Process for removing paraffin and unblocking oil well through thermal washing of hollow sucker rod

Abstract

The invention relates to the field of petroleum engineering, and provides a hollow sucker rod thermal washing paraffin removal blockage removal oil well process which comprises the steps that firstly, well washing fluid is output through a thermal washing pump truck or a high-pressure washing machine, then the well washing fluid is injected into an inner cavity of a hollow sucker rod through a high-pressure manifold, and then the well washing fluid flows in a circulation channel formed between the hollow sucker rod and an oil pipe; and the high-pressure jet flow can directly act on the wax blocks in the oil pipe to help to break and remove the sediments, the whole process needs to be continuously carried out, and it is ensured that the wax blocks accumulated in the oil pipe can be effectively removed through repeated circulation of the well washing fluid, so that the oil pipe is kept clean and smooth.

Description

Technical Field

[0001] This application relates to the field of oil engineering, and more specifically, to a process for hot washing paraffin removal and plugging removal from oil wells using hollow sucker rods. Background Art

[0002] In oil and gas field oil production, ensuring stable and high-yield oil wells is an ongoing pursuit goal. However, the phenomenon of wax deposition is widespread. Wax deposition can significantly increase the fluid flow resistance in the tubing, leading to an increase in the load of mechanical production equipment. This not only reduces the equipment efficiency but also shortens the trouble-free period of the oil well, greatly increasing the oil well production cost. As a common means to prevent and remove wax deposition in the tubing at present, hot washing paraffin removal plays a key role in oil production.

[0003] However, there are many drawbacks in the existing hot washing paraffin removal technology. In the actual operation process, there are also serious safety and environmental protection problems. For example, when the tubing string is removed to the surface for paraffin removal of the surface tubing string after well killing, if the sum of the washing and well killing pump pressure and the static liquid column pressure is greater than the formation pressure, well leakage is very likely to occur. This will not only cause pollution to the production oil layer, reduce the paraffin removal effect, but also may have a negative impact on the surrounding environment. In addition, when operating in a water-sensitive formation, in order to protect the production oil layer, a clay stabilizer or swelling inhibitor needs to be added to the water, which makes the operation procedure cumbersome and complex, further increasing the operation difficulty and cost.

[0004] Therefore, a scheme for hot washing paraffin removal and plugging removal from oil wells using hollow sucker rods is desired. Summary of the Invention

[0005] This application aims at the shortcomings in the prior art and provides a process for hot washing paraffin removal and plugging removal from oil wells using hollow sucker rods.

[0006] According to one aspect of this application, there is provided a process for hot washing paraffin removal and plugging removal from oil wells using hollow sucker rods, which includes: Step 1: Outputting well washing fluid through a hot washing pump truck or a high-pressure cleaner.

[0007] Step 2: Injecting the well washing fluid into the inner cavity of the hollow sucker rod through a high-pressure manifold, and the hollow sucker rod and the tubing form a circulation channel.

[0008] Step 3: The well washing fluid flows out from the bottom end of the hollow sucker rod and forms a high-pressure jet, wherein the high-pressure jet acts on the wax block in the tubing.

[0009] Step 4: Continuously execute Step 1 to Step 3 to remove the wax block in the tubing.

[0010] Due to the adoption of the above technical solutions, this application has remarkable technical effects: The hot washing and wax removal and plugging removal process for hollow sucker rods provided by this application first outputs washing fluid through a hot washing pump truck or a high-pressure cleaner. Then, the washing fluid is injected into the inner cavity of the hollow sucker rod through a high-pressure pipe manifold. Next, the washing fluid flows in the circulation channel formed between the hollow sucker rod and the tubing, and sprays out from the bottom of the hollow sucker rod in the form of a high-pressure jet. The high-pressure jet directly acts on the wax blocks in the tubing to help break and remove these sediments. The whole process needs to be carried out continuously to ensure that the wax blocks accumulated inside the tubing can be effectively removed by repeatedly circulating the washing fluid, so as to keep the tubing clean and unobstructed. Brief Description of the Drawings

[0011] By describing the embodiments of this application in more detail in conjunction with the drawings, the above and other objects, features, and advantages of this application will become more obvious. The drawings are used to provide a further understanding of the embodiments of this application, and constitute a part of the specification. Together with the embodiments of this application, they are used to explain this application, and do not constitute a limitation to this application. In the drawings, the same reference numerals generally represent the same components or steps.

[0012] Figure 1 It is a flowchart of the hot washing and wax removal and plugging removal process for hollow sucker rods according to an embodiment of this application.

[0013] Figure 2 It is a flowchart of step 3 in the hot washing and wax removal and plugging removal process for hollow sucker rods according to an embodiment of this application.

[0014] Figure 3 It is a schematic diagram of data flow of step 3 in the hot washing and wax removal and plugging removal process for hollow sucker rods according to an embodiment of this application.

[0015] Figure 4 It is a flowchart of step 35 in the hot washing and wax removal and plugging removal process for hollow sucker rods according to an embodiment of this application. Detailed Description of the Embodiments

[0016] Next, exemplary embodiments according to this application will be described in detail with reference to the drawings. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments of this application. It should be understood that this application is not limited by the exemplary embodiments described here.

[0017] During the oil and gas field exploitation process, how to ensure the stable and high-yield production of oil wells has always been an important goal. However, the wax deposition phenomenon is common in actual production. This phenomenon will significantly increase the flow resistance of the fluid in the tubing, thereby increasing the load of the mechanical oil production equipment, resulting in a decline in equipment efficiency, shortening the trouble-free period of the oil well, and significantly increasing the exploitation cost. As a common means to deal with the problem of tubing wax deposition at present, the hot washing and wax removal technology plays a crucial role in oilfield production.

[0018] However, there are still many deficiencies in the existing hot wax cleaning technology, and there are relatively prominent safety and environmental protection hazards in actual operation. For example, when the tubing string is lifted out for surface wax cleaning after well killing, if the sum of the pressure of the washing and killing pump and the static liquid column pressure exceeds the formation pressure, well leakage may occur. This will not only pollute the production oil layer, weaken the wax cleaning effect, but also have an adverse impact on the surrounding environment. In addition, when operating in a water-sensitive formation, in order to protect the production oil layer, it is usually necessary to add clay stabilizers or swelling inhibitors to the water, which makes the operation process more complex and increases the difficulty and cost of construction.

