hidden node to control the degree to which the state of the previous cell was forgotten. σ represents a sigmoid function, and has a value range of (0,1), and can map a real number to an interval of (0,1). tanh represents a hyperbolic tangent function. A represents the LSTM model.
The invention takes the relevant device parameter set selected aiming at each target device operation parameter as the input variable of an LSTM (Long Short-term memory network) model, takes each target device operation parameter as the output variable (also called predictive variable), and takes the historical data set as a model with the following ratio of 8:2, dividing the ratio into a training set and a testing set, and inputting the training set and the testing set into an LSTM model for training.
In the LSTM model training process, the genetic algorithm is used for selecting and optimizing the LSTM model hyper-parameters, and the network hyper-parameters are determined through verification and inspection; the network hyper-parameters comprise step frequency, the number of network nodes, an activation function, an optimizer and a learning rate.
Step 104: and storing the relevant equipment parameter set of each equipment operation parameter and the corresponding network hyper-parameter into a database to obtain a general database.
The invention stores the relevant device parameter set of each target device operation parameter and the corresponding network hyper-parameter value into the database by designing the relational database, generalizes the whole process and establishes the general database.
Step 105: and in the online prediction stage, a relevant device parameter set of the device operation parameters to be predicted and the corresponding network hyper-parameters are obtained from the general database and substituted into the LSTM model to obtain the LSTM pre-training model.
Fig. 4 is a schematic diagram of an LSTM model building and training process provided by the embodiment of the present invention. The process data set of fig. 4 is the values of the operating parameters of the equipment at each time for each piece of equipment stored in the general purpose database. T, P and Y represent input variables such as temperature and pressure and output variables obtained from the DCS system, respectively, and Y represents a prediction variable output after the LSTM model is learned.
Memory cells C of LSTM can be continually self-renewing and perfected in multiple simulations. In the testing process, the data set is divided into a training set and a testing set according to the proportion of 8:2, the training set is used for network learning, and the testing set is used for testing network performance. The LSTM captures the time variation and characteristics of input variables such as T, P and the like and the relation between the input variables and the output variable Y, thereby outputting the predicted variable Y by utilizing the trained LSTM model t+1 ,...,y t+n By comparison with the actual variable Y t+1 ,…,Y t+n And comparing and verifying the accuracy of the LSTM prediction model.
C in FIG. 4 t ，h t Are all hidden states of the LSTM: c t The memory unit is used for memorizing the state of the neuron, so that the LSTM unit has the capacity of storing, reading, resetting and updating long-distance historical information; h is t Is a hidden layer, is C t The content after attenuation is a hidden state inside the LSTM model.
Step 106: and training the LSTM pre-training model by adopting the historical data of the operation parameters of the equipment to be predicted and the operation parameters of the equipment in the relevant equipment parameter set, and obtaining the LSTM prediction model of the operation parameters of the equipment to be predicted after training.
In order to ensure the accuracy of the model prediction result, in the online prediction stage, the latest period of historical data of the operation parameters of each device to be predicted is acquired each time for training, the parameters such as the weight and the offset of each node in the model are updated, and the LSTM prediction model of the operation parameters of the device to be predicted is obtained after training is completed. The network hyper-parameters such as learning rate, optimizer, hidden layer node number and the like are determined in the model pre-training process, and the determination is not changed. Each training of the online prediction phase only changes the parameters of the model, such as weights, biases, etc.
After the trained LSTM prediction model is obtained, the predicted value of the operation parameter of the equipment to be predicted can be obtained only by inputting the latest historical data of the relevant equipment parameter set into the trained LSTM prediction model. And comparing the obtained predicted value with a threshold value, judging whether the equipment operation parameters exceed a normal allowable range, and giving an alarm if the equipment operation parameters exceed the normal allowable range.
Step 107: and obtaining a predicted value of the operation parameter of the equipment to be predicted by adopting the LSTM prediction model, and carrying out safety early warning according to a comparison result of the predicted value and the corresponding equipment operation parameter threshold value.
