Automatic slurry treatment method and system, computer equipment and storage medium

Abstract

The invention relates to an automatic mud processing method and system, computer equipment and a storage medium. The method comprises the steps that mud parameter data are collected in real time, and the mud parameter data are cleaned to obtain standard parameter data; analyzing the standard parameter data based on a preset mud time change model, and predicting short-term dynamic change and long-term trend to obtain a data prediction result; obtaining a current operation scene, performing matching analysis on the current operation scene according to the data prediction result, and dynamically switching a corresponding operation mode; generating a plurality of processing strategies based on a data prediction result, performing analogue simulation, determining a final processing strategy according to an analogue simulation result, and further adjusting corresponding slurry processing parameters; and monitoring the slurry treatment parameters in real time, identifying abnormal factors causing an abnormal state through the causal relationship model under the condition that the abnormal state is monitored, and generating a parameter adjustment scheme for the abnormal factors. The slurry treatment device has the effect of improving the slurry treatment efficiency.

Description

Technical Field

[0001] This application relates to the technical field of mud treatment, and in particular to an automated mud treatment method, system, computer equipment, and storage medium. Background Technology

[0002] Currently, mud treatment is an indispensable technical link in marine engineering drilling operations. The properties of mud (such as density, viscosity, pressure, etc.) have an important impact on the stability and safety of well control. Automated mud treatment technology aims to improve mud treatment efficiency and reduce operational risks through real-time monitoring and optimized control.

[0003] Existing mud treatment technologies typically employ real-time monitoring and control methods based on fixed rules or single parameters. Mud performance prediction is primarily based on static models, which struggle to fully consider the dynamic changes and complex time-dependent characteristics of mud parameters. Furthermore, in terms of anomaly identification and handling, current technologies often lack in-depth analysis of the correlations and causal relationships between mud parameters, only allowing for simple adjustments to surface anomalies, making it difficult to achieve accurate and efficient anomaly localization and parameter optimization.

[0004] The existing technical solutions mentioned above have the following drawbacks: the existing mud treatment methods are difficult to fully capture the dynamic changes and abnormal states of mud parameters, resulting in low efficiency in mud treatment, and therefore there is room for improvement. Summary of the Invention

[0005] To improve the efficiency of mud treatment, this application provides an automated mud treatment method, system, computer equipment, and storage medium.

[0006] The above-mentioned objective of this application is achieved through the following technical solution: An automated mud treatment method, the automated mud treatment method comprising: Real-time acquisition of mud parameter data, followed by cleaning and standardization of the mud parameter data to obtain standard parameter data; The standard parameter data are analyzed based on a preset mud time variation model to predict the short-term dynamic changes and long-term trends of the mud parameter data, and the data prediction results are obtained. The current work scenario is obtained, and the current work scenario is matched and analyzed based on the data prediction results. The corresponding operation mode is then dynamically switched based on the matching and analysis results. Multiple processing strategies are generated based on the data prediction results, and simulations are performed on the multiple strategies. The final processing strategy is determined based on the simulation results, and the corresponding mud processing parameters are adjusted according to the final processing strategy. The mud treatment parameters are monitored in real time, and when an abnormal state is detected, the abnormal factors causing the abnormal state are identified through a causal relationship model, and parameter adjustment schemes are generated for the abnormal factors.

[0007] By adopting the above technical solutions, real-time acquisition of mud parameter data, followed by cleaning and standardization, effectively eliminates noise and outliers in the mud parameters, improving the accuracy and reliability of subsequent analysis and ensuring that subsequent time-varying analysis is based on high-quality data. Analyzing standard parameter data based on a pre-defined mud time-varying model allows for the prediction of short-term dynamic changes and long-term trends in mud parameters, accurately capturing dynamic change patterns and identifying potential anomalies in advance. This provides precise decision-making support for adjusting operating modes and generating treatment strategies. By acquiring the current operational scenario, matching and analyzing the scenario based on data prediction results, and dynamically switching the corresponding operating mode based on the matching analysis results, the system can quickly respond to the needs of different operational scenarios and dynamically adjust its operating logic, thereby improving the flexibility and safety of the mud treatment process. Finally, by generating multiple treatment strategies based on data prediction results and simulating these strategies, the final treatment strategy can be determined based on the simulation results. This allows for a comprehensive evaluation of mud parameter control strategies, selecting the optimal treatment scheme, and achieving precise optimization of mud performance and efficient resource utilization.

[0008] In one example, this application can be further configured as follows: the construction of the preset mud time variation model specifically includes: Acquire historical operation data, extract time series features of mud parameters from the historical operation data, and analyze the corresponding short-term dynamic features and long-term trend features based on the time series features; Based on the aforementioned short-term dynamic characteristics, a long short-term memory network is used to train the time series data to establish a short-term time change model. Based on the aforementioned long-term trend characteristics, a Transformer model incorporating a global attention mechanism is used to train the time series data to establish a long-term time variation model. The short-term time variation model and the long-term time variation model are integrated to construct the complete mud time variation model.

[0009] By adopting the above technical solutions, historical operational data is acquired, and time-series features of mud parameters are extracted from this data. Based on these time-series features, corresponding short-term dynamic characteristics and long-term trend characteristics are analyzed. This allows for a comprehensive extraction of historical variation patterns of mud parameters, providing reliable data support for model training and improving the training accuracy of the time-varying model. By using a Long Short-Term Memory (LSTM) network to train the time-series data based on short-term dynamic features, short-term variation patterns of mud parameters can be accurately captured, improving the accuracy and timeliness of short-term predictions. By using a Transformer model incorporating a global attention mechanism to train the time-series data based on long-term trend characteristics, long-term trend changes and global correlations of mud parameters can be effectively captured, improving the accuracy of long-term predictions and adaptability to complex working conditions. Finally, by integrating the short-term and long-term time-varying models, a complete mud time-varying model is constructed, comprehensively considering both short-term dynamic changes and long-term trend changes, enhancing the model's comprehensive predictive ability for mud parameters, and providing more reliable decision support for mud treatment.

