Methods, systems, and readable media for long-distance multi-stage pump mud delivery in dredged reclamation construction

Abstract

This application discloses a method, system, and readable medium for long-distance multi-stage pump mud transportation in reclamation construction, belonging to the field of mud transportation control technology. The method first divides the transportation pipeline into segments of unit length and statistically stores fault damage data for each segment, using a clustering algorithm to obtain the fault distribution pattern. Then, it completes the division of pipeline control segments based on constraints, collects the pumping parameters of the mud pumps in each control segment, and the mud characteristics data at both ends to construct feature data groups. A dedicated segmented control model for each control segment is established using a deep learning algorithm to clarify the correlation between pumping parameters, mud characteristics, and pipeline faults. Using fault damage type as the optimization objective and mud velocity difference as a constraint, combined with real-time mud data and the segmented control model, the operating parameters of the mud pumps in each control segment are precisely adjusted. This application abandons the traditional single-model overall control mode, achieving refined segmented management of the pipeline, effectively adapting to different transportation conditions, stabilizing mud transportation operations, and significantly reducing the probability of pipeline fault damage.

Description

Technical Field

[0001] This application belongs to the field of mud transportation technology, and relates to a method, system and readable medium for mud transportation by a multi-stage pump with long-distance pumping in dredging construction. Background Technology

[0002] Currently, in the field of long-distance (>10km) mud transportation, the sharp reduction in pipeline life, insufficient system stability, and uncontrolled energy consumption are technical problems that urgently need to be solved.

[0003] Currently, in the field of long-distance mud transportation, the pressure limit of a single-stage pump cannot meet the requirements of long-distance transportation. Therefore, it is necessary to work in concert with multiple pump sets. In short, in addition to the mud pump at the end of the pipeline, a relay pump is set at every set distance on the transportation pipeline. This ensures that the pump is pressurized at every set distance in the transportation pipeline, so that the mud can be transported stably and quickly.

[0004] To ensure stable transmission of mud within the pipeline, current mud transport systems are equipped with intelligent control systems to control the coordinated operation of each mud pump. The working process is roughly as follows: various sensors installed on the pipeline monitor parameters such as the flow rate and pressure of the mud in the pipeline in real time, and then output control parameters according to the set control model to adjust the output power of the mud pumps at various points on the pipeline.

[0005] However, practical experience has revealed several drawbacks in current control systems and methods. For instance, during reclamation operations, different types of mud (concentration, particle size, etc.) flow at varying speeds and bear different pressures in different sections of the pipeline under the same pumping pressure. When mud filled with large particles passes through bends, the resistance encountered during mud transport increases significantly. To address this, current technology typically increases the output power of nearby relay pumps. While this method can increase the driving force on the mud at these points, it also introduces two problems: increased system energy consumption and accelerated pipeline wear (i.e., the aforementioned issues of uncontrolled energy consumption and drastically reduced pipeline lifespan).

[0006] In summary, how to reduce pipeline wear and failures while ensuring stable mud transmission in pipelines is a pressing technical problem that needs to be solved. Summary of the Invention

[0007] To address the problem of poor coordination among mud pumps and severe pipeline wear and failures during long-distance mud transport, this application provides a method for long-distance multi-stage pump mud transport in reclamation construction. This method uses a clustering algorithm to analyze the types and probabilities of failures in different sections of the transport pipeline, thereby segmenting the pipeline. Then, by combining the relationship between mud characteristics and failure types, a targeted control model is configured for each section of the pipeline, thus reducing pipeline wear and failure frequency while ensuring mud transport stability. To implement this mud transport method, this application also proposes a control system for long-distance multi-stage pump mud transport in reclamation construction. Finally, to facilitate the widespread application of this method, this application also proposes a computer-readable storage medium loaded with a computer-readable program module for implementing the above-mentioned mud transport method. The specific solutions are as follows: A method for conveying slurry using a multi-stage pump with long-distance pumping during hydraulic reclamation construction includes: The pipeline is divided into multiple unit length segments along its length and numbered. Fault damage data within each unit length segment is collected and stored in association with the unit length segment number. Clustering algorithms are used to analyze and obtain distribution data of various types of faults or damages, or combinations thereof, on the pipeline. Based on the above distribution data and the set constraints, the transport pipeline is divided into multiple control sections and numbered. Based on historical monitoring data, the pumping parameters of the mud pumps contained in each control section and the mud characteristic data at both ends of the control section are obtained. The above data are associated and stored as a feature data group and stored in correspondence with the control section number. Based on the feature data group corresponding to each control segment, a segmented control model is generated using deep learning algorithm to reflect the correlation between each pumping parameter, mud characteristics and control segment fault damage data. The segmented control model and the control segment number are stored together. Obtain the clustering results of the fault damage types corresponding to each control segment, take the above fault damage types as the control optimization target and the range of mud flow velocity difference between the two ends of the control segment as the control constraint, obtain the mud characteristic data in the current delivery pipeline, and adjust the pumping parameters of the mud pumps contained in the current control segment according to the segmented control model. The fault damage data includes fault damage type, fault damage coordinates, and fault damage degree. The mud properties include mud flow rate, pressure, concentration, particle size, viscosity, and critical flow rate; The pumping parameters include the pumping power and speed of the mud pump, as well as the variation of both over time. The above-mentioned constraints include: each control segment formed after division must include at least one mud pump, and the density concentration of the set fault damage type or combination thereof in each control segment exceeds the set value. The control constraint is that the flow rate at the mud inlet of the control section is not less than the flow rate at the mud outlet.

