A remote operation and maintenance control method and system for an emulsion pump

Abstract

The present invention discloses a remote operation and maintenance control method and system for an emulsion pump, which relates to the field of operation and maintenance control technology. The method comprises: collecting real-time status data of the emulsion pump through an edge device deployed in a target scenario; calling a baseline model and constructing an anomaly assessment model based on the identity information of the emulsion pump; training the anomaly assessment model with multi-source operation and maintenance samples as supervision; inputting real-time status data into the model, evaluating the anomaly type and probability, and generating an anomaly assessment result; making an operation and maintenance decision based on the evaluation result and preset operation and maintenance discrimination constraints, and generating corresponding control instructions, which are sent to the edge device for execution. This achieves the technical effect of providing scientific operation and maintenance decisions and improving the operation and maintenance efficiency and reliability of the emulsion pump.

Description

Technical Field

[0001] The present invention relates to the technical field of operation and maintenance control, and in particular to a remote operation and maintenance control method and system for an emulsion pump. Background Art

[0002] The operation and maintenance management of emulsion pumps mainly relies on manual inspections and regular maintenance. Traditional operation and maintenance methods cannot obtain equipment operating status data in real time, resulting in untimely fault diagnosis; they cannot accurately predict the type and probability of equipment abnormalities; operation and maintenance decisions lack scientific basis, and cannot accurately issue control instructions based on the actual operating status of the equipment, resulting in the equipment's operating capacity not being fully utilized, and also bringing great pressure to operation and maintenance personnel. Summary of the Invention

[0003] The present invention provides a remote operation and maintenance control method and system for an emulsion pump to solve the technical problems in the prior art of simple and extensive operation and maintenance decisions, low operation and maintenance efficiency, and high operation and maintenance pressure, thereby achieving the technical effect of providing scientific operation and maintenance decisions and improving the operation and maintenance efficiency and reliability of the emulsion pump.

[0004] In a first aspect, the present invention provides a remote operation and maintenance control method for an emulsion pump, wherein the method comprises:

[0005] Data is collected by edge devices deployed in the target scene to obtain real-time status data of the target emulsion pump, wherein the edge device is communicatively connected to the sensor network deployed in the target emulsion pump.

[0006] Based on the identification information of the target emulsion pump, the benchmark model is called in combination with the benchmark model library, and an anomaly assessment model is constructed according to the calling results.

[0007] Multi-source operation and maintenance samples are obtained using the identity identification information as a call constraint, and the anomaly assessment model is trained using the multi-source operation and maintenance samples as supervision.

[0008] The real-time status data is input into the anomaly assessment model for anomaly assessment, and the predicted anomaly type and predicted anomaly probability are obtained, and output as an anomaly assessment result, wherein the predicted anomaly type and the predicted anomaly probability correspond one to one.

[0009] An operation and maintenance decision is made based on the abnormality assessment result and the preset operation and maintenance discrimination constraints, and an operation and maintenance control instruction is generated based on the operation and maintenance decision result, and the operation and maintenance control instruction is sent to the edge device for control execution, wherein the operation and maintenance discrimination constraints include emergency stop constraint limits and degradation constraint limits.

[0010] In a second aspect, the present invention further provides an emulsion pump remote operation and maintenance control system, wherein the system comprises:

[0011] The data acquisition module is used to collect data through an edge device deployed in a target scene to obtain real-time status data of a target emulsion pump, wherein the edge device is communicatively connected to a sensor network deployed on the target emulsion pump.

[0012] The benchmark model calling module is used to call the benchmark model based on the identity identification information of the target emulsion pump in combination with the benchmark model library, and to build an anomaly assessment model according to the calling results.

[0013] The anomaly assessment model training module is used to obtain multi-source operation and maintenance samples using the identity identification information as a call constraint, and to train the anomaly assessment model using the multi-source operation and maintenance samples as supervision.

[0014] The anomaly assessment module is used to input the real-time status data into the anomaly assessment model for anomaly assessment, obtain the predicted anomaly type and the predicted anomaly probability, and output them as an anomaly assessment result, wherein the predicted anomaly type and the predicted anomaly probability correspond one to one.

[0015] An operation and maintenance decision execution module is used to make operation and maintenance decisions based on the abnormality assessment results and preset operation and maintenance discrimination constraints, and generate operation and maintenance control instructions based on the operation and maintenance decision results, and send the operation and maintenance control instructions to the edge device for control execution, wherein the operation and maintenance discrimination constraints include emergency stop constraint limits and degradation constraint limits.

[0016] The present invention discloses a remote operation and maintenance control method and system for an emulsion pump, comprising: acquiring real-time status data of a target emulsion pump by collecting data through an edge device deployed in a target scene, wherein the edge device is communicatively connected to a sensor network deployed on the target emulsion pump; calling a benchmark model based on the identity identification information of the target emulsion pump in combination with a benchmark model library, and constructing an anomaly assessment model according to the call result; acquiring multi-source operation and maintenance samples with the identity identification information as a call constraint, and training an anomaly assessment model with the multi-source operation and maintenance samples as supervision; inputting real-time status data into the anomaly assessment model for anomaly assessment, and obtaining predicted anomaly types and The abnormality probability is predicted and output as the abnormality assessment result, wherein the predicted abnormality type corresponds to the predicted abnormality probability one-to-one; an operation and maintenance decision is made according to the abnormality assessment result and the preset operation and maintenance discrimination constraint, and an operation and maintenance control instruction is generated based on the operation and maintenance decision result, and the operation and maintenance control instruction is sent to the edge device for control execution, wherein the operation and maintenance discrimination constraint includes an emergency stop constraint limit and a degradation constraint limit. The remote operation and maintenance control method and system of an emulsion pump disclosed in the present invention solves the technical problems of simple and extensive operation and maintenance decision-making, low operation and maintenance efficiency, and high operation and maintenance pressure, and achieves the technical effect of providing scientific operation and maintenance decision-making and improving the operation and maintenance efficiency and reliability of the emulsion pump. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1This is a flow chart of a remote operation and maintenance control method for an emulsion pump according to the present invention;

[0018] Figure 2 This is a structural schematic diagram of an emulsion pump remote operation and maintenance control system of the present invention.