[0019] Based on this, the present application proposes a hot wax cleaning and plugging removal process for oil wells using hollow sucker rods. Figure 1 FIG. is a flow chart of the hot wax cleaning and plugging removal process for oil wells using hollow sucker rods according to an embodiment of the present application. As Figure 1 shown, the hot wax cleaning and plugging removal process for oil wells using hollow sucker rods according to an embodiment of the present application includes: Step 1, outputting a well washing fluid through a hot washing pump truck or a high-pressure cleaner; Step 2, injecting the well washing fluid into the inner cavity of the hollow sucker rod through a high-pressure manifold, and the hollow sucker rod and the tubing form a circulation channel; Step 3, the well washing fluid flows out from the bottom end of the hollow sucker rod and forms a high-pressure jet, wherein the high-pressure jet acts on the wax block in the tubing; Step 4, continuously executing Steps 1 to 3 to remove the wax block in the tubing.

[0020] In Step 1, a well washing fluid is output through a hot washing pump truck or a high-pressure cleaner. It should be understood that during the oil production process in oil and gas fields, paraffin deposits are likely to form on the inner wall of the tubing, and these deposits will gradually accumulate and cause a decline in the production capacity of the oil well. By outputting a high-temperature and high-pressure well washing fluid through a hot washing pump truck or a high-pressure cleaner, these deposited wax blocks can be effectively melted. Specifically, the well washing fluid output by the hot washing pump truck or the high-pressure cleaner is usually hot water or hot oil that has been heated. These heat carriers have a relatively high temperature, which can effectively improve the heat conduction efficiency of the wax deposition area inside the tubing, thereby achieving the purpose of quickly melting the wax blocks.

[0021] In step 2, the well-washing fluid is injected into the inner cavity of the hollow sucker rod through the high-pressure manifold, and the hollow sucker rod and the tubing form a circulation channel. It should be understood that the internal structure of the oil well is complex. In order for the well-washing fluid to accurately reach the wax deposition position in the tubing, a specific delivery path is required. The hollow sucker rod is located inside the tubing. Injecting the well-washing fluid into the inner cavity of the hollow sucker rod through the high-pressure manifold can ensure that the well-washing fluid can reach the target area stably and efficiently, avoiding problems such as leakage and dispersion of the well-washing fluid during transportation, and thus improving the utilization efficiency of the well-washing fluid. Since the well-washing fluid circulates in the circulation channel formed by the hollow sucker rod and the tubing, it can continuously scour and act on the wax blocks in the tubing. Compared with the traditional wax removal method, this cyclic scouring method can more comprehensively and deeply remove the wax blocks at different positions in the tubing, ensure the smoothness of the tubing, reduce the flow resistance of the fluid in the tubing, and improve the oil production efficiency. Particularly, a high-pressure hydraulic jet head is arranged at the bottom end of the hollow sucker rod, and a sealing blowout preventer device is arranged in the circulation channel formed by the hollow sucker rod and the tubing. By arranging the sealing blowout preventer device in the closed circulation channel formed by the hollow sucker rod and the tubing in this application, the accidental ejection of liquid can be effectively prevented, thereby ensuring the operation safety. At the same time, the closed circulation channel design enables the maximum transfer of heat energy to the inside of the tubing, ensuring that the well-washing fluid directly acts on the wax blocks and significantly improving the wax melting efficiency. In addition, the well-washing fluid only flows in the space above the sucker rod pump, avoiding direct contact with the production oil layer, preventing the loss of the well-washing fluid in low-pressure wells and protecting the oil layer from pollution. Moreover, this design does not require pulling out the tubing string by killing the well, is particularly suitable for operations in water-sensitive formations, does not require adding clay stabilizers or swelling inhibitors, not only protects the production oil layer, but also simplifies the operation process, reduces the operation difficulty and cost.

[0022] In step 3, the well-washing fluid flows out from the bottom end of the hollow sucker rod and forms a high-pressure jet, wherein the high-pressure jet acts on the wax blocks in the tubing. It should be understood that the wax blocks in the tubing are usually attached to positions such as the inner wall, seriously affecting the normal transportation of crude oil. With its strong impact force, the high-pressure jet can directly act on these wax blocks. By controlling the nozzle diameter of the high-pressure hydraulic jet head, its high-pressure energy can be fully utilized to break and peel off the wax blocks, so that they are detached from the tubing wall, thereby effectively removing the wax blocks and dredging the tubing. However, the nozzle diameter of the traditional high-pressure hydraulic jet head is usually fixed, or can only be adjusted in a limited and extensive manner based on manual experience. This fixed or manual adjustment method cannot adaptively adjust the nozzle diameter according to the change of the state parameters of the high-pressure jet during the wax removal process, and it is difficult to achieve precise control of the wax removal process.

[0023] Accordingly, in Step 3: the well-washing fluid flows out from the bottom end of the hollow sucker rod and forms a high-pressure jet, the technical concept of the present application is to use a sensor assembly to collect the state parameters (temperature value, pressure value, flow rate value, and jet velocity value) of the high-pressure jet to obtain a time queue of the high-pressure jet state parameters, and use an artificial intelligence-based data analysis and processing method to group the high-pressure jet state parameters. Then, sequence encoding is performed on the time queue of the grouped temperature values and the time queue composed of pressure values, flow rate values, and jet velocity values to obtain temperature time-series features and impact time-series features, so as to adaptively control the nozzle diameter of the high-pressure hydraulic jet head according to the principal component-guided ablation collaborative representation between the two. The present application can collect the state parameters of the high-pressure jet in real time and adaptively control the nozzle diameter of the high-pressure hydraulic jet head, overcoming the drawbacks brought by the fixed or manually adjusted nozzle diameter in the traditional technology, and can ensure the optimal paraffin removal effect under different working conditions.

[0024] Figure 2 FIG. is a flowchart of Step 3 in the hot washing paraffin removal and plugging removal process for oil wells using a hollow sucker rod according to an embodiment of the present application. Figure 3 FIG. is a schematic diagram of data flow in Step 3 in the hot washing paraffin removal and plugging removal process for oil wells using a hollow sucker rod according to an embodiment of the present application. As Figure 2 and Figure 3 shown, Step 3 includes: Step 31, using a sensor assembly to collect the state parameters of the high-pressure jet to obtain a time queue of the high-pressure jet state parameters, where the high-pressure jet state parameters include temperature value, pressure value, flow rate value, and jet velocity value; Step 32, performing variable grouping on the time queue of the high-pressure jet state parameters to obtain a time queue of temperature values and a time queue composed of pressure values, flow rate values, and jet velocity values; Step 33, performing sequence encoding on the time queue of the temperature values to obtain a temperature time-series feature implicit encoding vector; Step 34, performing sequence encoding on the time queue composed of pressure values, flow rate values, and jet velocity values to obtain an impact time-series feature implicit encoding vector; Step 35, performing principal component-guided temperature ablation-impact ablation collaborative analysis on the impact time-series feature implicit encoding vector and the temperature time-series feature implicit encoding vector to obtain a temperature ablation-impact ablation collaborative interaction feature vector; Step 36, based on the temperature ablation-impact ablation collaborative interaction feature vector, obtaining an optimization instruction, where the optimization instruction is used to indicate increasing the nozzle diameter of the high-pressure hydraulic jet head, decreasing the nozzle diameter of the high-pressure hydraulic jet head, or maintaining the nozzle diameter of the high-pressure hydraulic jet head.