[0008] The above technical solution controls the slurry pumps on the pipeline by pre-segmenting the pipeline according to fault damage data. Each segment of the pipeline corresponds to a different segmented control model based on pumping parameters, slurry characteristics, and other data. Unlike existing technologies that use a single control model to control the entire long-distance pipeline, this control method is more flexible and adaptable to different transportation scenarios. It places greater emphasis on parameter adjustment in each control segment during the control process, thereby reducing the probability of pipeline fault damage while maintaining the stability of slurry transportation operations. Furthermore, due to the refined operation of segmented control, the stability of slurry transportation can be maintained by fine-tuning the slurry transportation status in each control segment, further reducing the probability of fault damage.

[0009] Furthermore, clustering algorithms are used to analyze and obtain distribution data of various faults or damages, or combinations thereof, on the pipeline, including: Establish and associate the correspondence between various mud characteristic data and the probability of occurrence of various types of faults and damages; Obtain the mud characteristic data in the current delivery pipeline; Based on the analysis of the above mud characteristics data, the expected probability of various faults and damages to the pipeline caused by the mud during its transmission along the pipeline is obtained. Fault damage types or combinations thereof with expected probabilities greater than a set value are selected as reference characteristics, and clustering is performed based on the reference characteristics to obtain fault damage distribution data on the pipeline.

[0010] The above technical solution can identify key reference data for clustering based on mud characteristic data under current working conditions, thereby making the clustering results more targeted and better able to meet the current mud transportation needs.

[0011] Furthermore, obtain the mud characteristic data in the current delivery pipeline, including: Ultrasonic detectors and pressure sensors are used to detect and obtain data on the flow rate, concentration, particle size, and pressure of the mud in the delivery pipeline; Based on the above data on flow rate, concentration, particle size, and pressure, the viscosity and critical flow rate of the mud were analyzed and calculated.

[0012] The above technical solutions can accurately obtain various mud characteristic data.

[0013] Furthermore, a water injection device and a three-way pipe are connected to the control section, and a first solenoid valve and a second solenoid valve are respectively installed between the water injection device and the three-way pipe and the delivery pipe. The mud transport method further includes: Obtain the pumping parameters and mud characteristic data of the current control segment, and calculate the probability of fault damage occurring in the current control segment based on the segmented control model; If the probability of failure or damage exceeds a set threshold and the probability of failure or damage cannot be reduced by adjusting the pumping parameters, then the first solenoid valve and / or the second solenoid valve are turned on to inject a set amount of water into the control section and / or discharge a set amount of mud.

[0014] Through the above technical solution, when the mud characteristics data such as mud concentration or pressure in the conveying pipeline are likely to cause failure or damage, the system changes the mud characteristics data by injecting water or releasing mud, thereby maintaining stable mud conveying while reducing the probability of failure or damage.

[0015] Furthermore, the mud conveying method also includes: The mud characteristic data in each control section is detected and acquired. The mud characteristic data is arranged according to the arrangement order of the sampling points on the delivery pipeline. The flow rate, pressure, concentration and viscosity data in the mud characteristic data are extracted and formed into multiple mud characteristic data queues. Abnormal data in the above mud characteristic data queues were obtained and data sampling sites were marked based on the finite difference method analysis. Identify the sampling point where the abnormal data is located, and adjust the pumping parameters of the control segment where the sampling point is located based on the type of abnormal data and the segmented control model.

[0016] The above technical solutions can identify abnormal sites on the pipeline from a macroscopic perspective, and at the same time, regulate them within the scope of a single control section to ensure the stable and efficient operation of long-distance pipelines.