[0019] Explanation of the accompanying symbols: data acquisition module 11, benchmark model calling module 12, anomaly assessment model training module 13, anomaly assessment module 14, operation and maintenance decision execution module 15. DETAILED DESCRIPTION

[0020] The above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods of the specification to better understand the above technical solution. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments of the present invention. It should be understood that the present invention is not limited to the example embodiments used only to explain the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. In addition, it should be noted that, for the convenience of description, only the parts related to the present invention, rather than all, are shown in the drawings.

[0021] Example 1

[0022] Figure 1 The figure is a flow chart of a remote operation and maintenance control method of an emulsion pump according to the present invention, wherein the method comprises:

[0023] S100: Acquire real-time status data of a target emulsion pump by collecting data through an edge device deployed in a target scene, wherein the edge device is communicatively connected to a sensor network deployed in the target emulsion pump.

[0024] Specifically, edge devices refer to miniaturized, intelligent devices deployed in industrial sites or target scenarios, equipped with data collection, processing, and transmission capabilities. A sensor network is a system composed of multiple sensors used to monitor the operating parameters of emulsion pumps (such as pressure, temperature, and flow) in real time. The communication connection between edge devices and sensor networks means that the two can exchange data via wired or wireless communication protocols (such as Industrial Ethernet, LoRa, and NB-IoT), enabling real-time collection of emulsion pump status data. Specifically, the real-time status data of the target emulsion pump acquired by the sensor is initially processed and transmitted via the edge device in the target scenario.

[0025] Specifically, first, edge devices are deployed in the target scenario and connected to the sensor network on the emulsion pump for communication. Then, the sensor network collects the operating status data of the emulsion pump (such as pressure, temperature, flow, vibration, etc.) in real time based on preset collection parameters, and transmits this data to the edge device. Next, the edge device performs preliminary processing on the collected data, such as data cleaning and formatting, and then uploads the processed data to the remote operation and maintenance management platform for subsequent anomaly assessment.

[0026] Among them, the remote operation and maintenance management platform is a centralized system used to receive, process and analyze real-time data from edge devices, and perform target emulsion pump status monitoring, fault diagnosis and operation and maintenance decisions based on this data. The remote operation and maintenance management platform is deployed on the cloud or local server and can uniformly manage multiple emulsion pumps.

[0027] Through the above process, real-time and accurate target emulsion pump operation data is provided for subsequent abnormality assessment, which reduces manual intervention, improves the degree of automation of operation and maintenance management, reduces the workload and operation and maintenance costs of operation and maintenance personnel, and ensures that operation and maintenance can be analyzed and decided based on the latest status information.

[0028] S200: Based on the identification information of the target emulsion pump, a benchmark model is called in combination with the benchmark model library, and an anomaly assessment model is constructed according to the call result.

[0029] Specifically, the target emulsion pump's identification information refers to the characteristic information used to uniquely identify the emulsion pump, such as the device model, serial number, and manufacturer information. The benchmark model library is a pre-built model collection that contains a variety of prediction models with different model structures and model complexity, such as decision trees, neural networks, and support vector machines. Model complexity reflects the model's computing resource requirements and fitting capabilities. Models with higher complexity can capture more complex patterns, but also require more computing resources and data. Through the design of the benchmark model library, a diverse selection of models is provided for the construction of subsequent anomaly assessment models to meet the anomaly assessment needs in different scenarios.

[0030] Specifically, according to the identity identification information of the target emulsion pump, a model matching the device is selected from the benchmark model library as the basis, wherein the anomaly assessment model is further constructed based on the called benchmark model, and is used to perform anomaly assessment on the operating status of the emulsion pump. For example, for a high-end emulsion pump that often has complex faults, a deep learning model with higher complexity may be selected; while for an ordinary emulsion pump with a relatively simple fault mode, a lightweight decision tree model may be selected. The benchmark model library provides a variety of options for the construction of the anomaly assessment model, ensuring that the model can adapt to the operation and maintenance needs of different emulsion pumps.

[0031] Optionally, the baseline model library can be dynamically adjusted according to the actual needs of the device. For example, when the device is upgraded or the failure mode changes, it can quickly switch to a more appropriate model, enhancing flexibility and scalability. In the above steps, the pre-built baseline model library avoids the complex process of building the model from scratch, reduces the time and resource consumption of model training, and improves overall operational efficiency.

[0032] In some embodiments, based on the identification information of the target emulsion pump, a benchmark model is called in combination with a benchmark model library, and an anomaly assessment model is constructed according to the call result, including:

[0033] Based on the identity identification information interactive emulsion knowledge base, determine the existing anomaly types; obtain the historical operation and maintenance logs of the target scenario, and perform statistical analysis on the historical operation and maintenance logs to obtain the anomaly characteristic indicators corresponding to each anomaly type, wherein the anomaly characteristic indicators include anomaly rate, anomaly proportion, and anomaly missed false positive rate; based on the anomaly rate, define the call model number constraint corresponding to each anomaly type, and based on the anomaly missed false positive rate, define the call model complexity constraint corresponding to each anomaly type; using the call model number constraint and the call model complexity constraint, traverse the benchmark model library to call the benchmark model of each anomaly type, and construct the anomaly assessment model based on the call results.

[0034] Specifically, the emulsion knowledge base is a database containing knowledge related to emulsion pumps, including equipment parameters, common fault types, fault characteristics and other information; the abnormal characteristic index is a quantitative indicator used to describe abnormal characteristics, including the abnormality rate (the frequency of abnormality), the abnormality proportion (the proportion of abnormality in the total operating time), and the abnormality missed misjudgment rate (the probability that the abnormality is not detected or misjudged).

[0035] Specifically, the calling model number constraint is a limit on the number of models determined based on the anomaly rate, which is used to control the complexity of the model and the computing resource requirements; the calling model complexity constraint is a limit on the model complexity determined based on the anomaly missed false positive rate, which is used to balance the accuracy and computing efficiency of the model.

[0036] Specifically, based on the identity identification information of the target emulsion pump, combined with the emulsion knowledge base and historical operation and maintenance logs, an adaptive anomaly assessment model is constructed. First, the emulsion knowledge base is queried through the identity identification information to determine the possible anomaly types of the target emulsion pump; then, the historical operation and maintenance log data of the target scenario where the emulsion pump is located is obtained, and the characteristic indicators related to each anomaly type are analyzed and extracted, including the anomaly rate, anomaly proportion and anomaly missed misjudgment rate. The above-mentioned anomaly characteristic indicators are used to describe the regularity of anomaly occurrence and the difficulty of diagnosis.

[0037] Specifically, the number of model calls for each type of anomaly is defined based on the anomaly rate to ensure that the number of model calls is consistent with the probability of anomaly occurrence; then, the complexity constraint of the model call is defined based on the anomaly omission and false positive rate to ensure the accuracy and reliability of the analysis process, where the number of model calls is proportional to the anomaly rate, and the anomaly omission and false positive rate is proportional to the model complexity constraint; in other words, the more frequently anomalies occur, the more models are called to ensure a full analysis of abnormal behavior. The higher the anomaly omission and false positive rates, the more difficult it is to evaluate this type of anomaly, and a model with higher complexity is required to improve the evaluation accuracy.

[0038] Specifically, according to the constraints on the number of calling models and the complexity of calling models, the benchmark model library is traversed, and the most appropriate combination of multiple benchmark models is selected for each anomaly type to be called, and the calling results are used to construct an anomaly assessment model to achieve accurate assessment and real-time monitoring of the abnormal state of the emulsion pump; among them, the number of models and the model complexity (including the average complexity or the range of complexity) of the combination of benchmark models corresponding to each anomaly type must meet the constraints on the number of calling models and the complexity of calling models.

[0039] The above-mentioned method steps enable rapid identification of anomaly types by combining the emulsion pump's identity information with the knowledge base. Historical operation and maintenance logs are used to generate characteristic indicators, providing a quantitative basis for subsequent model invocation. The dual constraints of model number and model complexity optimize the efficiency and accuracy of model selection, ensuring the reliability of the evaluation results. The resulting anomaly assessment model can monitor and diagnose the operating status of emulsion pumps in real time, providing data support for emulsion pump operation and maintenance. It also improves the accuracy and predictive power of anomaly identification, thereby extending the service life of the target emulsion pump and reducing downtime risks and operation and maintenance costs.

[0040] S300: Acquire multi-source operation and maintenance samples using the identity identification information as a call constraint, and train the anomaly assessment model using the multi-source operation and maintenance samples as supervision.

[0041] Specifically, multi-source operation and maintenance samples refer to operation and maintenance sample data from different sources, including original and homologous operation and maintenance sample data. Among them, original source samples refer to operation and maintenance data directly from the scene where the target emulsion pump is located, and homologous samples refer to operation and maintenance data from other similar scenes or emulsion pumps. Homologous samples are similar to the target scene in characteristics and operating conditions. The multi-source operation and maintenance samples are obtained by sample screening based on identity identification information, providing richer data resources for model training and avoiding the data shortage or bias problems that may be caused by a single data source.

[0042] Specifically, the multi-source operation and maintenance samples include sample pump status data, anomaly assessment data, operation and maintenance record data, etc., which are used to provide learning samples for the anomaly assessment model to learn the ability to recognize anomalies, so that the anomaly assessment model can better adapt to the actual operation of the target emulsion pump and improve the model's recognition accuracy and generalization ability for anomalies.

[0043] In some embodiments, obtaining multi-source operation and maintenance samples using the identity identification information as a call constraint, and using the multi-source operation and maintenance samples as supervision to train the anomaly assessment model includes:

[0044] According to the identity identification information, the homologous constraint and the original source constraint are determined, and the multi-source operation and maintenance samples are extracted accordingly, wherein the homologous constraint is a model constraint, and the original source constraint is a time window constraint; the multi-source operation and maintenance samples are labeled and divided to obtain a classified sample set, wherein the classified sample set includes a plurality of sample subsets corresponding to the anomaly type one by one, and each of the sample subsets includes a negative sample and a positive sample; according to the model scale of the benchmark model, the sample extraction constraint is defined, and random sampling is performed in the plurality of sample subsets; the benchmark models are trained separately according to the sampling results, and the benchmark models that have been trained are integrated to obtain the anomaly assessment model.

[0045] Specifically, the homologous constraint refers to the constraint condition based on the target emulsion pump model, which is used to screen the operation and maintenance samples of other emulsion pumps with the same model as the target emulsion pump; the origin constraint refers to the constraint condition based on the time window, which is used to screen the operation and maintenance samples of the target emulsion pump within a specific time period.

[0046] Specifically, multi-source operation and maintenance samples are classified according to the anomaly type to form multiple sample subsets. Then, the samples in each subset are further divided into positive samples (true anomaly samples) and negative samples (false normal samples and false anomaly samples). Then, according to the scale of each benchmark model (such as model complexity, training data requirements, etc.), the constraints for extracting samples from the classified sample set (such as the number of extracted samples) are defined, and random sampling is performed in each sample subset to obtain a data set suitable for model training (i.e., sampling results).

[0047] Furthermore, the sampling results are used to train multiple benchmark models, and the trained benchmark models are integrated and trained to form a final anomaly assessment model to improve the performance and stability of the model. The anomaly assessment model includes multiple groups of trained benchmark models specialized for various abnormal conditions, which can accurately predict and evaluate various types of anomalies, thereby providing efficient and comprehensive support for the intelligent operation and maintenance of emulsion pumps, and helping to achieve all-round monitoring of the operating status of the target emulsion pump and risk warning.

[0048] The above-mentioned method steps enable the model to learn more comprehensive anomaly features through labeling and classification of multi-source samples, thereby improving the accuracy of anomaly assessment; combining homologous and native samples enhances generalization capabilities; defining sample extraction constraints based on model scale avoids data overload or insufficiency, thereby improving training efficiency and resource utilization; integrating the results of multiple benchmark models through ensemble training methods reduces the bias that may be introduced by a single model, improves the stability and reliability of the model, and thereby ensures that the acquired anomaly assessment model can adapt to emulsion pumps of different models and operating conditions, further improving the adaptability of anomaly assessment.

[0049] S400: Input the real-time status data into an anomaly assessment model to perform anomaly assessment, obtain a predicted anomaly type and a predicted anomaly probability, and output them as an anomaly assessment result, wherein the predicted anomaly type and the predicted anomaly probability correspond one to one.

[0050] Specifically, the abnormality assessment results include multiple possible predicted abnormality types and corresponding predicted abnormality probabilities. The predicted abnormality type refers to the possible fault type of the target emulsion pump predicted by the model, and the predicted abnormality probability refers to the possibility of this type of fault occurring.

[0051] Specifically, the real-time status data is input into the trained anomaly assessment model. The model performs anomaly assessment based on the input data and predicts the possible fault types and probabilities of the target emulsion pump. Each predicted anomaly type corresponds to a predicted anomaly probability, which indicates the confidence level of the anomaly.

[0052] Through the model's reasoning process, potential problems of the target emulsion pump can be evaluated and predicted in real time, and the severity and urgency of the problem can be assessed, providing a basis for subsequent decision-making.

[0053] S500: Make an operation and maintenance decision based on the abnormality assessment result and the preset operation and maintenance judgment constraint, and generate an operation and maintenance control instruction based on the operation and maintenance decision result, and send the operation and maintenance control instruction to the edge device for control execution, wherein the operation and maintenance judgment constraint includes an emergency stop constraint limit and a degradation constraint limit.

[0054] Specifically, the operation and maintenance judgment constraint is a pre-set threshold value for judging the degree of operation risk of the target emulsion pump, including the emergency stop constraint limit and the degradation constraint limit. When the operation risk of the target emulsion pump exceeds the emergency stop constraint limit, it can be considered that the equipment operation needs to be stopped immediately to avoid greater losses; when the operation risk of the target emulsion pump is lower than the emergency stop constraint limit but higher than the degradation constraint limit, it means that the target emulsion pump needs to be downgraded to reduce the risk; the above decision-making based on the abnormal assessment results and the preset operation and maintenance judgment constraints avoids the subjectivity and untimeliness of human judgment, and improves the scientific nature and reliability of operation and maintenance decisions.

[0055] Specifically, the operation and maintenance control instructions are specific instruction information generated based on the operation and maintenance decision results, which are used to guide the edge device to perform specific operations on the emulsion pump, such as emergency stop, degradation, etc., and can be used to notify the operation and maintenance personnel to respond to the operation and maintenance.

[0056] By converting abnormality assessment results into specific operation and maintenance operations, we ensure that timely measures can be taken when abnormalities occur in the target emulsion pump to avoid the expansion of the fault. At the same time, the operating capacity of the target emulsion pump can be maintained as much as possible through measures such as degradation.

[0057] In some embodiments, making an operation and maintenance decision based on the abnormality assessment result and preset operation and maintenance discrimination constraints includes:

[0058] Based on the anomaly assessment result, the predicted anomaly probability is traversed and extracted and normalized to obtain an anomaly weighted coefficient set; the interactive expert system defines the risk severity of each anomaly category to obtain an anomaly risk severity set; based on the anomaly weighted coefficient set and the anomaly risk severity set, a weighted calculation is performed to obtain the risk coefficient corresponding to the anomaly assessment result.

[0059] Specifically, the abnormality assessment results are converted into quantitative risk coefficients to provide a scientific basis for operation and maintenance decisions. First, the predicted abnormality probability is extracted from the abnormality assessment results and normalized to obtain an abnormality weighted coefficient set. Normalization facilitates the comparison of different abnormal probabilities on the same scale; then, through the interactive expert system, based on the impact range and impact degree of each abnormality category, the risk severity is defined for each abnormality category to form an abnormality risk severity set. In other words, the abnormality risk severity set reflects the impact of the abnormality type on the operating safety of the target emulsion pump; then, the abnormality weighted coefficient set and the abnormality risk severity set are weighted and calculated to obtain a comprehensive risk coefficient. This risk coefficient reflects the overall risk level of the current operating status of the target emulsion pump, thereby providing a quantitative basis for subsequent operation and maintenance decisions.

[0060] In the above method steps, normalization and weighted calculation convert complex abnormality assessment results into a unified risk coefficient, making risk assessment more scientific and quantitative; with the help of the risk severity defined by the expert system, expert knowledge is integrated into the risk assessment, which makes up for the shortcomings of relying solely on data and improves the accuracy and reliability of risk assessment. The obtained risk coefficient provides a clear quantitative basis for operation and maintenance decisions, avoids the subjectivity and inconsistency of human judgment, and improves the scientific nature and effectiveness of operation and maintenance decisions.

[0061] In some embodiments, making an operation and maintenance decision based on the abnormality assessment result and a preset operation and maintenance discrimination constraint further includes:

[0062] Based on the emergency stop constraint limit and the degradation constraint limit, the risk coefficient is determined and a decision is triggered. When the risk coefficient is greater than the emergency stop constraint limit, an emergency stop operation is triggered; when the risk coefficient is less than the emergency stop constraint limit and greater than the degradation constraint limit, a degradation decision is executed on the target emulsion pump; when the risk coefficient is less than the degradation constraint limit, a height limit decision is executed on the target emulsion pump; and the operation and maintenance control instruction is generated according to the operation and maintenance decision result, wherein the operation and maintenance control instruction includes instruction priority and instruction parameters.

[0063] Specifically, the emergency stop constraint limit is a preset risk factor threshold. When the operating risk (risk factor) of the target emulsion pump exceeds this threshold, it is determined that the target emulsion pump needs to be stopped immediately to avoid greater losses; the downgrade constraint limit is another preset risk factor threshold. When the operating risk of the target emulsion pump is lower than the emergency stop constraint limit but higher than this threshold, it is determined that the target emulsion pump needs to be downgraded to reduce the risk.

[0064] Specifically, if the risk factor is greater than the emergency stop constraint limit, the emergency stop operation is triggered and the equipment operation is stopped immediately; if the risk factor is less than the emergency stop constraint limit but greater than the degradation constraint limit, the degradation decision is executed and the operating parameters of the equipment are reduced; when the risk factor is lower than the degradation constraint limit, the operating status of the target emulsion pump can be considered relatively safe, but the operating parameters of the target emulsion pump (such as power, load, etc.) still need to be limited to avoid further deterioration of the equipment status and ensure that the target emulsion pump operates within a safe range.

[0065] Furthermore, specific operation and maintenance control instructions are generated based on the decision results, including instruction priority and instruction parameters. The instruction priority is used to determine the execution order of the instructions, and the instruction parameters define the specific operation content (such as stop, speed reduction, power limit, etc.); then, the generated operation and maintenance control instructions are sent to the edge device, and the edge device performs specific control operations to ensure that the device can take timely measures when an abnormality occurs to avoid the expansion of the fault. At the same time, the operation capacity of the equipment is maintained as much as possible through degradation and height limit decisions.

[0066] The above-mentioned method and steps, through emergency stop operation, can immediately stop the equipment operation when the equipment operation risk is too high, effectively avoid the expansion of equipment failure, and ensure the safety of equipment and operators; through degradation and height limit decision-making, when the equipment operation risk is low but intervention is required, appropriate measures can be taken instead of direct shutdown, thereby reducing equipment downtime, providing operation and maintenance personnel with a more sufficient operation and maintenance window, and thus improving operation and maintenance flexibility and efficiency.

[0067] In some implementations, performing a degradation decision on a target emulsion pump further includes:

[0068] A principal component analysis is performed based on the multi-source operation and maintenance samples to determine a degradation parameter sequence, wherein the degradation parameter sequence has a degradation step mark and a degradation probability mark; an objective function is constructed according to the abnormality assessment model and a degradation decision based on the degradation parameter sequence is performed based on the optimization algorithm; the degradation loss is evaluated according to the adaptive degradation decision result; the backup status data of the standby emulsion pump of the target emulsion pump is interactively obtained, and the backup margin is calculated according to the backup status data; the degradation cost is evaluated in combination with the degradation loss and the backup margin, wherein the degradation cost includes lossless degradation and lossy degradation; the adaptive degradation decision result and the degradation cost are output as the operation and maintenance decision result.

[0069] Specifically, principal component analysis is used to extract key degradation parameters from multi-source operation and maintenance samples to reduce data complexity and obtain multiple control parameters (i.e., degradation parameter sequence) that mainly contribute to the performance degradation of the target emulsion pump (reducing the load level of the pump); the degradation parameter sequence is a parameter sequence determined by principal component analysis, which is used to guide the degradation operation of the emulsion pump. The degradation parameter sequence includes multiple degradation parameters and corresponding degradation steps (the magnitude of each degradation) and degradation probabilities (the possibility of performing degradation operations). Among them, the greater the impact of the degradation parameters, the shorter the degradation step and the higher the degradation probability.

[0070] Specifically, the objective function is constructed based on the anomaly assessment model and the degradation parameter sequence to guide the degradation decision. Exemplarily, the objective function is obtained by transfer learning and knowledge distillation of the anomaly assessment model to map the working parameters of the downgraded target emulsion pump to the corresponding downgraded risk coefficient.

[0071] Specifically, using the anomaly assessment model as the basis, the key features and parameters in the model are transferred to the new objective function through transfer learning, which can quickly initialize the objective function and enable it to have preliminary risk assessment capabilities; the subsequent knowledge distillation is used to extract the knowledge of the complex model after transfer learning and pass it to the objective function, further optimizing the performance of the objective function, so that the objective function has better generalization ability and accuracy while maintaining simplicity.

[0072] Specifically, the risk coefficient output by the objective function is used to assess the operational risk of the equipment after the downgrade operation, providing a quantitative basis for downgrade decisions. For example, an optimization algorithm (such as a genetic algorithm or simulated annealing) is used to solve the objective function and generate an adaptive downgrade decision plan. If the risk coefficient after downgrade is still above the preset threshold, further iteration is required to adjust the downgrade strategy.

[0073] Specifically, the degradation loss refers to the performance loss caused during the degradation operation; the backup status data is the working status information of the standby emulsion pump, which is used to evaluate the availability and remaining working capacity of the standby pump, that is, the backup margin. The backup margin reflects the supplementary performance that the standby emulsion pump can provide when the current main pump fails.

[0074] Specifically, the degradation cost refers to the overall evaluation result of the degradation operation after comprehensively considering the degradation loss and backup margin, which is divided into lossless degradation (no significant impact on equipment and production) and lossy degradation (with certain negative impact). For example, if the degradation loss is less than or equal to the backup margin, the current degradation is considered to be lossless degradation. Conversely, if the degradation loss is greater than the backup margin, it means that the pumping capacity of the target scenario has decreased after the current degradation, which is lossy degradation and may require priority processing by operation and maintenance personnel.

[0075] The above method and steps extract key features through principal component analysis to ensure that the degradation decision is not only accurate but also scientific; based on the optimization algorithm, it can find the optimal degradation path in a complex environment, thereby reducing the loss and performance impact on the equipment during the production process; by real-time monitoring of the status data of the standby emulsion pump and evaluating its backup margin, it can ensure that the normal operation of the equipment is not seriously affected during the degradation operation; finally, combining the degradation loss and backup margin to evaluate the degradation cost can provide comprehensive decision-making support for operation and maintenance personnel, so that the degradation operation can minimize production losses and maintain the stability and safety of equipment operation.

[0076] In some embodiments, the method further comprises:

[0077] Continuous monitoring is performed based on the edge device to obtain feedback status information; iterative operation and maintenance decisions are made based on the feedback status information, and if the risk coefficient is still greater than or equal to the degradation constraint limit after a preset number of consecutive iterations, feedback optimization is performed on the objective function.

[0078] Specifically, through continuous monitoring and iterative decision-making, the operation and maintenance strategy is dynamically adjusted, and the objective function is optimized when necessary. First, the operating status of the emulsion pump is continuously monitored in real time through the edge device to obtain the latest feedback status information, including the real-time operating parameters of the equipment and abnormal assessment results; then, based on the feedback status information, the operating risk of the equipment is periodically re-evaluated, and operation and maintenance decisions are made based on the risk coefficient. If the risk coefficient is still greater than or equal to the degradation constraint limit after a preset number of consecutive iterations, it means that the current objective function cannot well reflect the mapping relationship between the risk status of the target emulsion pump and the degraded operating parameters, which leads to the inability to effectively reduce the equipment risk. Therefore, the objective function is further feedback optimized and its parameters or structure are adjusted to improve its assessment ability and adaptability to equipment risks.

[0079] Through the above-mentioned continuous monitoring and iterative decision-making, the operation and maintenance strategy can be dynamically adjusted to adapt to changes in the equipment operating status, thereby improving the flexibility and adaptability of operation and maintenance. At the same time, if the risk factor is still higher than the degradation constraint limit, the objective function is optimized to further improve its performance and accuracy, thereby ensuring the effectiveness of operation and maintenance decisions.

[0080] In summary, the remote operation and maintenance control method for an emulsion pump provided by the present invention has the following technical effects:

[0081] Data is collected by edge devices deployed in the target scenario to obtain real-time status data of the target emulsion pump, wherein the edge device is communicatively connected to the sensor network deployed on the target emulsion pump; based on the identity identification information of the target emulsion pump, a benchmark model is called in combination with a benchmark model library, and an anomaly assessment model is constructed based on the call result; multi-source operation and maintenance samples are obtained using the identity identification information as a call constraint, and the anomaly assessment model is trained using the multi-source operation and maintenance samples as supervision; real-time status data is input into the anomaly assessment model for anomaly assessment, and the predicted anomaly type and predicted anomaly probability are obtained, and the anomaly assessment result is output, wherein the predicted anomaly type and the predicted anomaly probability correspond one to one; an operation and maintenance decision is made based on the anomaly assessment result and the preset operation and maintenance discrimination constraint, and an operation and maintenance control instruction is generated based on the operation and maintenance decision result, and the operation and maintenance control instruction is sent to the edge device for control execution, wherein the operation and maintenance discrimination constraint includes an emergency stop constraint limit and a degradation constraint limit, thereby achieving the technical effect of providing scientific operation and maintenance decision-making and improving the operation and maintenance efficiency and reliability of the emulsion pump.

[0082] Example 2

[0083] Figure 2 This is a schematic diagram of the structure of an emulsion pump remote operation and maintenance control system of the present invention. For example, Figure 1 The flow chart of the remote operation and maintenance control method of an emulsion pump of the present invention can be shown as follows: Figure 2 The structure shown is implemented.

[0084] Based on the same concept as the remote operation and maintenance control method of an emulsion pump in the embodiment, the present invention also provides an emulsion pump remote operation and maintenance control system comprising:

[0085] The data acquisition module 11 is used to collect data through an edge device deployed in a target scene to obtain real-time status data of a target emulsion pump, wherein the edge device is communicatively connected to a sensor network deployed in the target emulsion pump.

[0086] The benchmark model calling module 12 is used to call the benchmark model based on the identification information of the target emulsion pump in combination with the benchmark model library, and to build an anomaly assessment model according to the calling result.

[0087] The anomaly assessment model training module 13 is configured to obtain multi-source operation and maintenance samples using the identity identification information as a call constraint, and train the anomaly assessment model using the multi-source operation and maintenance samples as supervision.

[0088] The anomaly assessment module 14 is used to input the real-time status data into the anomaly assessment model for anomaly assessment, obtain the predicted anomaly type and the predicted anomaly probability, and output them as an anomaly assessment result, wherein the predicted anomaly type and the predicted anomaly probability correspond one to one.

[0089] The operation and maintenance decision execution module 15 is used to make operation and maintenance decisions based on the abnormality assessment results and the preset operation and maintenance judgment constraints, and generate operation and maintenance control instructions based on the operation and maintenance decision results, and send the operation and maintenance control instructions to the edge device for control execution, wherein the operation and maintenance judgment constraints include emergency stop constraint limits and degradation constraint limits.

[0090] The benchmark model calling module 12 includes:

[0091] The abnormality type determination unit is used to determine the type of abnormality that exists based on the identity recognition information interaction emulsion knowledge base.

[0092] The abnormal characteristic indicator acquisition unit is used to obtain the historical operation and maintenance logs of the target scenario, and perform statistical analysis on the historical operation and maintenance logs to obtain the abnormal characteristic indicators corresponding to each abnormality type, wherein the abnormal characteristic indicators include abnormality rate, abnormality proportion, and abnormality missed misjudgment rate.

[0093] The call constraint definition unit is used to define the call model number constraint corresponding to each exception type according to the exception rate, and define the call model complexity constraint corresponding to each exception type according to the exception missed false positive rate.

[0094] The anomaly assessment model construction unit is used to traverse the benchmark model library to call the benchmark model of each anomaly type based on the call model number constraint and the call model complexity constraint, and construct the anomaly assessment model according to the call result.

[0095] In some embodiments, the anomaly assessment model training module 13 includes:

[0096] The multi-source operation and maintenance sample extraction unit is used to determine the homologous constraint and the original source constraint according to the identity identification information, and extract the multi-source operation and maintenance samples accordingly, wherein the homologous constraint is a model constraint and the original source constraint is a time window constraint.

[0097] The classification sample set acquisition unit is used to mark and divide the multi-source operation and maintenance samples to obtain a classification sample set, wherein the classification sample set includes a plurality of sample subsets corresponding to the abnormality types one by one, and each of the sample subsets includes negative samples and positive samples.

[0098] The sample extraction constraint definition unit is used to define sample extraction constraints according to the model scale of the benchmark model and perform random sampling in a plurality of the sample subsets accordingly.

[0099] The anomaly assessment model training unit is used to train the benchmark models respectively according to the sampling results, and integrate the trained benchmark models to obtain the anomaly assessment model.

[0100] In some embodiments, the operation and maintenance decision execution module 15 includes:

[0101] The abnormality weighted coefficient set acquisition unit is used to traverse and extract the predicted abnormality probability based on the abnormality assessment result and perform normalization to obtain the abnormality weighted coefficient set.

[0102] The abnormal risk severity set acquisition unit is used to define the risk severity of each abnormal category through the interactive expert system and acquire the abnormal risk severity set.

[0103] The risk coefficient calculation unit is used to obtain the risk coefficient corresponding to the abnormal assessment result by weighted calculation based on the abnormal weight coefficient set and the abnormal risk severity set.

[0104] In some embodiments, the operation and maintenance decision execution module 15 includes:

[0105] The risk coefficient determination and decision triggering unit is used to determine the risk coefficient and trigger a decision based on the emergency stop constraint limit and the degradation constraint limit, and trigger an emergency stop operation when the risk coefficient is greater than the emergency stop constraint limit.

[0106] The degradation decision execution unit is configured to execute a degradation decision on the target emulsion pump when the risk coefficient is less than the emergency stop constraint limit and greater than the degradation constraint limit.

[0107] The height limit decision execution unit is used to execute the height limit decision on the target emulsion pump when the risk coefficient is less than the degradation constraint limit.

[0108] An operation and maintenance control instruction generating unit is used to generate the operation and maintenance control instruction according to the operation and maintenance decision result, wherein the operation and maintenance control instruction includes an instruction priority and an instruction parameter.

[0109] In some implementations, the downgrade decision execution unit in the operation and maintenance decision execution module 15 includes:

[0110] The degradation parameter sequence determining unit is configured to perform principal component analysis based on the multi-source operation and maintenance samples to determine a degradation parameter sequence, wherein the degradation parameter sequence has a degradation step size mark and a degradation probability mark.

[0111] A degradation decision execution unit is used to construct an objective function according to the abnormality assessment model and the degradation parameter sequence to perform a degradation decision based on an optimization algorithm.

[0112] The degradation loss evaluation unit is used to evaluate the degradation loss according to the adaptive degradation decision result.

[0113] The backup status data acquisition and backup margin calculation unit is used to interactively acquire the backup status data of the backup emulsion pump of the target emulsion pump and calculate the backup margin according to the backup status data.

[0114] The degradation cost evaluation unit is configured to evaluate the degradation cost by combining the degradation loss and the backup margin, wherein the degradation cost includes lossless degradation and lossy degradation.

[0115] The operation and maintenance decision result output unit is used to output the adaptive degradation decision result and the degradation cost as the operation and maintenance decision result.

[0116] In some embodiments, the system also includes an iteration and feedback optimization unit, which is used to: perform continuous monitoring based on the edge device to obtain feedback status information; make iterative operation and maintenance decisions based on the feedback status information, and if the risk coefficient is still greater than or equal to the degradation constraint limit after a preset number of consecutive iterations, feedback optimization is performed on the objective function.

[0117] It should be understood that the embodiments mentioned in this specification focus on their differences from other embodiments. The specific embodiments in the aforementioned embodiment one are also applicable to the remote operation and maintenance control system of an emulsion pump described in embodiment two. For the sake of brevity of the specification, no further elaboration will be given here.

[0118] It should be understood that the embodiments disclosed in the present invention and the above description can enable those skilled in the art to use the present invention to implement the present invention. At the same time, the present invention is not limited to the embodiments mentioned above. It should be understood that those skilled in the art can still modify the technical solutions described in the above embodiments or replace some of the technical features therein with equivalents; and such modifications or replacements do not deviate from the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention and are all included in the scope of protection of the present invention.

Claims

1. A remote operation and maintenance control method for an emulsion pump, characterized in that: The method comprises: Acquire real-time status data of a target emulsion pump by collecting data through an edge device deployed in a target scenario, wherein the edge device is communicatively connected to a sensor network deployed in the target emulsion pump; Based on the identification information of the target emulsion pump, the benchmark model is called in combination with the benchmark model library, and an anomaly assessment model is constructed based on the call results; Acquire multi-source operation and maintenance samples using the identity identification information as a call constraint, and train the anomaly assessment model using the multi-source operation and maintenance samples as supervision; Input the real-time status data into the anomaly assessment model to perform anomaly assessment, obtain the predicted anomaly type and the predicted anomaly probability, and output them as an anomaly assessment result, wherein the predicted anomaly type and the predicted anomaly probability correspond one to one; An operation and maintenance decision is made based on the abnormality assessment result and the preset operation and maintenance discrimination constraint, and an operation and maintenance control instruction is generated based on the operation and maintenance decision result, and the operation and maintenance control instruction is issued to the edge device for control execution, wherein the operation and maintenance discrimination constraint includes an emergency stop constraint limit and a degradation constraint limit; Based on the identification information of the target emulsion pump, the benchmark model is called in combination with the benchmark model library, and an anomaly assessment model is constructed based on the call results, including: Determining the type of anomaly present based on the identity information interactive emulsion knowledge base; Obtain historical operation and maintenance logs for the target scenario, and perform statistical analysis on the logs to obtain abnormal characteristic indicators corresponding to each abnormality type, wherein the abnormal characteristic indicators include abnormality rate, abnormality proportion, and abnormality missed false positive rate; Based on the exception rate, define the number of call models corresponding to each exception type, and based on the exception missed false positive rate, define the complexity constraint of the call model corresponding to each exception type; The benchmark model library is traversed to call the benchmark model of each anomaly type according to the call model number constraint and the call model complexity constraint, and the anomaly assessment model is constructed according to the call results.

2. The remote operation and maintenance control method for an emulsion pump according to claim 1, characterized in that: Acquiring multi-source operation and maintenance samples using the identity identification information as a call constraint, and training the anomaly assessment model using the multi-source operation and maintenance samples as supervision, including: Determine a homologous constraint and an original source constraint according to the identity identification information, and extract the multi-source operation and maintenance samples accordingly, wherein the homologous constraint is a model constraint, and the original source constraint is a time window constraint; Marking and dividing the multi-source operation and maintenance samples to obtain a classified sample set, wherein the classified sample set includes a plurality of sample subsets corresponding one-to-one to the anomaly type, and each of the sample subsets includes a negative sample and a positive sample; Defining sample extraction constraints according to the model scale of the benchmark model, and performing random sampling in a plurality of the sample subsets accordingly; The benchmark models are trained separately according to the sampling results, and the trained benchmark models are integrated to obtain the anomaly assessment model.

3. The remote operation and maintenance control method for an emulsion pump according to claim 2, characterized in that: Making an operation and maintenance decision based on the abnormality assessment results and the preset operation and maintenance discrimination constraints, including: Based on the anomaly assessment result, traverse and extract the predicted anomaly probability and perform normalization to obtain an anomaly weighted coefficient set; The interactive expert system defines the risk severity of each abnormal category and obtains the abnormal risk severity set; According to the abnormality weighting coefficient set and the abnormality risk severity set, a weighted calculation is performed to obtain the risk coefficient corresponding to the abnormality assessment result.

4. The remote operation and maintenance control method for an emulsion pump according to claim 3, characterized in that: Making an operation and maintenance decision based on the abnormality assessment results and the preset operation and maintenance discrimination constraints, including: Based on the emergency stop constraint limit and the degradation constraint limit, the risk coefficient is determined and a decision is triggered, and when the risk coefficient is greater than the emergency stop constraint limit, an emergency stop operation is triggered; When the risk coefficient is less than the emergency stop constraint limit and greater than the degradation constraint limit, executing a degradation decision on the target emulsion pump; When the risk coefficient is less than the degradation constraint limit, executing a height limit decision on the target emulsion pump; The operation and maintenance control instruction is generated according to the operation and maintenance decision result, wherein the operation and maintenance control instruction includes an instruction priority and an instruction parameter.

5. The remote operation and maintenance control method for an emulsion pump according to claim 4, characterized in that: Perform degradation decisions on targeted emulsion pumps, also including: Performing principal component analysis based on the multi-source operation and maintenance samples to determine a degradation parameter sequence, wherein the degradation parameter sequence has a degradation step size mark and a degradation probability mark; Constructing an objective function according to the abnormality assessment model and performing a degradation decision based on the degradation parameter sequence based on an optimization algorithm; Evaluate the degradation loss based on the adaptive degradation decision result; interactively acquiring backup status data of a backup emulsion pump of the target emulsion pump, and calculating a backup margin based on the backup status data; evaluating a degradation cost based on the degradation loss and the backup margin, wherein the degradation cost includes lossless degradation and lossy degradation; The adaptive degradation decision result and the degradation cost are output as the operation and maintenance decision result.

6. The remote operation and maintenance control method for an emulsion pump according to claim 5, characterized in that: The method further comprises: Continuously monitor the edge device to obtain feedback status information; An iterative operation and maintenance decision is made according to the feedback status information. If the risk coefficient is still greater than or equal to the degradation constraint limit after a preset number of consecutive iterations, feedback optimization is performed on the objective function.

7. An emulsion pump remote operation and maintenance control system, characterized in that: The system is used to execute the remote operation and maintenance control method of an emulsion pump according to any one of claims 1 to 6, and the system includes: A data acquisition module, configured to acquire real-time status data of a target emulsion pump by acquiring data through an edge device deployed in a target scenario, wherein the edge device is communicatively connected to a sensor network deployed on the target emulsion pump; The benchmark model calling module is used to call the benchmark model based on the identification information of the target emulsion pump and the benchmark model library, and to build an anomaly assessment model based on the call results; An anomaly assessment model training module is used to obtain multi-source operation and maintenance samples using the identity identification information as a call constraint, and train the anomaly assessment model using the multi-source operation and maintenance samples as supervision; An anomaly assessment module, configured to input the real-time status data into an anomaly assessment model for anomaly assessment, obtain a predicted anomaly type and a predicted anomaly probability, and output the result as an anomaly assessment result, wherein the predicted anomaly type and the predicted anomaly probability correspond one to one; An operation and maintenance decision execution module is used to make operation and maintenance decisions based on the abnormality assessment results and preset operation and maintenance discrimination constraints, and generate operation and maintenance control instructions based on the operation and maintenance decision results, and send the operation and maintenance control instructions to the edge device for control execution, wherein the operation and maintenance discrimination constraints include emergency stop constraint limits and degradation constraint limits.

Images