[0025] In step 31, the sensor assembly is used to collect the state parameters of the high-pressure jet to obtain a time queue of the high-pressure jet state parameters, and the high-pressure jet state parameters include temperature value, pressure value, flow rate value, and jet velocity value. It should be understood that during the process of wax removal from the oil pipe, the distribution, hardness, and thickness of the wax blocks are not uniform. By collecting the temperature value, pressure value, flow rate value, and jet velocity value of the high-pressure jet, the working state of the jet during the wax removal operation can be grasped in real time. For example, through the temperature value, the heat energy transfer of the well-washing fluid can be understood. If the temperature is too low, the wax blocks may not be effectively melted. The pressure value can reflect the impact force of the jet on the wax blocks. If the pressure is insufficient, it is difficult to break the wax blocks. That is to say, these high-pressure jet state parameters provide a direct basis for judging the wax removal effect. By collecting these state parameters of the high-pressure jet, the current state of the high-pressure jet can be understood, and then the high-pressure jet can be dynamically adjusted according to the real-time changes, which can achieve precise control of the wax removal process and is conducive to improving the wax removal efficiency and quality.

[0026] The following is a detailed elaboration of a specific implementation process of "using the sensor assembly to collect the state parameters of the high-pressure jet to obtain a time queue of the high-pressure jet state parameters, and the high-pressure jet state parameters include temperature value, pressure value, flow rate value, and jet velocity value": First, in order to accurately obtain the state parameters of the high-pressure jet, including temperature value, pressure value, flow rate value, and jet velocity value, high-quality sensors suitable for specific environmental conditions must be selected. For example, when selecting a temperature sensor, considering that the well-washing fluid may be in a high-temperature environment, thermocouples or resistance temperature detectors are usually selected. Such sensors can not only withstand high temperatures but also have high precision, thus ensuring the authenticity and reliability of the measured data. For the selection of a pressure sensor, it is necessary to consider whether it can withstand a high-pressure environment and whether it has the ability to respond quickly. Such sensors are generally installed at the outlet of the high-pressure pump or at the key connection parts of the hollow sucker rod to monitor the pressure change of the well-washing fluid in real time. At the same time, in order to monitor the flow rate of the well-washing fluid, electromagnetic flow meters or other types of flow sensors are usually used. These devices need to be installed on the pipeline between the high-pressure pump and the high-pressure manifold to ensure that the amount of liquid passing through can be accurately measured. It should be noted that since it is difficult to directly measure the velocity of the high-pressure jet, advanced devices such as laser Doppler velocimeters or particle image velocimetry systems can be used.

[0027] Once the required sensor types are determined and the installation is completed, the next step is to ensure that these sensors can stably transmit data to the central control system. All the data collected by the sensors need to pass through an effective transmission channel to reach the control center for processing. This usually involves the selection of wireless or wired transmission methods. In some application scenarios, especially when there is a long distance between the sensors and the control system, the wireless transmission method is favored for its flexibility. However, wireless transmission may face problems such as signal interference. Therefore, in some scenarios with extremely high requirements for data transmission stability, wired transmission is more applicable. Regardless of which transmission method is adopted, it is necessary to ensure the real-time and reliability of data transmission, so as to ensure that the basic data for subsequent analysis is up-to-date and accurate.

[0028] Throughout the process, the sensors are not only responsible for collecting data, but also need to convert this data into a format recognizable by the central control system. For this purpose, the sensors usually have a built-in microprocessor that can perform preliminary processing on the raw data, such as filtering and smoothing, to remove noise and unnecessary fluctuations. In addition, to facilitate subsequent data integration and analysis, the data output by the sensors is often formatted into a unified time series form. This means that parameters such as temperature, pressure, flow rate, and jet velocity at each time point will be recorded and arranged in chronological order. This time series data is not only convenient for storage and management, but also provides convenience for subsequent data mining.

[0029] After the data collection is completed, the next step is to transmit it to the central control system. This process seems simple, but actually involves multiple complex links. First, the data starts from each sensor, and after preprocessing, it is transmitted to the receiving end through a specified communication protocol. During this period, any problem in any step may lead to data loss or error. To avoid this situation, a redundancy design is usually added to the transmission link, such as setting up a backup channel or adding a retransmission mechanism. At the same time, to improve data security, encryption measures are also taken to protect sensitive information from being stolen. When the data successfully reaches the central control system, it will enter the next stage - data analysis and processing. This involves a large amount of computing resources and algorithm support, aiming to extract valuable information from the massive data and then guide the control operation of the on-site high-pressure water jet head.

[0030] In step 32, the time queue of the high-pressure jet state parameters is grouped by variables to obtain a time queue of temperature values ​​and a time queue consisting of pressure values, flow values ​​and jet velocity values. Accordingly, it is considered that the temperature is significantly different from the other three parameters (pressure, flow, jet velocity) in physical properties and influencing mechanisms. The temperature value mainly reflects the thermal properties of the well washing fluid, and its changes are mainly related to the heating, heat dissipation and heat exchange process of the well washing fluid with the wax block. The pressure value, flow value and jet velocity value more reflect the fluid mechanics properties of the well washing fluid, which are interrelated and jointly determine the impact and flushing effect of the high-pressure jet on the wax block. Therefore, in order to more accurately capture the respective change rules, in the technical solution of the present application, the time queue of the high-pressure jet state parameters is grouped by variables to obtain a time queue of temperature values ​​and a time queue consisting of pressure values, flow values ​​and jet velocity values.

[0031] In step 33, the time queue of the temperature value is sequence-encoded to obtain the implicit coding vector of the temperature time series feature. Specifically, in an embodiment of the present application, the step S33 includes: using a sequence encoder based on a bidirectional LSTM model to sequence-encode the time queue of the temperature value to obtain the implicit coding vector of the temperature time series feature. Accordingly, considering that the temperature value is constantly changing with the passage of time, and there is an inherent dependency relationship among the changes over time, such as the temperature value at a certain moment may be affected by the temperature at the previous moment or the previous moments and various factors in the wax removal process. Based on this, in the technical solution of the present application, the time queue of the temperature value is sequence-encoded to capture and mine the implicit dependency between each time point, and obtain the implicit coding vector of the temperature time series feature. In particular, in an example of the present application, the time queue of the temperature value is sequence-encoded to obtain the implicit coding vector of the temperature time series feature using a sequence encoder based on a bidirectional LSTM model. It should be understood that in the oil well wax removal process, the change of the temperature value is not only affected by the past moment, but may also be related to the situation at the future moment. For example, as the wax block gradually dissolves, the temperature at the current moment may show a specific change trend due to the accelerated dissolution of the subsequent wax blocks. The bidirectional LSTM model consists of a forward LSTM and a backward LSTM. The forward LSTM processes data from the beginning of the sequence to capture information about past time steps; the backward LSTM processes data from the end of the sequence to capture information about future time steps. This enables the model to fully capture the dependencies between the temperature values ​​in the time queue, thereby more completely exploring the potential patterns of temperature changes.