[0017] Furthermore, the types of fault damage include: pipe blockage, pipe wear, pipe leakage, and connection failure; The mud transport method further includes: Real-time monitoring of mud characteristics data in each control section of the conveying pipeline, and calculation of the degree of blockage and average wear of the conveying pipeline; If the congestion level is higher than the rated value, the control segment and its corresponding segmented control model, which are clustered according to the congestion level, will be activated. If the wear level is higher than the rated value, switch to the control segment and its corresponding segmented control model that are clustered according to the wear level; The degree of blockage is calculated by analyzing the average pressure, average flow velocity, and output power of each mud pump in the delivery pipeline.

[0018] The above technical solutions can achieve dynamic adaptation between control segments and operating conditions.

[0019] Furthermore, methods for generating segmented control models also include: A deep learning model is constructed by combining a temporal convolutional neural network with an attention mechanism to perform temporal feature extraction and key parameter weighting on the feature data set. At preset intervals, newly added pumping parameters, mud characteristics, and fault damage data are collected as training samples to perform incremental iterative optimization on the segmented control model.

[0020] The above technical solutions can continuously improve the accuracy of parameter control in segmented control models.

[0021] A multi-stage pump mud conveying system for long-distance reclamation construction includes: The pipeline segmentation and numbering module is used to divide the pipeline along its length into multiple unit length segments and number them sequentially. It also collects fault damage data, including fault damage type, fault damage coordinates, and fault damage degree, within each unit length segment and establishes an association between the fault damage data and the unit length segment number. The fault distribution analysis module has a built-in mapping relationship between mud characteristics and the probability of occurrence of various fault damages. It is used to acquire mud characteristic data in the conveying pipeline in real time, analyze the expected probability of various fault damages to the pipeline during mud transmission, filter fault damage types or combinations thereof with expected probabilities greater than a set value as reference characteristics, and combine clustering algorithms to analyze the distribution data of various fault damages or combinations thereof on the conveying pipeline. The control segment division module is used to divide the delivery pipeline into multiple control segments and number them based on fault damage distribution data and set constraints. The feature data storage module is configured to collect pumping parameters of mud pumps in each control section and mud characteristic data at both ends of the control section based on historical monitoring data, associate the pumping parameters and mud characteristic data to form a feature data group, and store it in association with the corresponding control section number. The segmented model construction module is configured to train and analyze the feature data groups corresponding to each control segment based on deep learning algorithms, construct and store a segmented control model to reflect the correlation between pumping parameters, mud characteristics and control segment fault damage data, and associate and bind the segmented control model with the corresponding control segment number. The pump group parameter control module is used to obtain the clustering results of the fault damage types of each control section, take the fault damage type as the control optimization target, and take the flow velocity at the mud inlet end of the control section not less than the flow velocity at the output end as the control constraint condition, call the corresponding segment control model to dynamically adjust the pumping parameters of the mud pump in the current control section, so as to realize the segmented fine control of long-distance mud transportation. The mud characteristics include mud flow rate, pressure, concentration, particle size, viscosity, and critical flow rate. The set constraints include that each control segment after division contains at least one mud pump, and the density concentration of the set fault damage type or combination thereof in each control segment exceeds a set value. The pumping parameters include the pumping power and speed of the mud pump, as well as the variation of both over time.

[0022] Furthermore, the types of fault damage include pipe blockage, pipe wear, pipe leakage, and connection failure; The mud conveying system also includes an adaptive switching module for operating conditions; The adaptive switching module is used to monitor the mud characteristics data of each control section in real time, calculate the degree of blockage in the pipeline based on the average pressure, average flow velocity and output power of each mud pump in the pipeline, and at the same time measure the average wear of the pipeline in real time. The adaptive working condition switching module is equipped with a model switching control unit: When the degree of blockage in the pipeline exceeds the rated value, the control segment corresponding to the blockage degree cluster and the matching segmented control model are activated. When the wear level of the pipeline exceeds the rated value, it automatically switches to the control segment corresponding to the wear level cluster and the matching segmented control model.

[0023] The above technical solutions can achieve dynamic adaptation between the control model and real-time conveying conditions.

[0024] A computer-readable storage medium having a program module loaded thereon, which, when executed by a processor, is used to implement the long-distance multi-stage pump mud transport method for reclamation construction as described above.

[0025] The above technical solutions will help promote and apply the mud transportation method proposed in this application.