[0010] In one example, this application can be further configured as follows: the analysis of the standard parameter data based on a preset mud time variation model, predicting the short-term dynamic changes and long-term trends of the mud parameter data, and obtaining data prediction results, specifically includes: Based on the short-term time variation model, short-term time series analysis is performed on the standard parameter data to predict the short-term variation trend of the mud parameter data; Based on the long-term time variation model, long-term trend analysis is performed on the standard parameter data to predict the long-term variation trend of the mud parameter data; The data prediction results are obtained based on the short-term and long-term trends of mud parameter data.

[0011] By adopting the above technical solutions, and through short-term time series analysis of standard parameter data based on short-term time variation models, the short-term trend of mud parameter data can be predicted. This allows for the rapid capture of sudden changes in mud parameters within a short period, providing an accurate basis for real-time adjustments. Furthermore, by performing long-term trend analysis of standard parameter data based on long-term time variation models, the long-term trend of mud parameter data can be predicted. This allows for an accurate grasp of the overall development trend of mud parameters, providing guidance for planned adjustments in the mud treatment process and effectively reducing the risk of parameter fluctuations under long-term operating conditions. Finally, by obtaining data prediction results based on the short-term and long-term trends of mud parameter data, short-term and long-term change information can be integrated, providing a more comprehensive decision-making basis for subsequent scenario matching and processing strategy generation, thereby improving the response speed and predictive capability of the mud treatment system.

[0012] In one example, this application can be further configured as follows: the step of performing a matching analysis on the current work scenario based on the data prediction results, and dynamically switching the corresponding operating mode based on the matching analysis results, specifically includes: If the matching analysis result indicates that the current operation scenario is within the safe threshold range, switch to normal operation mode and adjust the mud treatment parameters using the standard control frequency; If the matching analysis result indicates that the current working scenario is within the safety warning threshold range, switch to high-pressure mode, increase the data prediction frequency, and prioritize the adjustment of key mud parameters. If the matching analysis result indicates that the current operation scenario exceeds the safety warning threshold range, switch to emergency mode, adjust the mud treatment parameters according to the preset safety adjustment strategy, and trigger a safety alarm.

[0013] By adopting the above technical solutions, when the matching analysis results indicate that the current operating scenario is within the safe threshold range, switching to normal operation mode can maintain the stability of mud parameters and reduce unnecessary resource consumption, thereby improving the operating efficiency of the mud treatment process. When the matching analysis results indicate that the current operating scenario is within the safety warning threshold range, switching to high-pressure mode can quickly respond to impending abnormal parameters and prioritize the control of key mud parameters, thereby preventing potential operational risks from escalating further. When the matching analysis results indicate that the current operating scenario exceeds the safety warning threshold range, switching to emergency mode can adjust mud treatment parameters according to a preset safety adjustment strategy and trigger a safety alarm, enabling rapid emergency measures to be taken to reduce safety hazards caused by abnormal conditions and timely reminding operators to intervene manually, thereby ensuring the safety and continuity of drilling operations.

[0014] In one example, this application can be further configured as follows: Simulating multiple strategies and determining the final processing strategy based on the simulation results specifically includes: A multi-objective optimization algorithm is used to simulate multiple processing strategies. Based on the data prediction results and the introduction of external environmental changes, dynamic balancing is performed, and the weight of each processing strategy is calculated to generate a comprehensive score. The processing strategy with the highest overall score is selected as the final processing strategy.

[0015] By adopting the above technical solution and employing a multi-objective optimization algorithm, multiple treatment strategies are simulated. Based on data prediction results and incorporating changes in the external environment, dynamic balancing is performed, and the weight of each treatment strategy is calculated to generate a comprehensive score. This allows for a comprehensive evaluation of the advantages and disadvantages of different treatment strategies, especially considering multi-dimensional influencing factors in complex environments, thereby ensuring that the selection of treatment strategies is more scientific and reasonable. By selecting the treatment strategy with the highest comprehensive score as the final treatment strategy, the strategy with the highest resource utilization and best control effect can be quickly selected from many alternatives, thereby improving the economy and accuracy of mud treatment.

[0016] In one example, this application can be further configured as follows: upon detecting an abnormal state, identifying the abnormal factors causing the abnormal state through a causal relationship model, and generating a parameter adjustment scheme for the abnormal factors, specifically including: Based on historical mud parameter data, a causal relationship model between mud parameters is established using a Bayesian network. If the abnormal state is detected, the change chain of the mud parameter data is analyzed based on the causal relationship model to determine the abnormal factors that cause the abnormal state. When the abnormal factor is a equipment malfunction, adjust the corresponding pump pressure parameters to reduce pressure fluctuations. When the abnormal factor is a change in the external environment, increase the amount of mud weighting agent injected to stabilize the mud density.