[0032] The following is a detailed elaboration of a specific implementation process of "using a sequence encoder based on a bidirectional LSTM model to perform sequence encoding on the time queue of the temperature values to obtain the implicit encoding vector of the temperature time series features": First, the data preparation stage is the foundation of the entire implementation process. The main task of this stage is to perform necessary preprocessing on the temperature data of the well flushing fluid collected in real time. Since the data directly obtained by the sensor may have problems such as noise and missing values, a series of measures need to be taken to ensure the quality of the data. For example, the sliding average method or Kalman filtering technology can be used to smooth the original temperature time series, removing unnecessary fluctuations and outliers. For the missing parts in the data, linear interpolation or other appropriate interpolation methods can be used to fill them. In addition, before inputting the data into the bidirectional LSTM model, the batch size and time step need to be adjusted according to the specific situation. This is because the LSTM model processes data in time series of fixed length, and reasonably setting these parameters not only helps to improve the accuracy of the model, but also significantly affects its running efficiency. Through the above preprocessing steps, it can be ensured that the data input into the model is both representative and has good stability.

[0033] Next is the model construction stage. The bidirectional LSTM is an improved version of the LSTM network that allows information to flow from the past to the future and also enables future context information to be passed back to past nodes in reverse. This design is particularly suitable for tasks that require considering context information before and after, such as the temperature time series analysis in this application. Specifically, constructing a sequence encoder based on bidirectional LSTM mainly includes the following parts: First is the input layer, which is responsible for receiving the preprocessed temperature time series data; second is the bidirectional LSTM layer, which is the core part of the model and contains LSTM units in both the forward and backward directions. They process the time series data in the forward and reverse orders respectively, aiming to capture the long-term dependencies in the time series; then is the fully connected layer, which is used to further integrate the information output by the bidirectional LSTM layer and map it to the required dimension to form the final implicit encoding vector; finally is the output layer, which generates the final implicit encoding vector of the temperature time series features. In the specific implementation process, deep learning frameworks such as TensorFlow or PyTorch can be selected to build this model. These frameworks provide rich API support, making it relatively simple to define complex neural network structures and also facilitating debugging and optimization.

[0034] Once the model is built, it enters the training phase. The training process is essentially a process of continuously adjusting the model parameters to minimize the prediction error, usually using the mean squared error (MSE) or cross-entropy loss as the evaluation metric. To accelerate the convergence speed and improve the model performance, some hyperparameters need to be set, such as the learning rate, batch size, number of iterations, etc. A reasonable hyperparameter configuration can not only accelerate the training process but also effectively avoid overfitting. To avoid overfitting, technical means such as dropout and early stopping can be adopted. During the training process, the model will try to extract useful feature information from the input time series, which often requires a large amount of computing resources, especially when dealing with large-scale datasets. Therefore, GPUs are usually selected to accelerate the training process. It is worth noting that when training a bidirectional LSTM model, not only the accuracy of the model should be concerned, but also its generalization ability, that is, how the model performs on unseen data. Only when the model can maintain high accuracy and stability on new data can its value be truly realized.

[0035] After completing the model training, the time queue data of the preprocessed temperature values needs to be input into the bidirectional LSTM model. The model will perform layer-by-layer calculations on the input time series data of the temperature values according to the learned weights and biases, and finally output the hidden encoding vector of the temperature time series features. This hidden encoding vector contains the key feature information in the original temperature time series and can be used for subsequent data analysis steps.

[0036] In step 34, the time queue composed of the pressure value, flow value, and jet velocity value is sequence-encoded to obtain the hidden encoding vector of the impact time series features. Specifically, in the embodiment of the present application, the step S34 includes: using the sequence encoder based on the bidirectional LSTM model to sequence-encode the time queue composed of the pressure value, flow value, and jet velocity value to obtain the hidden encoding vector of the impact time series features. It should be understood that pressure, flow, and jet velocity are interrelated and have time dependence. For example, a change in pressure at a certain moment may affect the flow and jet velocity at subsequent moments. Therefore, in order to capture the complex dependence relationships and change patterns of these parameters in the time dimension, the present application sequence-encodes the time queue composed of the pressure value, flow value, and jet velocity value to effectively capture this implicit time series dynamic change information and obtain the hidden encoding vector of the impact time series features. In particular, in an example of the present application, the sequence encoder based on the bidirectional LSTM model is used to sequence-encode the time queue composed of the pressure value, flow value, and jet velocity value to comprehensively capture the complex bidirectional time dependence relationships in the time queues of these three parameters and obtain the hidden encoding vector of the impact time series features.

[0037] In step 35, principal component-guided temperature ablation-impact ablation collaborative analysis is performed on the impact timing feature implicit coding vector and the temperature timing feature implicit coding vector to obtain a temperature ablation-impact ablation collaborative interaction feature vector. Specifically, Figure 4 is a flowchart of step 35 in the hot washing and paraffin removal and plugging removal process of a hollow sucker rod according to an embodiment of the present application. As Figure 4 shown, the step S35 includes: step S351, performing principal component feature analysis on the impact timing feature implicit coding vector and the temperature timing feature implicit coding vector to obtain a set of impact timing feature principal component implicit coding vectors and a set of temperature timing feature principal component implicit coding vectors; step S352, calculating an impact-temperature timing feature difference embedding compensation implicit coding weight vector between the set of impact timing feature principal component implicit coding vectors and the set of temperature timing feature principal component implicit coding vectors; step S353, calculating the position-wise mean vectors of the set of impact timing feature principal component implicit coding vectors and the set of temperature timing feature principal component implicit coding vectors to obtain an impact timing principal component representation implicit coding vector and a temperature timing principal component representation implicit coding vector; step S354, based on the impact-temperature timing feature difference embedding compensation implicit coding weight vector, performing collaborative interaction on the impact timing principal component representation implicit coding vector and the temperature timing principal component representation implicit coding vector to obtain the temperature ablation-impact ablation collaborative interaction feature vector.