The invention adopts Principal Component Analysis (PCA) method to calculate T corresponding to all equipment operating parameters 2 And taking the statistical threshold as an equipment operation parameter threshold. The PCA method is usually tried out when the dimension of the data is reduced, the PCA replaces the original data with a plurality of main aspects of the data, and the main aspects firstly ensure that a large amount of information in the original data is contained and are not related to each other. T is 2 Based on an early warning index established by a Principal Component Analysis (PCA) method from the statistical viewpoint, the T corresponding to the operation parameter of each device is calculated by performing off-line modeling on a large amount of DCS off-line data 2 And (5) counting a threshold value. Using PCA and T 2 Providing an equipment operation parameter threshold value for the online early warning judgment in the step 107, comparing each parameter predicted value with the threshold value, and performing safety early warning if the predicted value exceeds the threshold value.
When the device is early-warned, the device of the abnormal propagation path and the root cause node can be displayed according to the SDG directed graph, and the problem that which device should be operated to solve when the abnormality occurs can be solved. SDG is an analytical graph that expresses the interplay relationships between process variables, reflecting the characteristics of the devices involved and the overall topology of the system.
The oil and gas gathering and transportation station equipment parameter early warning method provided by the invention solves the problem of prediction accuracy of equipment operation parameters in an oil and gas station library by using a long-term and short-term memory network, and a user can monitor the state of each equipment in real time and accurately predict the equipment operation parameters. When the method is realized by adopting program codes, one prediction code of the oil and gas gathering and transportation station can realize intelligent prediction and early warning on the operation parameters of all equipment, namely, a universal LSTM prediction model is provided, so that the universality of the early warning method is improved, and the resources of computing equipment are saved.
The process of the present invention is illustrated in detail below by means of a specific example.
The invention provides a parameter early warning method for oil and gas gathering and transportation station equipment, which mainly comprises the processes of key variable selection, principal component analysis, network hyper-parameter storage, LSTM algorithm prediction and the like. Assuming that 30 out-feed oil temperature and other equipment operation parameter (parameter for short) historical data samples are totally 34014 from 0 in 19 th 6 th to 21 st in 23 th 6 th, the method provided by the embodiment of the invention mainly comprises the following steps:
1) Data pre-processing
And sequencing the historical data of each parameter according to a time sequence, identifying and processing the abnormal values by adopting a Laeyda criterion method and a Showville method, constructing a process data set, and accelerating the training efficiency and precision of a subsequent LSTM model.
2) Correlation coefficient calculation and feature variable selection
And for each equipment operation parameter in the process data set, calculating the correlation between each equipment operation parameter and all other equipment operation parameters by using the spearman rank correlation coefficient, and constructing a correlation coefficient set. And drawing a correlation heat chart among the variables according to the correlation coefficient set. Correlation thermodynamic diagrams were analyzed, where blue indicates positive correlation between variables, red indicates negative correlation between variables, and the size of the dots indicates the size of the absolute value of the correlation. According to the analysis result, 25 variables closely related to the temperature of the output oil are screened, including the oil inlet pressure of the No. 1 water separator, the oil-water interface of the No. 1 water separator, the gas outlet pressure of the No. 1 water separator, the oil inlet pressure of the No. 2 water separator, the oil-water interface of the No. 2 water separator, the oil level of the No. 2 heating furnace, the water jacket temperature of the No. 2 heating furnace, the inlet temperature of the No. 3 heating furnace, the liquid level of the No. 1 settling tank, the liquid level of the No. 3 settling tank, the liquid level of the No. 8 settling tank, the inlet temperature of an output pump, the pressure of the output oil, the flow of the output oil, the liquid level of the No. 8 light oil tank, the outlet temperature of the No. 1 air compressor and the like.
3) Model building and optimization
And according to the calculation of the correlation between the previous variables and the selection of the characteristic variables, constructing a relevant equipment parameter set according to the selected 25 characteristic variables, wherein the relevant equipment parameter set is used as an input variable of the LSTM model, and the temperature of the external oil is used as an output variable. 30 points from 6 months 19 days 0 to 6 months 23 days 21 are divided into 34014 data, and the data are recorded according to the following data: the scale of 2 is divided into training set and test set, and divided into 10 groups, each group is composed of 3402 continuous time sequence data, and comprises 2721 data for training and 682 data for testing, which are input into the LSTM model in 10 batches.