[0026] This application includes at least one of the following beneficial effects: (1) By adopting the method of segmenting the pipeline according to the fault damage data and matching each control segment with a dedicated control model, the limitation of the traditional single model control of the entire pipeline is broken through, the adaptability is stronger, and it can flexibly cope with various mud transportation scenarios. (2) Based on the characteristics of the conveyed mud, the segmented control module is selectively used to realize the refined segmented management and control of the pipeline and mud pump, and the pumping parameters of each control segment are adjusted in a targeted manner to effectively reduce the probability of pipeline failure and damage. (3) By fine-tuning the mud conveying status of each segment, the overall operation of long-distance mud conveying operation is stabilized, and the continuity of construction is improved. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of the mud transport method of this application; Figure 2 A schematic diagram illustrating a method for obtaining fault damage distribution data on a pipeline. Figure 3 This is a schematic diagram showing the connection of the functional modules of the mud conveying system.

[0028] Attached reference numerals: 100, Pipeline segment numbering module; 200, Fault distribution analysis module; 300, Control segment division module; 400, Feature data storage module; 500, Segment model construction module; 600, Pump group parameter control module; 700, Operating condition adaptive switching module. Detailed Implementation

[0029] The embodiments of this application are described in detail below, and examples of the embodiments are shown in the accompanying drawings.

[0030] In the description of this specification, the references to "certain embodiments," "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples" refer to specific features, structures, materials, or characteristics described in connection with the described embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0031] A method for conveying slurry using a multi-stage pump with long spacing during hydraulic reclamation construction, such as... Figure 1 As shown, the main steps include the following: S100: Divide the pipeline along its length into multiple unit length segments and number them, collect fault damage data within each unit length segment and store it in association with the unit length segment number; S200 uses clustering algorithms to analyze and obtain distribution data of various faults or damages or their combinations on the pipeline; S300, Based on the above distribution data and the set constraints, the conveying pipeline is divided into multiple control sections and numbered; S400: Based on historical monitoring data, obtain the pumping parameters of the mud pumps contained in each control section and the mud characteristic data at both ends of the control section, and store the above data as a feature data group and store it in relation to the control section number. S500, based on the characteristic data group corresponding to each control section, uses deep learning algorithm to analyze and generate a segmented control model to reflect the correlation between each pumping parameter, mud characteristics and control section fault damage data, and stores the segmented control model and control section number together. S600: Obtain the clustering results of the fault damage types corresponding to each control section, take the above fault damage types as the control optimization target and the range of mud flow velocity difference between the two ends of the control section as the control constraint, obtain the mud characteristic data in the current delivery pipeline, and adjust the pumping parameters of the mud pumps contained in the current control section according to the segmented control model.

[0032] In step S100 above, the fault damage data includes fault damage type, fault damage coordinates, and fault damage severity. Fault damage types include, but are not limited to, pipe blockage, pipe wear, pipe leakage, and connection failure.

[0033] In this embodiment of the application, the unit length is defined as 1m. A coordinate system is set up along the length of the long-distance conveying pipeline with the inlet of the pipeline as the origin. The above-mentioned fault damage coordinates are the locations of the faults or damages on the pipeline. For example, coordinate point 53.5 means that the location of the fault or damage is 53.5m away from the inlet of the conveying pipeline.

[0034] In step S200, the distribution data of various types of fault damage or their combinations on the conveying pipeline are obtained through clustering algorithm analysis, such as... Figure 2 As shown, it includes: S210, Establish and associate the correspondence between various mud characteristic data and the probability of occurrence of various faults and damages; S220, Obtain mud characteristic data in the current conveying pipeline; S230, Based on the above mud characteristic data analysis, the expected probability of the mud causing various faults and damages to the pipeline during its transmission along the pipeline is obtained. S240: Select fault damage types or combinations thereof with expected probabilities greater than a set value as reference characteristics, and perform clustering based on the reference characteristics to obtain fault damage distribution data on the pipeline.

[0035] The mud characteristics include mud flow rate, pressure, concentration, particle size, viscosity, and critical flow rate.

[0036] Step S210 above can obtain the corresponding relationship by analyzing historical monitoring data of the transport pipeline (including mud characteristic data and the frequency of fault damage within a set time period). For example, when the mud flow velocity in the transport pipeline is high and the particle size is large, the probability of pipeline wear is greater. The above technical solution can identify key reference data for clustering based on mud characteristic data under current operating conditions, thereby making the clustering results more targeted and better able to meet the current mud transport needs.

[0037] In step S210 above, the clustering process includes clustering according to fault damage type. In certain embodiments, it also includes multidimensional joint clustering, such as clustering multiple damage types and damage degrees together as reference characteristics. In the embodiments of this application, the clustering operation preferably adopts the DBSCAN algorithm. This algorithm is based on fault damage density clustering, automatically identifies dense regions, marks sparse points as noise, and has a significant clustering effect.