[0038] It should be understood that the impact timing feature implicit coding vector and the temperature timing feature implicit coding vector respectively contain the implicit feature information of the high-pressure jet impact parameters and temperature parameters changing with time, and there is a complex deep collaborative relationship between the two in the time dimension. For example, during the wax block dissolution process, an appropriate temperature can enhance the impact force of the jet on the wax block, and the jet impact helps the temperature transfer in the wax block. Therefore, in order to excavate and refine this interaction relationship to more accurately reflect the actual state of the paraffin removal process, in the technical solution of the present application, principal component-guided temperature ablation-impact ablation collaborative analysis is performed on the impact timing feature implicit coding vector and the temperature timing feature implicit coding vector to obtain a temperature ablation-impact ablation collaborative interaction feature vector. In particular, this method can remove those relatively unimportant information, extract the key comprehensive information that best represents the synergistic effect of the two, and at the same time, by introducing a compensation mechanism, it realizes the balanced interaction between features from different sources, thereby enhancing the synergistic interaction effect between the two.

[0039] Specifically, first, perform principal component feature analysis on the impact time-series feature implicit encoding vector and the temperature time-series feature implicit encoding vector to obtain a set of impact time-series feature principal component implicit encoding vectors and a set of temperature time-series feature principal component implicit encoding vectors. The above process can be expressed as: ; where is the impact time-series feature implicit encoding vector, is the principal component feature analysis operation, is the impact time-series feature sample covariance matrix calculated through , is the impact time-series feature principal component orthogonal matrix, are the respective impact time-series feature principal component implicit encoding vectors in the set of impact time-series feature principal component implicit encoding vectors, is the impact time-series feature diagonal matrix, are respectively and the corresponding weight values, is the transpose matrix of , is the temperature time-series feature implicit encoding vector, is the temperature time-series feature sample covariance matrix calculated through , is the temperature time-series feature principal component orthogonal matrix, are the respective temperature time-series feature principal component implicit encoding vectors in the set of temperature time-series feature principal component implicit encoding vectors, is the temperature time-series feature diagonal matrix, are respectively and the corresponding weight values, is the transpose matrix of .

[0040] It should be understood that the impact time-series feature implicit encoding vector and the temperature time-series feature implicit encoding vector contain a large amount of complex information, which is high-dimensional and redundant. High-dimensional data not only increases the computational complexity but also may interfere with the analysis and processing of key information. By performing principal component feature analysis on these two time-series feature implicit encoding vectors, the main directions in the original vector spaces of these two vectors can be found, and thus the feature dimensionality reduction of the original vectors can be achieved. Moreover, in the original impact and temperature time-series feature implicit encoding vectors, there is some information that has little impact on the wax removal process or is repetitive. By performing principal component feature analysis operations, the feature sparsification of the original feature vectors can be realized. Specifically, during the conversion process, the principal components are sorted according to the variance size. The first principal component has the largest variance, and the subsequent components successively capture the remaining variance. In this way, those features with smaller variances and little contribution to the overall information will be weakened or discarded. For example, in the set of principal component implicit encoding vectors of the impact time-series features, the main several principal component vectors centrally reflect the change characteristics of the impact parameters that have an important impact on wax removal, while those secondary and interfering features are diluted. The same is true for the set of principal component implicit encoding vectors of the temperature time-series features, which can make the key features related to temperature during the wax removal process more prominent, help analyze the effects of temperature and impact parameters on wax removal more accurately, and thus provide more accurate data support for the subsequent precise adjustment of wax removal process parameters.

[0041] Specifically, in the embodiment of the present application, the step 352 includes: constructing the set of principal component implicit encoding vectors of the impact time-series features and the set of principal component implicit encoding vectors of the temperature time-series features into a principal component aggregation implicit encoding feature map of the impact time-series features and a principal component aggregation implicit encoding feature map of the temperature time-series features. This process can be expressed as: ; where is each principal component implicit encoding vector in the set of principal component implicit encoding vectors of the impact time-series features, is each principal component implicit encoding vector in the set of principal component implicit encoding vectors of the temperature time-series features, is a shape reshaping operation, is the principal component aggregation implicit encoding feature map of the impact time-series features, is the principal component aggregation implicit encoding feature map of the temperature time-series features.

[0042] Calculate the differential embedding weights for the principal component aggregation implicit encoding feature map of the impact time-series features and the principal component aggregation implicit encoding feature map of the temperature time-series features to obtain a branch weight vector of the impact time-series features and a branch weight vector of the temperature time-series features. This process can be expressed as: ; where is the principal component aggregation implicit encoding feature map of the impact time-series features, is the principal component aggregation implicit coding feature map of the temperature time series features, is the average pooling operation, and are respectively the corresponding first weight matrix and second weight matrix, and are respectively the corresponding first weight matrix and second weight matrix, is the non-linear activation function, is the output activation function, is the impact time series feature branch weight vector, is the temperature time series feature branch weight vector.

[0043] Calculate the impact-temperature time series feature difference embedding compensation implicit coding weight vector between the impact time series feature branch weight vector and the temperature time series feature branch weight vector, and this process can be expressed as: ; where, is the impact time series feature branch weight vector, is the temperature time series feature branch weight vector, is subtraction by position point, is the absolute value operation, is the impact-temperature time series feature difference embedding compensation implicit coding weight vector.

[0044] It should be understood that the set of impact time series feature principal component implicit coding vectors and the set of temperature time series feature principal component implicit coding vectors are high-dimensional data, and it is very difficult to directly analyze and understand these high-dimensional data. By respectively constructing these two sets of principal component implicit coding vectors into three-dimensional feature maps, complex high-dimensional information can be presented in a more intuitive way. Specifically, each point in the feature map represents a mapping of a high-dimensional vector, and the distribution and mutual relationship of these points reflect the key features and associations in the original high-dimensional data. For example, in the principal component aggregation implicit coding feature map of the impact time series features, the change trends and mutual relationships of different impact parameter principal components during the paraffin removal process can be clearly seen, and this information is crucial for judging the effect of impact on wax block removal. Similarly, the principal component aggregation implicit coding feature map of the temperature time series features retains the key information of the temperature parameter principal components, which helps to analyze the influence of temperature on wax block dissolution.