According to the characteristics of the LSTM model, selecting and optimizing the super-parameters by using a genetic algorithm, and determining the network super-parameter step frequency, the number of network nodes, an activation function, an optimizer, the learning rate and the batch processing quantity by verifying and checking respectively as follows: 50. 75, tanh, adam, 0.1, 10.
4) Network hyper-parameter preservation
Designing a relational database, and enabling characteristic variables and network hyper-parameters which affect the temperature of the output oil to be as follows: the step frequency, the number of network nodes, the activation function, the optimizer and the learning rate are stored in a database, a general database is established, and corresponding parameters can be directly extracted from the general database when the general database is used.
5) Online model training and online parameter prediction
Calculating T of output oil temperature by offline modeling through data 2 And (3) counting a threshold, training and testing each group of data for 10 times under the condition that other conditions are unchanged and the same model parameters are ensured, predicting the temperature of the oil output outside the combined station by using a trained LSTM prediction model, and measuring the prediction effect of the model by using a loss function expressed by root mean square error. The prediction error is about 0.5 percent and 0.6 percent, and the technical requirement is met.
6) On-line early warning judgment
And comparing the obtained predicted value with a threshold value, judging whether the equipment operation parameters exceed a normal allowable range, and giving an alarm if the equipment operation parameters exceed the normal allowable range.
Based on the method provided by the invention, the invention also provides an oil and gas gathering and transportation station equipment parameter early warning system, which comprises the following steps:
the process data set construction module is used for acquiring and preprocessing equipment operation parameter historical data of all equipment in the oil and gas gathering and transportation station to form a process data set; all the equipment in the oil-gas gathering and transportation station comprises a three-phase separator, an external transportation pump, an oil storage tank and a heating furnace; the equipment operation parameters comprise three-phase separator pressure, three-phase separator liquid level, external conveying pump pressure, oil storage tank liquid level and heating furnace temperature;
a related device parameter set construction module, configured to calculate, for each device operating parameter in the process data set, a correlation between each device operating parameter and all other device operating parameters using a spearman rank correlation coefficient, and establish a related device parameter set for each device operating parameter;
the LSTM model pre-training module is used for pre-training the LSTM model according to the relevant equipment parameter set of each equipment operation parameter, selecting and optimizing the LSTM model hyper-parameters by using a genetic algorithm, and determining network hyper-parameters;
the general data storage module is used for storing the relevant equipment parameter set of each equipment operation parameter and the corresponding network hyper-parameter into a database to obtain a general database;
the pre-training model building module is used for acquiring a relevant equipment parameter set of the operation parameters of the equipment to be predicted and corresponding network hyper-parameters from the general database in an online prediction stage, and substituting the relevant equipment parameter set and the corresponding network hyper-parameters into the LSTM model to obtain the LSTM pre-training model;
the LSTM model training module is used for training the LSTM pre-training model by adopting historical data of the operation parameters of the equipment to be predicted and the operation parameters of the equipment in the relevant equipment parameter set, and obtaining the LSTM prediction model of the operation parameters of the equipment to be predicted after training is finished;
and the equipment parameter prediction early warning module is used for obtaining a predicted value of the operation parameter of the equipment to be predicted by adopting the LSTM prediction model and carrying out safety early warning according to a comparison result of the predicted value and a corresponding equipment operation parameter threshold value.