[0017] By adopting the above technical solutions and establishing a causal relationship model between mud parameters based on historical mud parameter data and a Bayesian network, the inherent causal relationships between mud parameters can be uncovered, forming a knowledge network that can guide anomaly analysis and thus providing a scientific basis for the accurate location of abnormal states. When an abnormal state is detected, the change chain of mud parameter data can be analyzed based on the causal relationship model to identify the abnormal factors causing the abnormal state, enabling rapid inference of the root cause of the abnormal state and avoiding unnecessary resource waste caused by blind adjustments. When the abnormal factor is equipment malfunction, the corresponding pump pressure parameters can be adjusted to reduce pressure fluctuations; when the abnormal factor is external environmental changes, the amount of mud weighting agent injected can be increased to stabilize mud density. Targeted adjustment plans can be generated according to different abnormal factors, thereby efficiently restoring the stability of mud parameters and ensuring the continuity and safety of drilling operations.

[0018] The second objective of this invention is achieved through the following technical solution: An automated mud treatment system, characterized in that the automated mud treatment system comprises: The data acquisition and processing module is used to acquire mud parameter data in real time, and to clean and standardize the mud parameter data to obtain standard parameter data. The time variation analysis module is used to analyze the standard parameter data based on a preset mud time variation model, predict the short-term dynamic changes and long-term trends of the mud parameter data, and obtain data prediction results. The job scenario matching module is used to obtain the current job scenario, perform matching analysis on the current job scenario based on the data prediction results, and dynamically switch the corresponding operation mode based on the matching analysis results. The strategy generation and optimization module is used to generate multiple processing strategies based on the data prediction results, simulate the multiple strategies, determine the final processing strategy based on the simulation results, and then adjust the corresponding mud treatment parameters according to the final processing strategy. The anomaly detection and adjustment module is used to monitor the mud treatment parameters in real time, and when an abnormal state is detected, it identifies the abnormal factors that cause the abnormal state through a causal relationship model and generates a parameter adjustment scheme for the abnormal factors.

[0019] By adopting the above technical solutions, real-time acquisition of mud parameter data, followed by cleaning and standardization, effectively eliminates noise and outliers in the mud parameters, improving the accuracy and reliability of subsequent analysis and ensuring that subsequent time-varying analysis is based on high-quality data. Analyzing standard parameter data based on a pre-defined mud time-varying model allows for the prediction of short-term dynamic changes and long-term trends in mud parameters, accurately capturing dynamic change patterns and identifying potential anomalies in advance. This provides precise decision-making support for adjusting operating modes and generating treatment strategies. By acquiring the current operational scenario, matching and analyzing the scenario based on data prediction results, and dynamically switching the corresponding operating mode based on the matching analysis results, the system can quickly respond to the needs of different operational scenarios and dynamically adjust its operating logic, thereby improving the flexibility and safety of the mud treatment process. Finally, by generating multiple treatment strategies based on data prediction results and simulating these strategies, the final treatment strategy can be determined based on the simulation results. This allows for a comprehensive evaluation of mud parameter control strategies, selecting the optimal treatment scheme, and achieving precise optimization of mud performance and efficient resource utilization.

[0020] The above-mentioned objective three of this application is achieved through the following technical solution: A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the automated mud treatment method described above.

[0021] The fourth objective of this application is achieved through the following technical solution: A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described automated mud treatment method.

[0022] In summary, this application includes the following beneficial technical effects: 1. By collecting and cleaning mud parameter data in real time, noise and outliers can be effectively eliminated, improving the accuracy and reliability of subsequent analysis and ensuring that subsequent time-varying analysis is based on high-quality data. 2. By analyzing standard parameter data based on a preset mud time-varying model, short-term dynamic changes and long-term trends of mud parameters can be predicted, accurately capturing the dynamic change patterns of mud parameters and identifying potential abnormal trends in advance. This provides precise decision-making basis for adjusting operating modes and generating treatment strategies. 3. By acquiring the current operating scenario, matching and analyzing the current operating scenario based on data prediction results, and dynamically switching the corresponding operating mode based on the matching analysis results, the system can quickly respond to the needs of different operating scenarios and dynamically adjust the system's operating logic, thereby improving the flexibility and safety of the mud treatment process. 4. By generating multiple treatment strategies based on data prediction results and simulating these strategies, the final treatment strategy can be determined based on the simulation results. This allows for a comprehensive evaluation of mud parameter control strategies, selecting the optimal treatment scheme, and achieving precise optimization of mud performance and efficient resource utilization. 2. By acquiring historical operational data, time-series features of mud parameters are extracted. Based on these features, short-term dynamic characteristics and long-term trend characteristics are analyzed, comprehensively extracting historical variation patterns of mud parameters and providing reliable data support for model training, thereby improving the training accuracy of the time-varying model. By using a Long Short-Term Memory (LSTM) network to train the time-series data based on short-term dynamic features, short-term variation patterns of mud parameters can be accurately captured, improving the accuracy and timeliness of short-term predictions. By using a Transformer model incorporating a global attention mechanism to train the time-series data based on long-term trend characteristics, long-term trend changes and global correlations of mud parameters can be effectively captured, improving the accuracy of long-term predictions and adaptability to complex working conditions. By integrating short-term and long-term time-varying models, a complete mud time-varying model is constructed, comprehensively considering both short-term dynamic changes and long-term trend changes, enhancing the model's comprehensive predictive ability for mud parameters, and thus providing more reliable decision support for mud treatment. 3. By performing short-term time series analysis on standard parameter data based on short-term time variation models, the short-term trend of mud parameter data can be predicted, quickly capturing sudden changes in mud parameters within a short period, thus providing an accurate basis for real-time adjustments. By performing long-term trend analysis on standard parameter data based on long-term time variation models, the long-term trend of mud parameter data can be predicted, accurately grasping the overall development trend of mud parameters, providing guidance for planned adjustments in the mud treatment process, and effectively reducing the risk of parameter fluctuations under long-term operating conditions. By obtaining data prediction results based on the short-term and long-term trends of mud parameter data, short-term and long-term change information can be integrated, providing a more comprehensive decision-making basis for subsequent scenario matching and treatment strategy generation, thereby improving the response speed and predictive capability of the mud treatment system. Attached Figure Description