[0045] Accordingly, considering that the shock time-series feature principal component aggregated implicit encoding feature map and the temperature time-series feature principal component aggregated implicit encoding feature map respectively reflect the key information of shock parameters and temperature parameters during the paraffin removal process, but there are differences between them. These differences contain valuable information about the paraffin removal process. For example, in a certain paraffin removal stage, the roles of shock and temperature in wax block removal are different. In order to deeply explore the differences between the two feature maps at each position and comprehensively understand the difference performance of shock and temperature parameters during the paraffin removal process, in this application, it is necessary to perform differential embedding weight calculation processing on the shock time-series feature principal component aggregated implicit encoding feature map and the temperature time-series feature principal component aggregated implicit encoding feature map. That is, the shock time-series feature branch weight vector and the temperature time-series feature branch weight vector obtained through differential embedding weight calculation can highlight the differential information that has a key impact on the paraffin removal process in the two feature maps. For example, if at a certain moment, the role of shock in wax block fragmentation is more critical than temperature, then in the shock time-series feature branch weight vector, the weight value at the corresponding position will be larger, thus highlighting this differential information. This helps to quickly locate the parameter difference part that needs to be focused on during the paraffin removal process in subsequent analysis and provides a clear direction for optimizing the paraffin removal process.

[0046] It should be understood that the shock time-series feature branch weight vector and the temperature time-series feature branch weight vector respectively reflect the relative importance of shock- and temperature-related features during the paraffin removal process, but they are independent of each other. By calculating the shock-temperature time-series feature differential embedding compensation implicit encoding weight vector, the two can be associated, comprehensively considering shock and temperature features, so that they can play a more balanced role in subsequent processing. For example, in some paraffin removal stages, the role of shock in wax block fragmentation is prominent, while in other stages, temperature is more critical for wax melting. This weight vector can dynamically adjust the contribution ratio of the two according to these changes to ensure that the synergistic effect of shock and temperature during the paraffin removal process reaches the optimal.

[0047] Then, calculate the position-wise mean vectors of the set of shock time-series feature principal component implicit encoding vectors and the set of temperature time-series feature principal component implicit encoding vectors to obtain the shock time-series principal component representation implicit encoding vector and the temperature time-series principal component representation implicit encoding vector. The above process can be expressed as: ; where is the th shock time-series feature principal component implicit encoding vector in the set of shock time-series feature principal component implicit encoding vectors, is the th temperature time-series feature principal component implicit encoding vector in the set of temperature time-series feature principal component implicit encoding vectors, is the number of vectors in the set of principal component implicit encoding vectors of impact time series features and the set of principal component implicit encoding vectors of temperature time series features, is the principal component representation implicit encoding vector of the impact time series, is the principal component representation implicit encoding vector of the temperature time series.

[0048] It should be understood that the set of principal component implicit encoding vectors of impact time series features and the set of principal component implicit encoding vectors of temperature time series features usually contain a large amount of complex data. Directly performing subsequent aggregation and interaction operations on these high-dimensional and large-capacity data involves huge computational amounts and low efficiency. By calculating the position-wise mean vectors for the two sets respectively, averaging the feature values of multiple data points at the same position, the principal component representation implicit encoding vector of the impact time series and the principal component representation implicit encoding vector of the temperature time series can be obtained, which can effectively achieve data compression. For example, originally each set may contain thousands of vector data at different times and different dimensions. After calculating the mean vector, the data volume is significantly reduced, and in subsequent aggregation and interaction calculations, the computational complexity is significantly reduced, improving the efficiency of the entire process analysis and optimization. It should be particularly noted that this step focuses on the average of feature values at the same position and can extract the core information of each feature vector. Specifically, during the wax removal process, although the impact and temperature parameters at different times fluctuate, calculating the mean by position can eliminate the interference caused by some random fluctuations and prominently reflect the main trends and core features of the impact and temperature parameters changing over time. For example, when analyzing the effect of impact on the wax block, the principal component representation implicit encoding vector of the impact time series can more clearly show the core change trends of the impact parameters at each stage, helping to quickly judge the key impact points of the impact on the wax removal effect. Similarly, the principal component representation implicit encoding vector of the temperature time series can highlight the core information of the temperature change, facilitating the analysis of the effect of temperature on the dissolution of the wax block.

[0049] Finally, based on the impact-temperature time series feature difference embedding compensation implicit encoding weight vector, the principal component representation implicit encoding vector of the impact time series and the principal component representation implicit encoding vector of the temperature time series are interacted collaboratively to obtain the temperature ablation-impact ablation collaborative interaction feature vector. The above process can be expressed as: ; where, is a non-linear activation function, is the impact-temperature time series feature difference embedding compensation implicit encoding weight vector, is the principal component representation implicit encoding vector of the impact time series, is the principal component representation implicit encoding vector of the temperature time series, is the element-wise multiplication by position, is the point convolution encoding, is the temperature ablation-impact ablation collaborative interaction feature vector.

[0050] It should be understood that the impact-temperature time series feature difference embedding compensation implicit coding weight vector guides the adjustment of the interaction mode according to the differences between the implicit coding vectors characterized by the impact time series principal components and the implicit coding vectors characterized by the temperature time series principal components. That is, by performing collaborative interaction on the implicit coding vector characterized by the impact time series principal components and the implicit coding vector characterized by the temperature time series principal components based on the impact-temperature time series feature difference embedding compensation implicit coding weight vector, not only can the information of the original feature differences be effectively fused, but also by introducing a compensation mechanism, the sensitivity and adaptability of the interaction result to the original feature differences are ensured. That is, the obtained temperature ablation-impact ablation collaborative interaction feature vector can provide a more scientific and accurate basis for the decision-making of the paraffin removal process. Simply put, the model can judge whether the current paraffin removal process reaches the optimal state according to the impact and temperature collaborative effect reflected by this vector. If the ideal effect is not achieved, the paraffin removal process parameters can be accurately adjusted based on the information provided by the vector.

[0051] In step 36, based on the temperature ablation-impact ablation collaborative interaction feature vector, an optimization instruction is obtained, and the optimization instruction is used to represent increasing the nozzle diameter of the high-pressure water jet head, decreasing the nozzle diameter of the high-pressure water jet head, or maintaining the nozzle diameter of the high-pressure water jet head. Specifically, in the embodiment of the present application, step 36 includes: inputting the temperature ablation-impact ablation collaborative interaction feature vector into a temperature-impact parameter optimizer based on a classifier to obtain the optimization instruction. That is, classification processing is performed on the temperature ablation-impact ablation collaborative interaction feature vector obtained by performing principal component-guided collaborative analysis using the impact time series feature implicit coding vector and the temperature time series feature implicit coding vector, so as to accurately identify the optimal nozzle diameter setting under the current operating conditions by using the powerful classification ability of the classifier, so that the high-pressure jet can better adapt to the paraffin removal requirements and improve the paraffin removal efficiency. In particular, in a specific embodiment of the present application, inputting the temperature ablation-impact ablation collaborative interaction feature vector into a temperature-impact parameter optimizer based on a classifier to obtain the optimization instruction includes: performing fully connected coding on the temperature ablation-impact ablation collaborative interaction feature vector using the fully connected layer of the classifier to obtain a temperature ablation-impact ablation collaborative interaction fully connected coding feature vector; inputting the temperature ablation-impact ablation collaborative interaction fully connected coding feature vector into the Softmax classification function of the classifier to obtain the probability values of the temperature ablation-impact ablation collaborative interaction feature vector belonging to each classification label, where the classification labels include those for representing increasing the nozzle diameter of the high-pressure water jet head, those for representing decreasing the nozzle diameter of the high-pressure water jet head, and those for representing maintaining the nozzle diameter of the high-pressure water jet head; and determining the classification label corresponding to the largest of the probability values as the optimization instruction.

[0052] Preferably, inputting the temperature ablation-impulse ablation collaborative interaction feature vector into the temperature-impulse parameter optimizer based on a classifier to obtain an optimization instruction includes: calculating the low-order proximity and high-order similarity between the eigenvalue and the eigenvalue of the temperature ablation-impulse ablation collaborative interaction feature vector, so as to obtain a temperature ablation-impulse ablation collaborative interaction low-order proximity matrix and a temperature ablation-impulse ablation collaborative interaction high-order similarity matrix: ; where and respectively represent the eigenvalue and the eigenvalue of the temperature ablation-impulse ablation collaborative interaction feature vector, represents the eigenvalue at the position of the temperature ablation-impulse ablation collaborative interaction low-order proximity matrix, represents the eigenvalue at the position of the temperature ablation-impulse ablation collaborative interaction high-order similarity matrix.

[0053] Multiplying the temperature ablation-impulse ablation collaborative interaction feature vector with the temperature ablation-impulse ablation collaborative interaction low-order proximity matrix and the temperature ablation-impulse ablation collaborative interaction high-order similarity matrix respectively to obtain a first temperature ablation-impulse ablation collaborative interaction recursive feature vector: ; where represents the temperature ablation-impulse ablation collaborative interaction low-order proximity matrix, represents the temperature ablation-impulse ablation collaborative interaction high-order similarity matrix, represents matrix multiplication, represents the temperature ablation-impulse ablation collaborative interaction feature vector, represents the first temperature ablation-impulse ablation collaborative interaction recursive feature vector.

[0054] Multiplying the temperature ablation-impulse ablation collaborative interaction feature vector with the temperature ablation-impulse ablation collaborative interaction high-order similarity matrix and the temperature ablation-impulse ablation collaborative interaction low-order proximity matrix respectively to obtain a second temperature ablation-impulse ablation collaborative interaction recursive feature vector: ; where represents the second temperature ablation-impulse ablation collaborative interaction recursive feature vector.

[0055] Multiply the temperature ablation-impulse ablation collaborative interaction low-order proximity matrix with the autocorrelation matrix of the first temperature ablation-impulse ablation collaborative interaction recursive eigenvector and the second temperature ablation-impulse ablation collaborative interaction recursive eigenvector to obtain the first temperature ablation-impulse ablation collaborative interaction sparse expansion matrix , where represents the transpose of a vector.

[0056] Multiply the temperature ablation-impulse ablation collaborative interaction high-order similarity matrix with the autocorrelation matrix of the first temperature ablation-impulse ablation collaborative interaction recursive eigenvector and the second temperature ablation-impulse ablation collaborative interaction recursive eigenvector to obtain the second temperature ablation-impulse ablation collaborative interaction sparse expansion matrix .

[0057] After performing dot addition on the first temperature ablation-impulse ablation collaborative interaction sparse expansion matrix and the second temperature ablation-impulse ablation collaborative interaction sparse expansion matrix, multiply the result with the transposed vector of the temperature ablation-impulse ablation collaborative interaction eigenvector to obtain the optimized temperature ablation-impulse ablation collaborative interaction eigenvector , where represents dot addition.

[0058] Input the optimized temperature ablation-impulse ablation collaborative interaction eigenvector into the temperature-impulse parameter optimizer based on a classifier to obtain an optimization instruction.

[0059] Here, when the impulse timing feature implicit encoding vector represents the sample sequence timing encoding features of pressure value, flow value, and jet velocity value, and the temperature timing feature implicit encoding vector represents the single-sample timing encoding feature of temperature, during the feature principal component compensation interaction, the insufficient principal component compensation correlation correspondence caused by the source timing sample space distribution difference will lead to the sparse correlation interaction of the temperature ablation-impulse ablation collaborative interaction eigenvector, thereby reducing the accuracy of the optimization instruction obtained by inputting into the temperature-impulse parameter optimizer based on a classifier due to the lack of classification inference degree.

[0060] Therefore, for the temperature ablation-impulse ablation collaborative interaction feature vector, its low-order proximity and high-order similarity matrix representation are used as multi-scale feature correlation quantization characterization. Through recursive optimization, the coupling of the time-delay auto-encoding representation of the temperature ablation-impulse ablation collaborative interaction feature vector among different correlation modes is simulated respectively, so as to simulate the sparse activation mechanism of the feature distribution neuron group, and the distributed sub-dimensions in the feature component space are used for prospective sparse constraint expansion, realizing the multi-granularity predictable expression of the temperature ablation-impulse ablation collaborative interaction feature vector, effectively avoiding the time-sequence loop attenuation problem caused by the failure of joint inference, and improving the accuracy of the optimization instruction obtained by inputting the temperature ablation-impulse ablation collaborative interaction feature vector into the temperature-impulse parameter optimizer based on the classifier.

[0061] In summary, step 3 is clearly described. It uses artificial intelligence-based data processing technology to group the high-pressure jet state parameters. Then, sequence encoding is performed on the time queue of the grouped temperature values and the time queue composed of the pressure value, flow rate value, and jet velocity value to obtain the temperature time-sequence feature and the impulse time-sequence feature. Based on this, the nozzle diameter of the high-pressure hydraulic jet head is adaptively controlled according to the ablation collaborative representation guided by the principal component between the two. In this way, by real-time collecting the state parameters of the high-pressure jet and adaptively controlling the nozzle diameter of the high-pressure hydraulic jet head, the disadvantages brought by the fixed or manually adjusted nozzle diameter in the traditional technology can be overcome, and thus the optimal paraffin removal effect can be ensured under different working conditions.

[0062] In step 4, steps 1 to 3 are continuously executed to remove the wax blocks in the oil pipe. It should be understood that the wax blocks in the oil pipe are continuously broken and melted under the continuous impact of the high-pressure jet. Each cycle can further remove the residual wax, gradually reducing the wax blocks in the oil pipe until they are completely removed. This can effectively reduce the fluid flow resistance in the oil pipe, ensure that the crude oil can flow smoothly from the oil layer to the wellhead, and thus is conducive to maintaining the stable high production of the oil well.

[0063] In summary, the hot washing paraffin removal and plugging removal technology for the hollow sucker rod well based on the embodiments of the present application is clarified. First, the well washing fluid is output by a hot washing pump truck or a high-pressure cleaner. Then, the well washing fluid is injected into the inner cavity of the hollow sucker rod through the high-pressure pipe manifold. Then, the well washing fluid flows in the circulation channel formed between the hollow sucker rod and the oil pipe and sprays out in the form of a high-pressure jet from the bottom of the hollow sucker rod. The high-pressure jet will directly act on the wax blocks in the oil pipe to help break and remove these sediments. The whole process needs to be carried out continuously to ensure that the wax blocks accumulated inside the oil pipe can be effectively removed by repeatedly circulating the well washing fluid to keep the oil pipe clean and unobstructed.

Claims

1. A hollow sucker rod hot washing wax removal and oil well blockage removal process, characterized in that: include: Step 1: Output the well washing fluid through a hot wash pump truck or a high-pressure cleaning machine; Step 2: The well-washing fluid is injected into the inner cavity of the hollow sucker rod through the high-pressure manifold, and the hollow sucker rod and the oil pipe form a circulation channel; Step 3: The well-washing fluid flows out from the bottom end of the hollow sucker rod and forms a high-pressure jet, wherein the high-pressure jet acts on the wax block in the oil pipe; Step 4: Continuously execute steps 1 to 3 to remove the wax block in the oil pipe; wherein, step 3: The well-washing fluid flows out from the bottom end of the hollow sucker rod and forms a high-pressure jet, including: using a sensor component to collect the state parameters of the high-pressure jet to obtain a time queue of the high-pressure jet state parameters, the high-pressure jet state parameters include temperature value, pressure value, flow value and jet velocity value; variable grouping of the time queue of the high-pressure jet state parameters to obtain a time queue of the temperature value, And, a time queue composed of pressure values, flow values ​​and jet velocity values; sequence encoding the time queue of temperature values ​​to obtain an implicit coding vector of temperature time series characteristics; sequence encoding the time queue composed of pressure values, flow values ​​and jet velocity values ​​to obtain an implicit coding vector of impact time series characteristics; principal component-guided temperature ablation-impact ablation collaborative analysis of the impact time series characteristic implicit coding vector and the temperature time series characteristic implicit coding vector to obtain a temperature ablation-impact ablation collaborative interaction characteristic vector; based on the temperature ablation-impact ablation collaborative interaction characteristic vector, an optimization instruction is obtained, and the optimization instruction is used to indicate increasing the nozzle diameter of the high-pressure hydraulic jet head, reducing the nozzle diameter of the high-pressure hydraulic jet head or maintaining the nozzle diameter of the high-pressure hydraulic jet head.

2. The hollow sucker rod hot washing, wax removal and oil well blockage removal process according to claim 1 is characterized in that: A high-pressure hydraulic jet head is arranged at the bottom end of the hollow sucker rod.

3. The hollow sucker rod hot washing, wax removal and oil well blockage removal process according to claim 1 is characterized in that: A sealing blowout prevention device is arranged in the circulation channel formed by the hollow sucker rod and the oil pipe.

4. The hollow sucker rod hot washing, wax removal and oil well blockage removal process according to claim 3 is characterized in that: The time queue of the temperature values ​​is sequence-encoded to obtain an implicitly encoded vector of the temperature time series feature, including: using a sequence encoder based on a bidirectional LSTM model to sequence-encode the time queue of the temperature values ​​to obtain an implicitly encoded vector of the temperature time series feature.

5. The hollow sucker rod hot washing, wax removal and oil well blockage removal process according to claim 4 is characterized in that: The time queue composed of pressure values, flow values ​​and jet velocity values ​​is sequence-encoded to obtain an impact timing feature implicit coding vector, including: using the sequence encoder based on the bidirectional LSTM model to sequence-encode the time queue composed of pressure values, flow values ​​and jet velocity values ​​to obtain the impact timing feature implicit coding vector.

6. The hollow sucker rod hot washing, wax removal and oil well blockage removal process according to claim 5, characterized in that: The method comprises: performing principal component-guided temperature ablation-shock ablation collaborative analysis on the impact timing feature implicit coding vector and the temperature timing feature implicit coding vector to obtain a temperature ablation-shock ablation collaborative interaction feature vector, including: performing principal component feature analysis on the impact timing feature implicit coding vector and the temperature timing feature implicit coding vector to obtain a set of impact timing feature principal component implicit coding vectors and a set of temperature timing feature principal component implicit coding vectors; calculating the collision coefficient between the set of impact timing feature principal component implicit coding vectors and the set of temperature timing feature principal component implicit coding vectors; The shock-temperature timing feature difference is embedded into a compensation implicit coding weight vector; the positional mean vector of the set of the shock timing feature principal component implicit coding vectors and the set of the temperature timing feature principal component implicit coding vectors is calculated to obtain the shock timing principal component characterization implicit coding vector and the temperature timing principal component characterization implicit coding vector; based on the shock-temperature timing feature difference embedding compensation implicit coding weight vector, the shock timing principal component characterization implicit coding vector and the temperature timing principal component characterization implicit coding vector are collaboratively interacted to obtain the temperature ablation-shock ablation collaborative interaction feature vector.

7. The hollow sucker rod hot washing, wax removal and oil well blockage removal process according to claim 6, characterized in that: The method comprises: calculating the shock-temperature timing feature difference embedding compensation implicit coding weight vector between the set of the shock timing feature principal component implicit coding vectors and the set of the temperature timing feature principal component implicit coding vectors, comprising: constructing the set of the shock timing feature principal component implicit coding vectors and the set of the temperature timing feature principal component implicit coding vectors into a shock timing feature principal component aggregation implicit coding feature graph and a temperature timing feature principal component aggregation implicit coding feature graph; performing difference embedding weight calculation on the shock timing feature principal component aggregation implicit coding feature graph and the temperature timing feature principal component aggregation implicit coding feature graph to obtain a shock timing feature branch weight vector and a temperature timing feature branch weight vector; and calculating the shock-temperature timing feature difference embedding compensation implicit coding weight vector between the shock timing feature branch weight vector and the temperature timing feature branch weight vector.

8. The hollow sucker rod hot washing, wax removal and plugging removal process for oil wells according to claim 7, characterized in that: Based on the temperature ablation-shock ablation collaborative interaction feature vector, an optimization instruction is obtained, including: inputting the temperature ablation-shock ablation collaborative interaction feature vector into a classifier-based temperature-shock parameter optimizer to obtain the optimization instruction.

Images