The invention provides a parameter early warning method and system for oil gas gathering and transportation station equipment, based on a general long-term and short-term memory network, the main early warning process comprises seven parts of data preprocessing, correlation coefficient calculation and characteristic variable selection, network model building and optimization, network hyper-parameter storage, online model training, online parameter prediction, online early warning judgment and the like, and mainly comprises the following steps: processing the acquired data abnormal values, obtaining key variable and network model super parameter values by using correlation analysis, and storing a key variable set (namely a related equipment parameter set) and network super parameters into a database, thereby establishing different LSTM pre-training models for each equipment operation parameter; in the on-line prediction stage, parameters of the network model are obtained by reading the general database and are trained, the predicted value of the equipment operation parameters at the next moment is obtained by predicting the trained LSTM prediction model, the predicted value is compared with the allowable range threshold value, whether the equipment carries out safety early warning or not is judged, finally, intelligent prediction early warning on all the parameters is realized by one prediction code of the oil and gas gathering and transportation station, and the universality, the early warning efficiency and the early warning precision of the early warning method are greatly improved.
Further, the present invention also provides an electronic device, which may include: a processor, a communication interface, a memory, and a communication bus. The processor, the communication interface and the memory are communicated with each other through a communication bus. The processor may invoke a computer program in memory to execute the method for hydrocarbon gathering station equipment parameter early warning.
Further, the computer program in the memory described above may be stored in a computer-readable storage medium when it is implemented in the form of a software functional unit and sold or used as a separate product. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: u disk, removable hard disk, read only memory, random access memory, magnetic or optical disk, etc. for storing program codes.
Further, the present invention also provides a non-transitory computer readable storage medium, on which a computer program is stored, which when executed can implement the method for pre-warning the equipment parameter of the oil and gas gathering station.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the method part for description.
The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.
Claims (8)
1. An oil and gas gathering station equipment parameter early warning method is characterized by comprising the following steps:
acquiring historical data of equipment operation parameters of all equipment in an oil and gas gathering and transportation station and preprocessing the historical data to form a process data set; all the equipment in the oil-gas gathering and transportation station comprises a three-phase separator, an external transportation pump, an oil storage tank and a heating furnace; the equipment operation parameters comprise three-phase separator pressure, three-phase separator liquid level, external conveying pump pressure, oil storage tank liquid level and heating furnace temperature;
for each equipment operation parameter in the process data set, calculating the correlation between each equipment operation parameter and all other equipment operation parameters by using a spearman rank correlation coefficient, and establishing a relevant equipment parameter set of each equipment operation parameter;
pre-training the LSTM model according to the relevant equipment parameter set of each equipment operation parameter, selecting and optimizing the LSTM model hyper-parameters by using a genetic algorithm, and determining network hyper-parameters;
storing the relevant equipment parameter set of each equipment operation parameter and the corresponding network hyper-parameter into a database to obtain a general database;
in the online prediction stage, a relevant device parameter set of the device operation parameters to be predicted and a corresponding network hyper-parameter are obtained from the general database and are substituted into the LSTM model to obtain an LSTM pre-training model;
training the LSTM pre-training model by adopting historical data of the operation parameters of the equipment to be predicted and the operation parameters of the equipment in the relevant equipment parameter set, and obtaining the LSTM prediction model of the operation parameters of the equipment to be predicted after training;
and obtaining a predicted value of the operation parameter of the equipment to be predicted by adopting the LSTM prediction model, and carrying out safety early warning according to a comparison result of the predicted value and the corresponding equipment operation parameter threshold value.
2. The method for early warning of equipment parameters of an oil and gas gathering and transportation station according to claim 1, wherein the step of obtaining and preprocessing historical data of equipment operating parameters of all equipment in the oil and gas gathering and transportation station to form a process data set specifically comprises the steps of:
acquiring historical data of equipment operation parameters of all equipment in the oil and gas gathering and transportation station, and sequencing the historical data according to a time sequence to obtain a time sequence data set of the operation parameters of all the equipment;
and identifying and processing abnormal values in the time sequence data set by adopting a Lauda criterion method and a Showville method to obtain a process data set of all equipment operation parameters.
