CN-121457723-B - Wellhead device fault prediction system based on narrowband Internet of things

Abstract

The invention relates to the technical field of intelligent operation and maintenance and safety early warning of oil and gas equipment, in particular to a wellhead device fault prediction system based on a narrowband Internet of things. The system comprises a data acquisition module, a factor quantification module, a dynamic weight module, a fault prediction module, an uncertainty quantification module, a risk assessment module and a decision response module, wherein the data acquisition module is used for acquiring unstructured text data and structured time sequence data of a target wellhead device, the factor quantification module is used for determining a working condition mutation factor, the dynamic weight module is used for determining a dynamic weight factor, the fault prediction module is used for responding to the dynamic weight factor and determining direct fault probability, the uncertainty quantification module is used for determining an uncertainty index based on multiple prediction results of the fault prediction module, the risk assessment module is used for combining the direct fault probability and the uncertainty index to determine a comprehensive risk index, and the decision response module is used for generating a hierarchical early warning instruction. The invention realizes the transition from passive response to active prediction, obviously enhances the prospective, self-adaption and timeliness of fault prediction, and realizes early warning.

Inventors

• LIU JIALI

Assignees

• 西安力勘石油能源科技有限公司

Dates

Publication Date

20260508

Application Date

20251110

Claims (2)

1. 1. Wellhead fault prediction system based on narrowband internet of things, characterized by comprising: the data acquisition module is used for acquiring unstructured text data and structured time sequence data of the target wellhead device; the factor quantization module is used for determining the working condition mutation factor based on unstructured text data The method comprises the steps of processing unstructured text data through a pre-training language model, extracting key semantic units, obtaining corresponding reference influence weights, confidence levels and intensity modifier scores aiming at the key semantic units, combining the reference influence weights, the confidence levels and the intensity modifier scores, and calculating working condition mutation factors ; Dynamic weight module for mutating factor according to working condition Determining dynamic weighting factors The dynamic weight module is specifically used for calling the working condition mutation factors Calculating dynamic weight factor by preset self-adaptive regulating function The self-adaptive adjusting function specifically comprises the following steps: wherein, the method comprises the steps of, Is a preset reference physical loss weight, For the preset sensitivity coefficient to be a set, As a hyperbolic tangent function; The fault prediction module is used for responding to the dynamic weight factors and determining direct fault probability based on the structured time sequence data; The uncertainty quantization module is used for determining an uncertainty index based on multiple prediction results of the fault prediction module, and comprises a Dropout layer in the fault prediction module is started for multiple times in an reasoning stage and is propagated forward to obtain a group of prediction results; the risk assessment module is used for combining the direct fault probability and the uncertainty index to determine a comprehensive risk index, and comprises the steps of calling the direct fault probability, the uncertainty index and a preset uncertainty amplification coefficient, carrying out weighted amplification on the direct fault probability based on the product of the uncertainty index and the uncertainty amplification coefficient, and outputting the weighted amplification result as the comprehensive risk index; The decision response module is used for generating a grading early warning instruction according to the comprehensive risk index and a preset risk threshold value; the fault prediction module adopts a physical information neural network model, and a loss function of the physical information neural network model comprises data driving loss and physical information loss; The physical information loss is calculated by substituting the prediction result of the fault prediction module into a preset standard erosion model describing the erosion and abrasion process of the target wellhead device, so as to quantify the deviation degree of the prediction result to the standard erosion model.
2. 2. The wellhead fault prediction system based on the narrowband internet of things according to claim 1, wherein the decision response module is specifically configured to: Comparing the comprehensive risk index with a preset first risk threshold value and a preset second risk threshold value; outputting a first-level early warning instruction when the comprehensive risk index is lower than a first risk threshold; Outputting a second-level early warning instruction when the comprehensive risk index is between the first risk threshold and the second risk threshold; And outputting a third-level early warning instruction when the comprehensive risk index is higher than the second risk threshold.

Description

Wellhead device fault prediction system based on narrowband Internet of things Technical Field The invention relates to the technical field of intelligent operation and maintenance and safety early warning of oil and gas equipment, in particular to a wellhead device fault prediction system based on a narrowband Internet of things. Background How to accurately predict potential faults of the wellhead device, and realize the transition from passive response to active predictive maintenance, is a core technical challenge for guaranteeing safe and efficient operation of an oil and gas field; The existing fault prediction technology relies on structured sensor time sequence data, expert experience and prospective risk information contained in unstructured texts such as geological reports, operation instructions and the like are difficult to fuse, and therefore prediction model information is incomplete to input; meanwhile, the traditional model is usually static, the prediction logic of the traditional model cannot be dynamically adjusted according to the working condition of real-time change, and the accuracy and the robustness of the prediction are obviously reduced when the traditional model faces to the conditions of geological environment mutation, operation adjustment and the like; in addition, the existing method often provides a single fault probability prediction value, lacks quantitative evaluation of uncertainty of a prediction result, and is difficult to support reliable decision in a high-risk scene; Thus, there is a need for a new fault prediction scheme that can fuse multi-source heterogeneous data, achieve adaptive prediction, and quantify uncertainty. The above information disclosed in the above background section is only for enhancement of understanding of the background of the disclosure and therefore it may include information that does not form the prior art that is already known to those of ordinary skill in the art. Disclosure of Invention In order to solve the technical problems, the invention discloses a wellhead device fault prediction system based on a narrowband Internet of things, and specifically, the technical scheme of the invention comprises the following steps: the data acquisition module is used for acquiring unstructured text data and structured time sequence data of the target wellhead device; The factor quantization module is used for determining a working condition mutation factor based on unstructured text data; the dynamic weight module is used for determining a dynamic weight factor according to the working condition mutation factor; The fault prediction module is used for responding to the dynamic weight factors and determining direct fault probability based on the structured time sequence data; The uncertainty quantization module is used for determining an uncertainty index based on the multiple prediction results of the fault prediction module; The risk assessment module is used for combining the direct fault probability and the uncertainty index to determine a comprehensive risk index; and the decision response module is used for generating a grading early warning instruction according to the comprehensive risk index and a preset risk threshold value. Further, the factor quantization module is configured to determine a working condition mutation factor based on unstructured text data, and includes: Processing unstructured text data through a pre-training language model, and extracting key semantic units; Aiming at the key semantic units, obtaining corresponding reference influence weights, confidence degrees and intensity modifier scores; and calculating the working condition mutation factor by combining the reference influence weight, the confidence level and the intensity modifier score. Further, the fault prediction module adopts a physical information neural network model, and a loss function of the physical information neural network model comprises data driving loss and physical information loss. Further, the dynamic weight factor is used for adjusting the weight of the physical information loss in the loss function; The dynamic weight module is specifically used for: calling a working condition mutation factor; and calculating a dynamic weight factor through a preset self-adaptive adjusting function. Further, the physical information loss is calculated by substituting the prediction result of the failure prediction module into a preset standard erosion model describing the erosion and abrasion process of the target wellhead device, so as to quantify the deviation degree of the prediction result from the standard erosion model. Further, the uncertainty quantization module is configured to determine an uncertainty indicator based on a plurality of prediction results of the fault prediction module, and includes: Enabling a Dropout layer in the fault prediction module for multiple times in an reasoning stage and performing forward propagation to obtain a group of