[0023] Figure 1 This is a flowchart of an automated mud treatment method according to an embodiment of this application; Figure 2 This is a flowchart illustrating the implementation of a preset mud time variation model in an automated mud treatment method according to an embodiment of this application. Figure 3 This is a flowchart illustrating the implementation of step S20 in an automated mud treatment method according to an embodiment of this application. Figure 4 This is a flowchart illustrating the implementation of step S30 in an automated mud treatment method according to an embodiment of this application. Figure 5 This is a flowchart illustrating the implementation of step S40 in an automated mud treatment method according to an embodiment of this application. Figure 6 This is a flowchart illustrating the implementation of step S50 in an automated mud treatment method according to an embodiment of this application. Figure 7 This is a schematic diagram of an automated mud treatment system according to one embodiment of this application; Figure 8 This is a schematic diagram of a device according to one embodiment of this application. Detailed Implementation

[0024] The present application will be further described in detail below with reference to the accompanying drawings.

[0025] In one embodiment, such as Figure 1 As shown, this application discloses an automated mud treatment method, which specifically includes the following steps: S10: Collect mud parameter data in real time, and clean and standardize the mud parameter data to obtain standard parameter data.

[0026] Specifically, real-time operational data of the mud is collected by sensors, including raw parameters such as density, viscosity, flow rate, and pressure. The collected data is first processed for missing values. For short-term data gaps, linear interpolation is used to fill in the missing data, while for long-term data gaps, regression prediction is used to fill in the missing data based on historical trends. At the same time, outliers in the data are identified using statistical methods, such as using box plots or the three-standard-deviation method to remove outliers. Then, local weighted regression is used to replace outliers to avoid interference from outliers to the model. Finally, all parameters are normalized to a fixed range, such as [0,1], to eliminate the impact of differences in different parameter dimensions on subsequent calculations.

[0027] S20: Analyze standard parameter data based on a preset mud time variation model, predict the short-term dynamic changes and long-term trends of mud parameter data, and obtain data prediction results.

[0028] Specifically, based on the preset mud time variation model, standard parameter data are input into the short-term prediction module and the long-term prediction module. The short-term prediction module is responsible for predicting the trend of mud parameter changes within the next 30 seconds to 1 minute. For example, when historical data on density and flow velocity show continuous fluctuations in a short period of time, the short-term prediction module can capture the trend of an impending increase in density in advance through time series analysis of dynamic windows. The long-term prediction module, based on historical data with a larger time span, such as the trend of changes over 1 hour, combined with the periodic fluctuation characteristics of mud parameters, can derive the overall trend of mud performance changes within the next 1 hour through gradual fitting and trend line extraction of viscosity and pressure. Finally, the outputs of the short-term prediction module and the long-term prediction module are integrated and processed to generate a prediction data result that fuses short-term and long-term trends.

[0029] S30: Obtain the current work scenario, perform matching analysis on the current work scenario based on the data prediction results, and dynamically switch the corresponding operation mode based on the matching analysis results.

[0030] Specifically, based on the currently collected operational scenario parameters, such as formation pressure, mud flow rate, and density, the data prediction results are compared and analyzed. When the parameter values ​​of the current operational scenario match the predicted trend and are within a safe range, it is matched as a normal operational scenario. When the predicted trend shows that the mud density or pressure is close to the safety warning threshold, such as when the density is close to the maximum tolerance range but not exceeded, it is matched as a high-pressure scenario. When the predicted trend shows that the parameters have exceeded the safety threshold, such as when the pressure exceeds the safe range and causes a blowout risk, it is matched as an emergency scenario. Subsequently, the operating mode is switched based on the matching results. For example, when switching to a high-pressure scenario, the frequency of mud parameter acquisition is increased, and density and pressure-related parameters are adjusted first to balance well control pressure. In an emergency scenario, a preset safety adjustment strategy is executed and an alarm is triggered to alert the operators.

[0031] S40: Generate multiple processing strategies based on the data prediction results, simulate the multiple strategies, determine the final processing strategy based on the simulation results, and then adjust the corresponding mud processing parameters according to the final processing strategy.

[0032] Specifically, multiple treatment strategies are generated based on the trends in the data prediction results. For example, when the prediction shows that high mud density may lead to a blowout risk, the following treatment strategies can be generated: reduce pump pressure to reduce mud circulation speed, increase the proportion of diluent to reduce density, or a composite strategy of simultaneously reducing pump pressure and adjusting the proportion of diluent and weighting agent. Then, dynamic balance analysis is performed on each strategy through simulation. For example, after introducing the external environmental factor of formation pressure fluctuation, the simulation obtains the specific control effect of each strategy on density and pressure, while evaluating resource consumption and time cost. Based on the comprehensive score, the strategy with the best density control effect and the least consumption is selected as the final treatment strategy, and the actual parameter values ​​of pump pressure, diluent or weighting agent are adjusted according to the selected strategy.

[0033] S50: Real-time monitoring of mud treatment parameters, and in the event of an abnormal state, identification of the abnormal factors causing the abnormal state through a causal relationship model, and generation of parameter adjustment schemes for the abnormal factors.

[0034] Specifically, the real-time monitored mud parameters are compared with the normal parameter range. When a parameter exceeds the normal range, such as a sudden decrease in flow rate or an abnormal increase in density, an abnormal state detection is triggered. The monitored abnormal data is input into the causal relationship model. Through causal chain analysis, the key factors of the abnormality are identified. For example, a decrease in flow rate may be caused by low pump pressure or a sudden change in formation pressure, and an increase in density may be related to excessive weighting agent injection. Subsequently, corresponding parameter adjustment schemes are generated based on the identified abnormal factors. For example, when the factor is insufficient pump pressure, the pump pressure is increased to the normal range; when the factor is external pressure change, the pressure is balanced by increasing the proportion of weighting agent.

[0035] In one embodiment, such as Figure 2 As shown, step S20, namely the construction of the preset mud time variation model, specifically includes: S201: Obtain historical operation data, extract time series features of mud parameters from the historical operation data, and analyze the corresponding short-term dynamic features and long-term trend features based on the time series features.

[0036] Specifically, time series data of parameters such as density, flow rate, pressure and viscosity of mud are extracted from historical data records. The sliding window technique is used to segment the time series for analysis to extract short-term dynamic features, such as the frequency and amplitude of density fluctuations within 1 minute. At the same time, Fourier transform is used to perform spectral analysis on the overall trend of the time series data to extract long-term periodic change features, such as the high and low peak interval trend of mud pressure. Finally, the short-term dynamic features and long-term trend features are combined to provide input data for model training.

[0037] S202: Based on short-term dynamic characteristics, a long short-term memory network is used to train time series data to establish a short-term time change model.

[0038] Specifically, the extracted short-term dynamic features are input into a long short-term memory network. By utilizing its ability to model short-term dependencies in time series, a multi-layer recurrent network is used to capture the abrupt changes in mud parameters in a short period of time, such as the rate and time interval of short-term density increases. The model dynamically adjusts the weights during training to optimize prediction accuracy, and finally outputs a short-term time change model that can accurately predict changes in mud parameters in the next 30 seconds to 1 minute.

[0039] S203: Based on long-term trend characteristics, a Transformer model combined with a global attention mechanism is used to train time series data to establish a long-term time variation model.

[0040] Specifically, the extracted long-term trend features are input into the Transformer model, and its self-attention mechanism is used to assign weights to key change points of mud parameters over a long period of time. For example, the periodic peak changes of parameters such as density and pressure are given higher weights, so as to more accurately capture the long-term trend of mud parameters. During the training process, the multi-head attention mechanism is combined to realize the global correlation model between different parameters, and finally outputs a long-term time change model that can predict the changes of mud parameters in the next hour.

[0041] S204: Integrate short-term and long-term time variation models to construct a complete mud time variation model.

[0042] Specifically, the prediction results of the short-term time variation model and the long-term time variation model are weighted and fused. The short-term prediction results are given a higher weight to meet the needs of immediate adjustment, while the long-term prediction results are given a lower weight to provide trend reference. At the same time, the prediction results of the two types of models are integrated by weighted averaging. The prediction deviations that may occur during the fusion process are dynamically corrected to ensure that the fused time variation model can meet the needs of both short-term dynamic response and long-term trend analysis. Finally, a complete mud time variation model is constructed.

[0043] In one embodiment, such as Figure 3 As shown, in step S20, the standard parameter data is analyzed based on the preset mud time variation model to predict the short-term dynamic changes and long-term trends of the mud parameter data, and the data prediction results are obtained. Specifically, this includes: S21: Based on the short-term time variation model, perform short-term time series analysis on the standard parameter data to predict the short-term variation trend of mud parameter data.

[0044] Specifically, the standard parameter data acquired in real time is input into the short-term time variation model. Through recursive calculation of LSTM, the rate and magnitude of change of the current mud parameters in a short period of time are analyzed. For example, when the density continues to rise, the model can predict the peak position that the density will reach in the next 30 seconds. Then, based on the prediction results, the short-term change trend curve is output to provide a basis for rapid adjustment.

[0045] S22: Based on the long-term time variation model, perform long-term trend analysis on the standard parameter data to predict the long-term variation trend of mud parameter data.

[0046] Specifically, standard parameter data is input into a long-term time variation model. Through the global attention mechanism in the Transformer model, the periodic variation characteristics of mud parameters over past time periods are captured. For example, the gradual upward trend of pressure is extended and predicted to obtain the highest pressure value that may be reached in the next hour. At the same time, the possible location of future trough points is identified by combining the periodic variation pattern, and a long-term trend prediction curve is output to provide a long-term reference for operation planning and adjustment strategies.

[0047] S23: Based on the short-term and long-term trends of mud parameter data, obtain data prediction results.

[0048] Specifically, short-term and long-term trends are assigned different weights, and the fusion ratio of short-term and long-term prediction results is adjusted according to the needs of the operation scenario. For example, in a high-pressure operation environment, the weight of short-term trends can be increased to cope with rapid changes, while in a relatively stable environment, the weight of long-term trends is increased to provide a more comprehensive reference. Finally, the trend fusion algorithm outputs comprehensive data prediction results to support subsequent scenario matching and strategy adjustment.

[0049] In one embodiment, such as Figure 4 As shown, in step S30, the current work scenario is matched and analyzed based on the data prediction results, and the corresponding operating mode is dynamically switched according to the matching analysis results. This specifically includes: S31: If the matching analysis result indicates that the current operation scenario is within the safe threshold range, switch to normal operation mode and adjust the mud treatment parameters using the standard control frequency.

[0050] Specifically, when the matching analysis results show that parameters such as mud density and pressure are within the set safe range, such as when the density is maintained at the middle value of the preset range, the operation mode is set to normal mode. At the same time, the mud treatment parameters are periodically sampled and adjusted according to the standard frequency. For example, the diluent ratio is adjusted once every 1 minute to maintain the stability of the density, and the pump pressure is finely adjusted to keep the parameters running smoothly when there are slight fluctuations in density and pressure.

[0051] S32: If the matching analysis results indicate that the current operation scenario is within the safety warning threshold range, switch to high-pressure mode, increase the data prediction frequency, and prioritize the adjustment of key mud parameters.

[0052] Specifically, when the matching analysis results show that certain parameters are close to the safety threshold, such as when the pressure rises close to the maximum safe tolerance range, the operating mode is switched to high pressure mode. At the same time, the sampling frequency of mud parameters is increased, such as shortening the data sampling interval to 10 seconds. The changing trends of pressure and density are monitored in detail, and parameters that affect pressure are adjusted first, such as by reducing the injection speed of weighting agent or increasing the proportion of diluent to control density, thereby stabilizing the pressure within the safe range.

[0053] S33: If the matching analysis result indicates that the current operation scenario exceeds the safety warning threshold range, switch to emergency mode, adjust the mud treatment parameters according to the preset safety adjustment strategy, and trigger a safety alarm.

[0054] Specifically, when the matching analysis results show that the mud parameters have exceeded the safety threshold, such as a significant increase in density leading to a sharp increase in pressure, the system switches to emergency mode and rapidly reduces the density by executing preset safety adjustment strategies. For example, it immediately stops the injection of weighting agent and increases the proportion of diluent, while reducing the pump pressure to slow down the mud circulation speed. The system also notifies the operators by triggering audible and visual alarms. The alarm content includes the specific values ​​of the parameters that exceed the standard and suggested manual intervention measures to ensure operational safety.

[0055] In one embodiment, such as Figure 5 As shown, in step S40, multiple strategies are simulated, and the final processing strategy is determined based on the simulation results. Specifically, this includes: S41: Employ a multi-objective optimization algorithm to simulate multiple processing strategies, dynamically balance them based on data prediction results and external environmental changes, calculate the weight of each processing strategy, and generate a comprehensive score.

[0056] Specifically, multiple candidate processing strategies are generated based on the data prediction results, such as adjusting pump pressure and changing the ratio of diluent to weighting agent. These strategies are then input into a multi-objective optimization algorithm for simulation. External environmental factors, such as fluctuations in formation pressure or changes in equipment operating status, are introduced into the simulation. The optimization algorithm calculates the balance weight of each strategy in terms of stabilizing density, controlling pressure, and reducing resource consumption. Finally, each strategy is comprehensively scored, and the scoring results are used to select the optimal strategy.

[0057] S42: Select the processing strategy with the highest overall score as the final processing strategy.

[0058] Specifically, the scoring results of the multi-objective optimization algorithm are sorted, and the strategy with the highest comprehensive score is selected as the final processing strategy. For example, if the strategy with the highest score in the simulation is to simultaneously reduce the pump pressure and the diluent ratio to control density and pressure, then this strategy will be applied to the mud treatment operation. Subsequently, relevant parameters are adjusted according to this strategy and the control effect is monitored in real time to ensure that the treatment results meet the operational requirements.

[0059] In one embodiment, such as Figure 6 As shown, in step S50, when an abnormal state is detected, the abnormal factors causing the abnormal state are identified through a causal relationship model, and a parameter adjustment scheme is generated for the abnormal factors, specifically including: S51: Based on historical mud parameter data, establish a causal relationship model between mud parameters using a Bayesian network.

[0060] Specifically, a Bayesian network is constructed using historical mud parameter data. By mining the statistical correlations and causal relationships between parameters such as density, flow velocity, and pressure, for example, the direct relationship between pressure changes and density fluctuations is determined through the model. At the same time, the causal strength between each parameter is quantified, and finally a causal relationship model that can infer the source of anomalies is formed to support fault analysis and adjustment decisions under abnormal conditions.

[0061] S52: When an abnormal state is detected, analyze the change chain of mud parameter data based on the causal relationship model to determine the abnormal factors that lead to the abnormal state.

[0062] Specifically, the abnormal data detected is input into a Bayesian network, and the change chain of the abnormal data is analyzed through causal reasoning. For example, when the pressure rises abnormally, the model may identify that the density exceeding the standard is the direct cause, and the root cause of the density exceeding the standard is that the weighting agent is injected too quickly. Finally, the abnormal factor is determined to be the excessive injection of weighting agent, and a basis is provided for subsequent parameter adjustment.

[0063] S53: When the abnormal factor is equipment malfunction, adjust the corresponding pump pressure parameters to reduce pressure fluctuations. When the abnormal factor is external environmental change, increase the amount of mud weighting agent injected to stabilize mud density.

[0064] Specifically, when the causal relationship model analysis shows that the abnormal factor is equipment malfunction, such as insufficient pump pressure leading to a decrease in flow rate, the pump pressure parameters are adjusted first to reduce pressure fluctuations; when the abnormal factor is external environmental change, such as a sudden increase in formation pressure leading to a decrease in density, the amount of weighting agent injected is increased to quickly increase mud density and ensure that well control safety is restored to normal.

[0065] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0066] In one embodiment, an automated mud treatment system is provided, which corresponds one-to-one with the automated mud treatment methods described in the above embodiments. For example... Figure 7 As shown, the automated mud treatment system includes a data acquisition and processing module, a time variation analysis module, a work scenario matching module, and a strategy generation and optimization module. Detailed descriptions of each functional module are as follows: The data acquisition and processing module is used to acquire mud parameter data in real time, and to clean and standardize the mud parameter data to obtain standard parameter data. The time variation analysis module is used to analyze standard parameter data based on a preset mud time variation model, predict the short-term dynamic changes and long-term trends of mud parameter data, and obtain data prediction results. The job scenario matching module is used to obtain the current job scenario, perform matching analysis on the current job scenario based on the data prediction results, and dynamically switch the corresponding operation mode based on the matching analysis results. The strategy generation and optimization module is used to generate multiple processing strategies based on data prediction results, simulate multiple strategies, determine the final processing strategy based on the simulation results, and then adjust the corresponding mud processing parameters according to the final processing strategy. The anomaly detection and adjustment module is used to monitor mud treatment parameters in real time. When an abnormal state is detected, it identifies the abnormal factors that cause the abnormal state through a causal relationship model and generates parameter adjustment schemes for the abnormal factors.

[0067] Optionally, the construction of a preset mud time variation model includes: The historical data extraction module is used to acquire historical operation data, extract time series features of mud parameters from the historical operation data, and analyze the corresponding short-term dynamic features and long-term trend features based on the time series features. The short-term model training module is used to train time series data based on short-term dynamic features using a long short-term memory network to establish a short-term time change model. The long-term model training module is used to train a Transformer model that incorporates a global attention mechanism on time series data based on long-term trend features, thereby establishing a long-term time variation model. The model integration module is used to integrate short-term and long-term time variation models to construct a complete mud time variation model.

[0068] Optionally, the time variation analysis module specifically includes: The short-term analysis submodule is used to perform short-term time series analysis on standard parameter data based on a short-term time change model, and to predict the short-term change trend of mud parameter data. The long-term analysis submodule is used to perform long-term trend analysis on standard parameter data based on a long-term time variation model, and to predict the long-term variation trend of mud parameter data. The trend fusion submodule is used to obtain data prediction results based on the short-term and long-term trends of mud parameter data.

[0069] Optionally, the job scenario matching module specifically includes: The normal mode switching submodule is used to switch to normal operation mode when the matching analysis result shows that the current operation scenario is within the safe threshold range, and to adjust the mud treatment parameters using the standard control frequency. The high-pressure mode switching submodule is used to switch to high-pressure mode when the matching analysis results indicate that the current operation scenario is within the safety warning threshold range, thereby increasing the data prediction frequency and prioritizing the control of key mud parameters. The emergency mode switching submodule is used to switch to emergency mode when the matching analysis result shows that the current operation scenario exceeds the safety warning threshold range. It adjusts the mud treatment parameters according to the preset safety adjustment strategy and triggers a safety alarm.

[0070] Optionally, the strategy generation and optimization module specifically includes: The simulation optimization submodule is used to simulate multiple processing strategies using a multi-objective optimization algorithm, dynamically balance them based on data prediction results and external environmental changes, calculate the weight of each processing strategy, and generate a comprehensive score. The strategy selection submodule is used to select the processing strategy with the highest overall score as the final processing strategy.

[0071] Optionally, the anomaly detection and adjustment module specifically includes: The causal relationship modeling submodule is used to establish a causal relationship model between mud parameters based on historical mud parameter data and through a Bayesian network. The abnormal factor identification submodule is used to analyze the change chain of mud parameter data based on the causal relationship model when an abnormal state is detected, and to identify the abnormal factors that cause the abnormal state. The adjustment scheme generation submodule is used to adjust the corresponding pump pressure parameters to reduce pressure fluctuations when the abnormal factor is equipment malfunction, and to increase the amount of mud weighting agent injected to stabilize mud density when the abnormal factor is external environmental change.

[0072] Specific limitations regarding automated mud treatment systems can be found in the limitations of automated mud treatment methods described above, and will not be repeated here. Each module in the aforementioned automated mud treatment system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0073] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 8 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements an automated mud treatment method.

[0074] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the following steps: Real-time acquisition of mud parameter data, followed by cleaning and standardization of the mud parameter data to obtain standard parameter data; Based on the preset mud time variation model, the standard parameter data are analyzed to predict the short-term dynamic changes and long-term trends of mud parameter data, and the data prediction results are obtained. The system acquires the current work scenario, performs matching analysis on the current work scenario based on the data prediction results, and dynamically switches the corresponding operating mode based on the matching analysis results. Multiple processing strategies are generated based on the data prediction results, and simulations are performed on these strategies. The final processing strategy is determined based on the simulation results, and the corresponding mud processing parameters are adjusted accordingly. Real-time monitoring of mud treatment parameters is performed, and when abnormal conditions are detected, the abnormal factors causing the abnormal conditions are identified through a causal relationship model, and parameter adjustment schemes are generated for the abnormal factors.

[0075] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor: Real-time acquisition of mud parameter data, followed by cleaning and standardization of the mud parameter data to obtain standard parameter data; Based on the preset mud time variation model, the standard parameter data are analyzed to predict the short-term dynamic changes and long-term trends of mud parameter data, and the data prediction results are obtained. The system acquires the current work scenario, performs matching analysis on the current work scenario based on the data prediction results, and dynamically switches the corresponding operating mode based on the matching analysis results. Multiple processing strategies are generated based on the data prediction results, and simulations are performed on these strategies. The final processing strategy is determined based on the simulation results, and the corresponding mud processing parameters are adjusted accordingly. Real-time monitoring of mud treatment parameters is performed, and when abnormal conditions are detected, the abnormal factors causing the abnormal conditions are identified through a causal relationship model, and parameter adjustment schemes are generated for the abnormal factors.

[0076] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0077] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above.

[0078] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. An automated mud treatment method, characterized in that, The automated mud treatment method includes: Real-time acquisition of mud parameter data, followed by cleaning and standardization of the mud parameter data to obtain standard parameter data; The standard parameter data are analyzed based on a preset mud time variation model to predict the short-term dynamic changes and long-term trends of the mud parameter data, and the data prediction results are obtained. The current work scenario is obtained, and the current work scenario is matched and analyzed based on the data prediction results. The corresponding operation mode is then dynamically switched based on the matching and analysis results. Multiple processing strategies are generated based on the data prediction results, and simulations are performed on the multiple strategies. The final processing strategy is determined based on the simulation results, and the corresponding mud processing parameters are adjusted according to the final processing strategy. The mud treatment parameters are monitored in real time, and when an abnormal state is detected, the abnormal factors causing the abnormal state are identified through a causal relationship model, and parameter adjustment schemes are generated for the abnormal factors.

2. The automated mud treatment method according to claim 1, characterized in that, The construction of the preset mud time variation model specifically includes: Acquire historical operation data, extract time series features of mud parameters from the historical operation data, and analyze the corresponding short-term dynamic features and long-term trend features based on the time series features; Based on the aforementioned short-term dynamic characteristics, a long short-term memory network is used to train the time series data to establish a short-term time change model. Based on the aforementioned long-term trend characteristics, a Transformer model incorporating a global attention mechanism is used to train the time series data to establish a long-term time variation model. The short-term time variation model and the long-term time variation model are integrated to construct the complete mud time variation model.

3. The automated mud treatment method according to claim 2, characterized in that, The process of analyzing the standard parameter data based on a preset mud time variation model to predict the short-term dynamic changes and long-term trends of the mud parameter data, and obtaining data prediction results, specifically includes: Based on the short-term time variation model, short-term time series analysis is performed on the standard parameter data to predict the short-term variation trend of the mud parameter data; Based on the long-term time variation model, long-term trend analysis is performed on the standard parameter data to predict the long-term variation trend of the mud parameter data; The data prediction results are obtained based on the short-term and long-term trends of mud parameter data.

4. The automated mud treatment method according to claim 1, characterized in that, The step of performing a matching analysis on the current work scenario based on the data prediction results, and dynamically switching the corresponding operating mode based on the matching analysis results, specifically includes: If the matching analysis result indicates that the current operation scenario is within the safe threshold range, switch to normal operation mode and adjust the mud treatment parameters using the standard control frequency; If the matching analysis result indicates that the current working scenario is within the safety warning threshold range, switch to high-pressure mode, increase the data prediction frequency, and prioritize the adjustment of key mud parameters. If the matching analysis result indicates that the current operation scenario exceeds the safety warning threshold range, switch to emergency mode, adjust the mud treatment parameters according to the preset safety adjustment strategy, and trigger a safety alarm.

5. The automated mud treatment method according to claim 1, characterized in that, The step of simulating multiple strategies and determining the final processing strategy based on the simulation results specifically includes: A multi-objective optimization algorithm is used to simulate multiple processing strategies. Based on the data prediction results and the introduction of external environmental changes, dynamic balancing is performed, and the weight of each processing strategy is calculated to generate a comprehensive score. The processing strategy with the highest overall score is selected as the final processing strategy.

6. The automated mud treatment method according to claim 1, characterized in that, The step of identifying the abnormal factors causing the abnormal state through a causal relationship model and generating parameter adjustment schemes for the abnormal factors when an abnormal state is detected specifically includes: Based on historical mud parameter data, a causal relationship model between mud parameters is established using a Bayesian network. If the abnormal state is detected, the change chain of the mud parameter data is analyzed based on the causal relationship model to determine the abnormal factors that cause the abnormal state. When the abnormal factor is a equipment malfunction, adjust the corresponding pump pressure parameters to reduce pressure fluctuations. When the abnormal factor is a change in the external environment, increase the amount of mud weighting agent injected to stabilize the mud density.

7. An automated mud treatment system, characterized in that, The automated mud treatment system includes: The data acquisition and processing module is used to acquire mud parameter data in real time, and to clean and standardize the mud parameter data to obtain standard parameter data. The time variation analysis module is used to analyze the standard parameter data based on a preset mud time variation model, predict the short-term dynamic changes and long-term trends of the mud parameter data, and obtain data prediction results. The job scenario matching module is used to obtain the current job scenario, perform matching analysis on the current job scenario based on the data prediction results, and dynamically switch the corresponding operation mode based on the matching analysis results. The strategy generation and optimization module is used to generate multiple processing strategies based on the data prediction results, simulate the multiple strategies, determine the final processing strategy based on the simulation results, and then adjust the corresponding mud treatment parameters according to the final processing strategy. The anomaly detection and adjustment module is used to monitor the mud treatment parameters in real time, and when an abnormal state is detected, it identifies the abnormal factors that cause the abnormal state through a causal relationship model and generates a parameter adjustment scheme for the abnormal factors.

8. The automated mud treatment system according to claim 7, characterized in that, The construction of the preset mud time variation model specifically includes: The historical data extraction module is used to acquire historical operation data, extract time series features of mud parameters from the historical operation data, and analyze the corresponding short-term dynamic features and long-term trend features based on the time series features. The short-term model training module is used to train the time series data using a long short-term memory network based on the short-term dynamic features to establish a short-term time change model. The long-term model training module is used to train the time series data based on the long-term trend features using a Transformer model combined with a global attention mechanism to establish a long-term time change model. The model integration module is used to integrate the short-term time variation model and the long-term time variation model to construct a complete mud time variation model.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the automated mud treatment method as described in any one of claims 1 to 6.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the automated mud treatment method as described in any one of claims 1 to 6.

Images