3. The method of claim 1, wherein the step of calculating the correlation between each plant operating parameter and all other plant operating parameters using the spearman rank correlation coefficient for each plant operating parameter in the process data set and establishing the set of relevant plant parameters for each plant operating parameter comprises:
for each equipment operation parameter in the process data set, calculating the correlation between each equipment operation parameter and all other equipment operation parameters by using a spearman rank correlation coefficient, and determining that the two equipment operation parameters are closely related when the absolute value of the spearman rank correlation coefficient is greater than 0.8;
finding out other equipment operation parameters closely related to each equipment operation parameter, and establishing a related equipment parameter set of each equipment operation parameter.
4. The hydrocarbon gathering station equipment parameter early warning method as set forth in claim 1, wherein said pre-training the LSTM model according to the set of relevant equipment parameters for each equipment operating parameter, selecting and optimizing the LSTM model hyper-parameters using a genetic algorithm, determining network hyper-parameters, specifically comprising:
taking a relevant equipment parameter set of each equipment operation parameter as an input variable of the LSTM model, taking each equipment operation parameter as an output variable, and training the LSTM model;
in the LSTM model training process, selecting and optimizing the LSTM model hyper-parameters by using a genetic algorithm to determine network hyper-parameters; the network hyper-parameters comprise step frequency, the number of network nodes, an activation function, an optimizer and a learning rate.
5. The method for early warning the equipment parameters of the oil and gas gathering and transportation station according to claim 1, wherein before the early warning is performed according to the comparison result between the predicted value and the corresponding equipment operation parameter threshold value, the method for early warning the equipment operation parameters of the oil and gas gathering and transportation station by using the LSTM prediction model to obtain the predicted value of the equipment operation parameters to be predicted further comprises the following steps:
calculating T corresponding to all equipment operating parameters by adopting a principal component analysis method 2 And taking the statistical threshold as a device operation parameter threshold.
6. The utility model provides an oil gas gathering station equipment parameter early warning system which characterized in that includes:
the process data set construction module is used for acquiring and preprocessing equipment operation parameter historical data of all equipment in the oil and gas gathering and transportation station to form a process data set; all the equipment in the oil-gas gathering and transportation station comprises a three-phase separator, an external transportation pump, an oil storage tank and a heating furnace; the equipment operation parameters comprise three-phase separator pressure, three-phase separator liquid level, external conveying pump pressure, oil storage tank liquid level and heating furnace temperature;
a related device parameter set construction module, configured to calculate, for each device operating parameter in the process data set, a correlation between each device operating parameter and all other device operating parameters using a spearman rank correlation coefficient, and establish a related device parameter set for each device operating parameter;
the LSTM model pre-training module is used for pre-training the LSTM model according to the relevant equipment parameter set of each equipment operation parameter, selecting and optimizing the LSTM model hyper-parameters by using a genetic algorithm, and determining network hyper-parameters;
the general data storage module is used for storing the relevant equipment parameter set of each equipment operation parameter and the corresponding network hyper-parameter into a database to obtain a general database;
the pre-training model building module is used for acquiring a relevant equipment parameter set of the operation parameters of the equipment to be predicted and corresponding network hyper-parameters from the general database in an online prediction stage, and substituting the relevant equipment parameter set and the corresponding network hyper-parameters into the LSTM model to obtain the LSTM pre-training model;
the LSTM model training module is used for training the LSTM pre-training model by adopting historical data of the operation parameters of the equipment to be predicted and the operation parameters of the equipment in the relevant equipment parameter set, and obtaining the LSTM prediction model of the operation parameters of the equipment to be predicted after training is finished;
and the equipment parameter prediction early warning module is used for obtaining a predicted value of the operation parameter of the equipment to be predicted by adopting the LSTM prediction model and carrying out safety early warning according to a comparison result of the predicted value and a corresponding equipment operation parameter threshold value.
7. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor when executing the computer program implements the oil and gas gathering station device parameter alerting method of any of claims 1 to 5.