[0038] It should be noted that in step S210, using different types of fault damage or their combination data as reference characteristics for clustering will result in multiple sets of different fault damage distribution data, that is, multiple different control segment division methods.

[0039] In step S300, the set constraints include at least: 1. Each control segment formed after division includes at least one mud pump; 2. The density concentration of the set fault damage type or its combination in each control segment exceeds the set value.

[0040] The significance of the first constraint is that there are sufficient active adjustment methods to adjust the mud characteristics of the divided control sections; the significance of the second constraint is that the fault damage types in the divided control sections are relatively concentrated, and the mud characteristic data can be adjusted in a targeted manner through the segmented control model.

[0041] In step S500, the pumping parameters include the pumping power and speed of the mud pump, as well as the variation of both over time.

[0042] To continuously improve the parameter control accuracy of the segmented control model, the method for generating the segmented control model in this application further includes: constructing a deep learning model using a temporal convolutional neural network combined with an attention mechanism, and performing temporal feature extraction and key parameter weighting on the feature data set. Newly added pumping parameters, mud characteristics, and fault damage data are collected at preset intervals as training samples to incrementally iteratively optimize the segmented control model.

[0043] In steps S220 and S600, the mud characteristic data in the current conveying pipeline is obtained, specifically including: The A100 uses an ultrasonic detector and a pressure sensor to detect and obtain data on the flow rate, concentration, particle size, and pressure of the mud in the delivery pipeline. Based on the aforementioned flow rate, concentration, particle size, and pressure data, the viscosity and critical flow rate of the mud are analyzed and calculated using A200. For example, the friction differential pressure ΔP is obtained from a pressure sensor, the flow rate v is obtained from a flow meter, and the mud density ρm is obtained from a densitometer. Then, the Reynolds number is calculated to determine the mud flow regime, and the effective viscosity μea is derived through iterative solution.

[0044] The control constraint in step S600 above is that the flow rate at the mud inlet of the control section is not less than the flow rate at the mud outlet. In practical applications, if the inner diameters of the pipes at both ends of the control section are the same, the control constraint can be simplified to the mud flow velocity at the mud inlet of the control section being not less than the mud flow velocity at the mud outlet. Based on the above control constraint, the probability of mud blockage in the conveying pipeline can be reduced.

[0045] To facilitate the adjustment of the mud characteristics within the conveying pipeline, a water injection device and a tee pipe are connected to the control section. A first solenoid valve and a second solenoid valve are respectively installed between the water injection device and the tee pipe and the conveying pipeline. Based on the above configuration, the mud conveying method further includes: S710: Obtain pumping parameters and mud characteristic data for the current control segment, and calculate the probability of fault damage occurring in the current control segment based on the segmented control model. S711, if the probability of failure or damage exceeds a set threshold and the probability of failure or damage cannot be reduced by adjusting the pumping parameters, then the first solenoid valve and / or the second solenoid valve are turned on to inject a set amount of water into the control section and / or discharge a set amount of mud to adjust the mud characteristics in the pipeline.

[0046] Based on the above scheme, when the mud characteristics data such as mud concentration or pressure in the conveying pipeline are likely to cause failure or damage, the system changes the mud characteristics data by injecting water or releasing mud, thereby maintaining stable mud conveying while reducing the probability of failure or damage.

[0047] In addition to precise adjustment of individual control sections, the mud conveying method in this application embodiment further includes: S810, detect and acquire mud characteristic data in each control section, arrange the mud characteristic data according to the arrangement order of the sampling points on the conveying pipeline, extract the flow rate, pressure, concentration and viscosity data in the mud characteristic data, and form multiple mud characteristic data queues respectively.

[0048] S820, based on the finite difference method, abnormal data in the above-mentioned mud characteristic data queues are obtained and the data sampling sites are marked. In this application, the first-order finite difference method can be directly used to quickly identify abnormal data.

[0049] S830: Identify the sampling point where the abnormal data is located, and adjust the pumping parameters of the control segment where the sampling point is located based on the type of abnormal data and the segmented control model.

[0050] In practical applications, when mud is transported along a pipeline, its flow velocity and pressure exhibit linear changes. For example, after the mud is pumped out from the mud pump outlet, its flow velocity gradually decreases until it is pumped again by the next mud pump. By acquiring the changing trends of these mud characteristic data, abnormal data sampling points on the pipeline can be quickly identified. This allows for macroscopic identification of abnormal points on the pipeline, while simultaneously enabling control within a single control section to ensure the stable and efficient operation of long-distance transport pipelines.

[0051] Finally, in order to achieve dynamic adaptation between control segments and operating conditions, the mud conveying method in this embodiment further includes: B100 monitors the mud characteristics of each control section in the delivery pipeline in real time, and calculates the degree of blockage and the average wear of the delivery pipeline. B110, if the congestion level is higher than the rated value, activate the control segment and its corresponding segmented control model that are clustered according to the congestion level; B120, if the wear level is higher than the rated value, switches to the control segment and its corresponding segmented control model that are clustered by wear level.

[0052] The degree of blockage is calculated by analyzing the average pressure, average flow velocity, and output power of each mud pump within the pipeline. In practical applications, in addition to considering the degree of pipeline blockage and wear, the degree of pipeline leakage or vibration can also be considered, thereby generating corresponding control segment results and segmented control models. The specific implementation process will not be elaborated again.

[0053] To implement the above-mentioned method for conveying slurry using a multi-stage pump with long spacing during reclamation construction, this application also discloses a system for conveying slurry using a multi-stage pump with long spacing during reclamation construction, combined with... Figure 3 As shown, it mainly includes a pipeline segment numbering module 100, a fault distribution analysis module 200, a control segment division module 300, a feature data storage module 400, a segment model construction module 500, and a pump group parameter control module 600.

[0054] The pipeline segmentation and numbering module 100 is used to divide the pipeline along its length into multiple unit-length segments and number them sequentially. It then collects fault damage data within each unit-length segment, including the fault damage type, coordinates, and severity, and establishes a storage association between the fault damage data and the unit-length segment numbers. Fault damage types include pipeline blockage, pipeline wear, pipeline leakage, and connection failure. In practical applications, this pipeline segmentation and numbering module is a CAD-related program module.

[0055] The fault distribution analysis module 200 incorporates a mapping relationship between mud characteristics and the probability of various fault damages. It acquires real-time mud characteristic data within the pipeline, analyzes the expected probability of various fault damages caused to the pipeline during mud transport, filters fault damage types or combinations thereof with expected probabilities greater than a set value as reference characteristics, and combines this with clustering algorithms to obtain the distribution data of various fault damages or combinations on the pipeline. In practical applications, the fault distribution analysis module 200 incorporates the DBSCAN algorithm module and performs data clustering analysis and outputs results based on the set reference characteristics. The mud characteristics include mud flow rate, pressure, concentration, particle size, viscosity, and critical flow velocity.

[0056] The control segment division module 300 is used to divide the pipeline into multiple control segments and number them based on fault damage distribution data and set constraints. The set constraints include that each control segment after division contains at least one mud pump, and the density concentration of set fault damage types or combinations thereof within each control segment exceeds a set value.

[0057] The feature data storage module 400 is configured to collect pumping parameters of the mud pumps in each control section and mud characteristic data at both ends of the control section based on historical monitoring data. It associates the pumping parameters and mud characteristic data to form a feature data group and stores it in association with the corresponding control section number. The segmented model construction module 500 is configured to train and analyze the feature data groups corresponding to each control section based on a deep learning algorithm, construct and store a segmented control model reflecting the correlation between pumping parameters, mud characteristics, and control section fault damage data, and bind the segmented control model to the corresponding control section number. The pumping parameters include the pumping power and rotational speed of the mud pump, and the variation of both over time.

[0058] The pump group parameter control module 600 is used to obtain the clustering results of fault damage types in each control section. Taking the fault damage type as the control optimization target and the flow velocity at the mud inlet end of the control section not being less than the flow velocity at the output end as the control constraint, the corresponding segment control model is called to dynamically adjust the pumping parameters of the mud pump in the current control section, so as to realize the segmented fine control of long-distance mud transportation.

[0059] The optimized mud conveying system also includes an adaptive switching module 700. In this embodiment, the adaptive switching module 700 is used to monitor the mud characteristic data of each control section in real time, calculate the degree of blockage in the conveying pipeline based on the average pressure, average flow velocity and output power of each mud pump in the conveying pipeline, and simultaneously measure the average wear degree of the conveying pipeline in real time.

[0060] The 700 adaptive switching module is equipped with a model switching control unit: when the degree of pipe blockage is higher than the rated value, the control segment corresponding to the blockage degree cluster and the matching segmented control model are activated; when the degree of pipe wear is higher than the rated value, it automatically switches to the control segment corresponding to the wear degree cluster and the matching segmented control model, thereby realizing the dynamic adaptation of the control model to the real-time conveying conditions.

[0061] To facilitate the promotion and use of the mud conveying method described in this application, this application also discloses a computer-readable storage medium loaded with a program module. When the program module is executed by a processor, it is used to implement the long-distance multi-stage pump mud conveying method for reclamation construction as described above. The computer-readable storage medium includes, but is not limited to, disk storage, CD-ROM, optical storage, etc.

[0062] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A method for conveying mud slurry using a multi-stage pump with long spacing during hydraulic reclamation construction, characterized in that, include: The pipeline is divided into multiple unit length segments along its length and numbered. Fault damage data within each unit length segment is collected and stored in association with the unit length segment number. Clustering algorithms are used to analyze and obtain distribution data of various types of faults or damages, or combinations thereof, on the pipeline. Based on the above distribution data and the set constraints, the transport pipeline is divided into multiple control sections and numbered. Based on historical monitoring data, the pumping parameters of the mud pumps contained in each control section and the mud characteristic data at both ends of the control section are obtained. The above data are associated and stored as a feature data group and stored in correspondence with the control section number. Based on the feature data group corresponding to each control segment, a segmented control model is generated using deep learning algorithm to reflect the correlation between each pumping parameter, mud characteristics and control segment fault damage data. The segmented control model and the control segment number are stored together. Obtain the clustering results of the fault damage types corresponding to each control segment, take the above fault damage types as the control optimization target and the range of mud flow velocity difference between the two ends of the control segment as the control constraint, obtain the mud characteristic data in the current delivery pipeline, and adjust the pumping parameters of the mud pumps contained in the current control segment according to the segmented control model. The fault damage data includes fault damage type, fault damage coordinates, and fault damage degree. The mud properties include mud flow rate, pressure, concentration, particle size, viscosity, and critical flow rate; The pumping parameters include the pumping power and speed of the mud pump, as well as the variation of both over time. The above-mentioned constraints include: each control segment formed after division must include at least one mud pump, and the density concentration of the set fault damage type or combination thereof in each control segment exceeds the set value. The control constraint is that the flow rate at the mud inlet of the control section is not less than the flow rate at the mud outlet.

2. The method for conveying slurry using a multi-stage pump with long spacing during hydraulic reclamation construction according to claim 1, characterized in that, Distribution data of various faults or their combinations on the pipeline were obtained through clustering algorithm analysis, including: Establish and associate the correspondence between various mud characteristic data and the probability of occurrence of various types of faults and damages; Obtain the mud characteristic data in the current delivery pipeline; Based on the analysis of the above mud characteristics data, the expected probability of various faults and damages to the pipeline caused by the mud during its transmission along the pipeline is obtained. Fault damage types or combinations thereof with expected probabilities greater than a set value are selected as reference characteristics, and clustering is performed based on the reference characteristics to obtain fault damage distribution data on the pipeline.

3. The method for conveying slurry using a multi-stage pump with long spacing during hydraulic reclamation construction according to claim 2, characterized in that, Obtain the mud characteristic data in the current delivery pipeline, including: Ultrasonic detectors and pressure sensors are used to detect and obtain data on the flow rate, concentration, particle size, and pressure of the mud in the delivery pipeline; Based on the above data on flow rate, concentration, particle size, and pressure, the viscosity and critical flow rate of the mud were analyzed and calculated.

4. The method for conveying slurry using a multi-stage pump with long spacing during hydraulic reclamation construction according to claim 1, characterized in that, The control section is connected to a water injection device and a three-way pipe. A first solenoid valve and a second solenoid valve are respectively installed between the water injection device and the three-way pipe and the delivery pipeline. The mud transport method further includes: Obtain the pumping parameters and mud characteristic data of the current control segment, and calculate the probability of fault damage occurring in the current control segment based on the segmented control model; If the probability of failure or damage exceeds a set threshold and the probability of failure or damage cannot be reduced by adjusting the pumping parameters, then the first solenoid valve and / or the second solenoid valve are turned on to inject a set amount of water into the control section and / or discharge a set amount of mud.

5. The method for conveying slurry using a multi-stage pump with long spacing during hydraulic reclamation construction according to claim 1, characterized in that, The mud transport method further includes: The mud characteristic data in each control section is detected and acquired. The mud characteristic data is arranged according to the arrangement order of the sampling points on the delivery pipeline. The flow rate, pressure, concentration and viscosity data in the mud characteristic data are extracted and formed into multiple mud characteristic data queues. Abnormal data in the above mud characteristic data queues were obtained and data sampling sites were marked based on the finite difference method analysis. Identify the sampling point where the abnormal data is located, and adjust the pumping parameters of the control segment where the sampling point is located based on the type of abnormal data and the segmented control model.

6. The method for conveying slurry using a multi-stage pump with long spacing during hydraulic reclamation construction according to claim 1, characterized in that, The types of fault damage include: pipe blockage, pipe wear, pipe leakage, and connection failure; The mud transport method further includes: Real-time monitoring of mud characteristics data in each control section of the conveying pipeline, and calculation of the degree of blockage and average wear of the conveying pipeline; If the congestion level is higher than the rated value, the control segment and its corresponding segmented control model, which are clustered according to the congestion level, will be activated. If the wear level is higher than the rated value, switch to the control segment and its corresponding segmented control model that are clustered according to the wear level; The degree of blockage is calculated by analyzing the average pressure, average flow velocity, and output power of each mud pump in the delivery pipeline.

7. The method for conveying slurry using a multi-stage pump with long spacing during hydraulic reclamation construction according to claim 1, characterized in that, Methods for generating piecewise control models also include: A deep learning model is constructed by combining a temporal convolutional neural network with an attention mechanism to perform temporal feature extraction and key parameter weighting on the feature data set. At preset intervals, newly added pumping parameters, mud characteristics, and fault damage data are collected as training samples to perform incremental iterative optimization on the segmented control model.

8. A multi-stage pump mud conveying system for long-distance reclamation construction, characterized in that, include: The pipeline segment numbering module (100) is used to divide the pipeline into multiple unit length segments along its length and number them sequentially, to collect fault damage data containing fault damage type, fault damage coordinates, and fault damage degree within each unit length segment, and to establish an association storage relationship between the fault damage data and the unit length segment number. The fault distribution analysis module (200) has a built-in mapping relationship between mud characteristics and the probability of occurrence of various fault damages. It is used to obtain mud characteristic data in the conveying pipeline in real time, analyze the expected probability of various fault damages to the pipeline during mud transmission, filter fault damage types or combinations thereof with expected probabilities greater than the set value as reference characteristics, and combine clustering algorithms to analyze the distribution data of various fault damages or combinations thereof on the conveying pipeline. The control section division module (300) is used to divide the delivery pipeline into multiple control sections and number them based on the fault damage distribution data and set constraints. The feature data storage module (400) is configured to collect pumping parameters of mud pumps in each control section and mud characteristic data at both ends of the control section based on historical monitoring data, associate the pumping parameters and mud characteristic data to form a feature data group, and store it in association with the corresponding control section number. The segmented model construction module (500) is configured to train and analyze the feature data group corresponding to each control segment based on the deep learning algorithm, construct and store the segmented control model to reflect the relationship between pumping parameters, mud characteristics and control segment fault damage data, and associate and bind the segmented control model with the corresponding control segment number. The pump group parameter control module (600) is used to obtain the clustering results of the fault damage type of each control section, take the fault damage type as the control optimization target and take the flow velocity at the mud inlet end of the control section not less than the flow velocity at the output end as the control constraint condition, call the corresponding segment control model to dynamically adjust the pumping parameters of the mud pump in the current control section, and realize the segmented fine control of long-distance mud transportation. The mud characteristics include mud flow rate, pressure, concentration, particle size, viscosity, and critical flow rate. The set constraints include that each control segment after division contains at least one mud pump, and the density concentration of the set fault damage type or combination thereof in each control segment exceeds a set value. The pumping parameters include the pumping power and speed of the mud pump, as well as how these two parameters change over time.

9. The long-distance multi-stage pump mud conveying system for reclamation construction according to claim 8, characterized in that, The types of fault damage include pipe blockage, pipe wear, pipe leakage, and connection failure. The mud conveying system also includes an adaptive switching module (700); The adaptive switching module (700) is used to monitor the mud characteristic data of each control section in real time, calculate the degree of blockage in the pipeline based on the average pressure, average flow velocity and output power of each mud pump in the pipeline, and at the same time measure the average wear of the pipeline in real time. The adaptive switching module (700) is equipped with a model switching control unit: When the degree of blockage in the pipeline exceeds the rated value, the control segment corresponding to the blockage degree cluster and the matching segmented control model are activated. When the wear level of the pipeline exceeds the rated value, it automatically switches to the control segment corresponding to the wear level cluster and the matching segmented control model.

10. A computer-readable medium, characterized in that, It is loaded with a program module, which, when executed by the processor, is used to implement the long-distance multi-stage pump mud transportation method for reclamation construction as described in any one of claims 1-7.

Images