8. A non-transitory computer readable storage medium having stored thereon a computer program, wherein the computer program when executed implements the oil and gas gathering station equipment parameter warning method as claimed in any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211359817.9A CN115688581A (en) | 2022-11-02 | 2022-11-02 | Oil gas gathering and transportation station equipment parameter early warning method, system, electronic equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211359817.9A CN115688581A (en) | 2022-11-02 | 2022-11-02 | Oil gas gathering and transportation station equipment parameter early warning method, system, electronic equipment and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115688581A true CN115688581A (en) | 2023-02-03 |
Family
ID=85048559
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211359817.9A Pending CN115688581A (en) | 2022-11-02 | 2022-11-02 | Oil gas gathering and transportation station equipment parameter early warning method, system, electronic equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115688581A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117034157A (en) * | 2023-10-08 | 2023-11-10 | 广州健新科技有限责任公司 | Hydropower equipment fault identification method and system combining multimodal operation data |
| CN117474345A (en) * | 2023-12-28 | 2024-01-30 | 广州恩伟博科技有限公司 | Intelligent environment-friendly remote real-time monitoring and early warning method and system |
-
2022
- 2022-11-02 CN CN202211359817.9A patent/CN115688581A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117034157A (en) * | 2023-10-08 | 2023-11-10 | 广州健新科技有限责任公司 | Hydropower equipment fault identification method and system combining multimodal operation data |
| CN117034157B (en) * | 2023-10-08 | 2024-01-12 | 广州健新科技有限责任公司 | Hydropower equipment fault identification method and system combining multimodal operation data |
| CN117474345A (en) * | 2023-12-28 | 2024-01-30 | 广州恩伟博科技有限公司 | Intelligent environment-friendly remote real-time monitoring and early warning method and system |
| CN117474345B (en) * | 2023-12-28 | 2024-04-02 | 广州恩伟博科技有限公司 | Intelligent environment-friendly remote real-time monitoring and early warning method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115688581A (en) | Oil gas gathering and transportation station equipment parameter early warning method, system, electronic equipment and medium | |
| CN110348752B (en) | Large industrial system structure safety assessment method considering environmental interference | |
| CN113779496A (en) | Power equipment state evaluation method and system based on equipment panoramic data | |
| CN108921230A (en) | Method for diagnosing faults based on class mean value core pivot element analysis and BP neural network | |
| CN116167527B (en) | Pure data-driven power system static safety operation risk online assessment method | |
| CN112632765B (en) | Combat capability assessment method combining weighting method and SEM method | |
| CN113268883A (en) | Method for predicting corrosion rate of submarine crude oil pipeline based on PCA-ABC-SVM model | |
| CN116432123A (en) | Electric energy meter fault early warning method based on CART decision tree algorithm | |
| CN115618732A (en) | Nuclear reactor digital twin key parameter autonomous optimization data inversion method | |
| CN114548494A (en) | Visual cost data prediction intelligent analysis system | |
| CN117312972A (en) | Method for identifying health state of scraper conveyor speed reducer | |
| CN113762591A (en) | Short-term electric quantity prediction method and system based on GRU and multi-core SVM counterstudy | |
| CN113033898A (en) | Electrical load prediction method and system based on K-means clustering and BI-LSTM neural network | |
| CN115864644A (en) | Relay protection device state evaluation method, system, equipment and medium | |
| Gan | Discrete Hopfield neural network approach for crane safety evaluation | |
| CN111061191B (en) | Distributed oil-gas storage tank remote operation and maintenance method | |
| CN113849479A (en) | Comprehensive energy supply station oil tank leakage detection method based on instant learning and self-adaptive threshold | |
| Najar et al. | Comparative Machine Learning Study for Estimating Peak Cladding Temperature in AP1000 Under LOFW | |
| CN116720662B (en) | Distributed energy system applicability evaluation method based on set pair analysis | |
| CN111651938B (en) | Variable coal quality unit output prediction method based on thermodynamic calculation and big data | |
| Zhang et al. | Health Condition Assessment of Marine Systems Based on an Improved Radar Chart | |
| CN116595883B (en) | Real-time online system state correction method for numerical reactor | |
| Du et al. | Study of plant communities based on supervised learning model SVM | |
| CN117457102A (en) | Heat exchange equipment ammonium salt crystallization risk prediction method based on data driving | |
| CN116288532A (en) | Method and equipment for monitoring electrolytic tank |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |