Clean area external pump remote sampling control method and system based on double pump redundancy switching

Abstract

The application discloses a clean area external pump remote sampling control method and system based on double-pump redundancy switching, realizes accurate quantitative evaluation and millisecond-level fault identification of the double-pump state by constructing a health degree scoring model based on multi-dimensional sensing data, simultaneously, utilizes an adaptive valve timing control strategy based on real-time pressure difference feedback to intelligently adjust the switching rate to suppress the water hammer effect, overcomes the pipeline damage and flow fluctuation problems caused by the lack of fine control of the traditional double-pump switching, and in addition, realizes load balancing during the cooperative work of the double pumps in combination with a health degree dynamic distribution mechanism, avoids excessive equipment wear and tear, thereby significantly prolongs the service life of the system and guarantees the continuity and data integrity of the clean area sampling.

Description

Technical Field

[0001] This invention relates to the field of online monitoring, sampling and control technology, specifically to a method and system for remote sampling control of external pumps in clean areas based on dual-pump redundancy switching. Background Technology

[0002] External pump sampling systems for clean areas are widely used in fields with extremely high requirements for environmental cleanliness, such as pharmaceuticals, biosafety, and semiconductor manufacturing. Their core task is to continuously and stably collect air or liquid samples without disrupting the slightly positive pressure environment of the clean area. Existing external pump sampling systems for clean areas mostly operate in a single-pump independent mode. If the main pump suddenly fails or its performance deteriorates, the system will immediately lose its sampling capability, resulting in sampling interruption.

[0003] Furthermore, some existing systems that incorporate dual-pump redundancy lack fine-grained control during switching, often employing a fixed rate or a simple "off then on" logic, failing to detect the pressure difference between the two pumps in real time. When there is a significant pressure difference between the two pumps, a forced rapid switch can easily cause drastic pressure surges in the pipeline, inducing water hammer. This can not only damage delicate sampling pipelines and valve assemblies but also lead to drastic fluctuations in sampling flow, severely impacting the accuracy and representativeness of the sampling data. Simultaneously, existing technologies lack an adaptive load distribution mechanism based on equipment health, making it difficult to achieve load balancing when two pumps are working collaboratively, resulting in a shortened overall equipment lifespan. Summary of the Invention

[0004] The present invention aims to provide a remote sampling control method and system for external pumps in clean areas based on dual-pump redundancy switching, which aims to overcome the problems of low reliability in single-pump mode, water hammer effect and unstable flow during dual-pump switching due to large pressure difference, and lack of adaptive load balancing based on health status in existing systems.

[0005] To achieve the above objectives, the technical solution adopted by this invention is: a remote sampling control method for external pumps in clean areas based on dual-pump redundancy switching, comprising: The real-time multi-dimensional sensor data streams of the main pump and the standby pump are acquired, and after low-pass filtering, the operating deviation value and vibration amplitude normalization factor of each pump relative to the rated operating condition are calculated. Based on the operating deviation value and vibration amplitude normalization factor, a weighted fusion health score model is constructed to calculate the health score of the main pump and the health score of the standby pump. The health scores of the main pump and the standby pump are then compared with a preset dynamic threshold to determine the initial operating mode. Based on the initial operating mode, establish the mapping relationship between the terminal demand flow and the inlet pressure. Combine the health scores of the main pump and the standby pump to calculate the pressure gradient difference and flow compensation requirement of the two pumps. Generate and execute the health-based dual pump speed allocation command to monitor the instantaneous flow response. Based on the instantaneous flow response, monitor the mutation rate of dual pump operating parameters and determine the fault level. Calculate the flow buffer capacity requirement during the switching process based on the determination result, and execute valve timing control and pressure wave suppression strategies to verify the switching completion status and lock the new operating mode. In the new operating mode, a timestamp-based sample data stream synchronization mechanism is constructed, segmented encryption and hash integrity verification are implemented, and adaptive bandwidth allocation and packet loss retransmission strategies are executed to verify the data consistency of the remote receiver and update the status log. Based on the status log records, multidimensional abnormal features are identified and fault levels are classified. Local isolation and bypass redundancy activation are performed, self-diagnostic tests and component replacement verification are triggered to generate self-healing reports, full-cycle operation data is aggregated to build health profiles, operation trends are explored to predict potential failure risks, dynamic preventive maintenance plans are formulated, and control parameters are optimized.

[0006] Preferably, the acquisition of real-time multi-dimensional sensor data streams from the main pump and the standby pump includes: Raw data streams are synchronously acquired from a multi-source sensor array deployed on the main pump and the standby pump. The raw data streams include the main pump outlet pressure, the standby pump outlet pressure, the main pump motor phase current, the standby pump motor phase current, the main pump bearing vibration acceleration, and the standby pump bearing vibration acceleration. The acquired raw voltage signal is smoothed by using the sliding window length and sampling time interval to eliminate environmental noise interference and obtain the filtered sensor reading.

[0007] Preferably, the calculation of the operating deviation value and vibration amplitude normalization factor of each pump relative to the rated operating condition includes: The pressure deviation is obtained by subtracting the rated reference pressure from the measured main pump outlet pressure and the standby pump outlet pressure. The current deviation is obtained by subtracting the rated reference current from the measured main pump motor phase current and the standby pump motor phase current. The vibration amplitude normalization factor is obtained by dividing the measured vibration acceleration of the main pump bearing and the standby pump bearing by the upper limit of the safety threshold. The pressure deviation, current deviation, and vibration amplitude normalization factor directly reflect whether the pump body is overloaded, blocked, or mechanically worn.

[0008] Preferably, the construction of the weighted fusion health score model includes: Pressure stability, current stability, and vibration level are taken as three core dimensions, and weights are set for pressure dimension, current dimension, and vibration dimension, and the sum of the three is equal to 1. Using the pressure deviation, current deviation, vibration amplitude normalization factor, and weighting coefficient, the health score of the main pump and the health score of the standby pump are calculated respectively. The calculated health score is converted into a dimensionless value between 0 and 1. The closer the value is to 1, the healthier the pump is.

[0009] Preferably, determining the initial operating mode includes: If the health score of the main pump is greater than or equal to the preset lower health threshold and greater than the health score of the standby pump, then the main pump priority mode is confirmed. If the health score of the main pump is less than the preset lower health threshold or the health score of the standby pump is greater than the health score of the main pump, then the dual-pump collaborative mode or standby priority mode will be entered. The judgment results guide the opening and closing actions of the valves and establish the initial operating benchmark of the system.

[0010] Preferably, the traffic buffer capacity requirement during the calculation switching process includes: The required buffer capacity is calculated based on the switching time and the current target flow rate, combined with the fluid loss coefficient; the value of the fluid loss coefficient is adjusted according to the compressibility of the liquid in the pipeline, and the start-up timing of the standby pump is planned in advance.

[0011] Preferably, the valve timing control and pressure wave suppression strategy includes: Following the principle of opening before closing, first open the valve on the standby pump side, and then close the valve on the main pump side after the pressure is balanced. Calculate the pressure difference between the main pump outlet pressure and the standby pump outlet pressure in real time; When the pressure difference exceeds a preset pressure difference threshold, the valve opening change rate is reduced to slow down the propagation speed of the pressure wave. When the pressure difference does not exceed the preset pressure difference threshold, the valve opening change rate is increased to improve switching efficiency.

[0012] Preferably, the identification of multidimensional abnormal features and the classification of fault levels includes: Continuously monitor the pressure and current change rates of the main pump and standby pump to detect sudden failures; The integral result of the pressure mutation rate is calculated, and when the integral result exceeds the severe fault threshold, it is judged as a level one fault; When the integral result is between the general fault threshold, it is marked as a level 2 fault and an early warning observation is initiated.

[0013] Preferably, the prediction of potential failure risks in the mining operation trend includes: The rate of equipment degradation is determined by analyzing the slope of the health score over time. Estimate remaining lifespan by dividing the current health status and critical failure threshold by the rate of health decline. Analyze the periodicity of failure occurrences to identify whether failures are more likely to occur under certain temperature or humidity conditions.

[0014] On the other hand, this invention proposes a remote sampling control system for external pumps in clean areas based on dual-pump redundancy switching, comprising: The data acquisition unit is used to acquire real-time multi-dimensional sensor data streams from the main pump and the standby pump. After low-pass filtering, it calculates the operating deviation value of each pump relative to the rated operating condition and the vibration amplitude normalization factor. The status assessment unit is used to construct a weighted fusion health score model based on the operating deviation value and vibration amplitude normalization factor, calculate the main pump health score and the standby pump health score, and compare the main pump health score and the standby pump health score with a preset dynamic threshold to determine the initial operating mode. The flow control unit is used to establish a mapping relationship between the terminal demand flow and the inlet pressure according to the initial operating mode, calculate the pressure gradient difference between the two pumps and the flow compensation requirement by combining the health score of the main pump and the health score of the standby pump, and generate and execute the health-based dual pump speed allocation command to monitor the instantaneous flow response. The redundant switching unit is used to monitor the mutation rate of the dual pump operating parameters and determine the fault level based on the instantaneous flow response, calculate the flow buffer capacity requirement during the switching process based on the determination result, and execute valve timing control and pressure wave suppression strategies to verify the switching completion status and lock the new operating mode. The data transmission unit is used to build a timestamp-based sample data stream synchronization mechanism in the new operating mode, implement segmented encryption and hash integrity verification, and execute adaptive bandwidth allocation and packet loss retransmission strategies to verify the data consistency of the remote receiving end and update the status log. The fault-tolerant recovery unit is used to identify multi-dimensional abnormal features and classify fault levels based on the status log records, perform local isolation and bypass redundancy activation, and trigger self-diagnostic tests and component replacement verification to generate a self-healing report. The record management unit is used to aggregate full-cycle operation data to build health records, identify operational trends to predict potential failure risks, formulate dynamic preventive maintenance plans, and provide feedback to optimize control parameters.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention achieves precise quantitative assessment of dual-pump status and millisecond-level fault identification by constructing a health scoring model based on multi-dimensional sensor data. Simultaneously, it utilizes an adaptive valve timing control strategy based on real-time pressure difference feedback to intelligently adjust the switching rate to suppress water hammer effects, overcoming the problems of pipeline damage and flow fluctuations caused by the lack of fine control in traditional dual-pump switching. Furthermore, combined with a dynamic health allocation mechanism, it achieves load balancing during dual-pump collaborative operation, avoiding excessive equipment wear and tear, thereby significantly extending the system's lifespan and ensuring the continuity and integrity of cleanroom sampling. Attached Figure Description

[0016] Figure 1 This is a flowchart of the remote sampling control method for external pumps in clean areas based on dual-pump redundancy switching, as described in this invention. Figure 2 This is a block diagram of the cleanroom external pump remote sampling control system based on dual-pump redundancy switching, as described in this invention. Detailed Implementation

[0017] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0018] like Figure 1 As shown, this invention proposes a remote sampling control method for external pumps in clean areas based on dual-pump redundancy switching. This method achieves load balancing during dual-pump collaborative operation, avoids excessive equipment wear, significantly extends system lifespan, and ensures the continuity and data integrity of clean area sampling. Specifically, it includes: The real-time multi-dimensional sensor data streams of the main pump and the standby pump are acquired, and after low-pass filtering, the operating deviation value and vibration amplitude normalization factor of each pump relative to the rated operating condition are calculated. The process of acquiring real-time multi-dimensional sensor data streams from the main pump and the standby pump includes: synchronously acquiring raw data streams from multi-source sensor arrays deployed on the main pump and the standby pump. The raw data streams include the main pump outlet pressure, the standby pump outlet pressure, the main pump motor phase current, the standby pump motor phase current, the main pump bearing vibration acceleration, and the standby pump bearing vibration acceleration. The acquired raw voltage signals are smoothed using the sliding window length and sampling time interval to eliminate environmental noise interference and obtain filtered sensor readings.

[0019] By using low-pass filtering and sliding window smoothing, the interference of environmental noise on the original signal is effectively eliminated, significantly improving the calculation accuracy of the operating deviation value and the vibration amplitude normalization factor. This high signal-to-noise ratio data preprocessing lays a solid foundation for the subsequent construction of an accurate health scoring model, ensuring that the system can keenly capture early minor anomalies of the equipment, thereby avoiding misjudgment or missed reporting due to data distortion and ensuring the reliability of fault identification.

[0020] The calculation of the operating deviation and vibration amplitude normalization factor of each pump relative to the rated operating conditions includes: subtracting the rated reference pressure from the measured main pump outlet pressure and standby pump outlet pressure to obtain the pressure deviation; subtracting the rated reference current from the measured main pump motor phase current and standby pump motor phase current to obtain the current deviation; dividing the measured main pump bearing vibration acceleration and standby pump bearing vibration acceleration by the upper limit of the safety threshold to obtain the vibration amplitude normalization factor; and visually reflecting whether the pump body is overloaded, blocked, or mechanically worn based on the pressure deviation, current deviation, and vibration amplitude normalization factor.

[0021] This calculation method transforms multidimensional physical quantities into intuitive standardized indicators, enabling precise quantitative characterization of typical faults such as pump overload, blockage, and mechanical wear. It directly identifies abnormal fluid resistance and electrical load imbalance through pressure and current deviations, and sensitively captures early signs of mechanical wear using vibration normalization factors. This provides high-confidence feature inputs for health scoring, significantly improving the accuracy and response speed of fault diagnosis.

[0022] A weighted fusion health score model is constructed based on the operating deviation value and vibration amplitude normalization factor. The health scores of the main pump and the standby pump are calculated, and the health scores of the main pump and the standby pump are compared with the preset dynamic threshold to determine the initial operating mode. The weighted fusion health rating model includes: taking pressure stability, current stability, and vibration level as three core dimensions, setting weights for pressure, current, and vibration dimensions, and ensuring that the sum of the three is equal to 1; using pressure deviation, current deviation, vibration amplitude normalization factors, and weight coefficients to calculate the health rating of the main pump and the standby pump respectively; and converting the calculated health rating into a dimensionless value between 0 and 1, with the value closer to 1 indicating a healthier pump condition.

[0023] This weighted fusion model effectively overcomes the risk of misjudgment by a single parameter through the coordinated quantification of multi-dimensional indicators, and achieves an accurate profile of the equipment's health status. It transforms complex operating data into intuitive dimensionless scores, which not only provides an objective and quantitative basis for decision-making on the initial operating mode (selection of main / standby pumps), but also significantly improves the system's early identification capability for potential faults and its robustness against interference.

[0024] The determination of the initial operating mode includes: if the health score of the main pump is greater than or equal to the preset lower health threshold and is greater than the health score of the standby pump, then it is confirmed as the main pump priority mode; if the health score of the main pump is less than the preset lower health threshold or the health score of the standby pump is greater than the health score of the main pump, then it enters the dual-pump collaborative mode or the standby priority mode; the valve opening and closing actions are guided according to the determination results to establish the initial operating benchmark of the system.

[0025] This dynamic judgment mechanism enables intelligent decision-making based on the real-time health status of the equipment, ensuring that the system prioritizes the operation of the optimal pump group and maximizes the continuity and reliability of sampling. By automatically switching between main / standby modes or activating the dual-pump collaborative strategy, it effectively avoids the risk of single-point failure and provides a precise initial benchmark for valve timing control, significantly improving the system's adaptability and overall operational stability.

[0026] Based on the initial operating mode, a mapping relationship between terminal demand flow and inlet pressure is established. Combining the health scores of the main pump and the standby pump, the pressure gradient difference between the two pumps and the flow compensation requirement are calculated. A health-based dual-pump speed allocation command is generated and executed to monitor instantaneous flow response. By dynamically balancing the dual-pump speed allocation, the terminal flow demand is accurately matched and inlet pressure fluctuations are effectively suppressed. The health score is used to compensate for system deviations in real time, ensuring that the flow remains stable even when equipment performance degrades or load changes suddenly. This significantly improves the system's adjustment accuracy, response speed, and energy efficiency level of multi-pump collaborative operation.

[0027] Based on the instantaneous flow response, monitor the mutation rate of dual pump operating parameters and determine the fault level. Calculate the flow buffer capacity requirement during the switching process based on the determination results, and execute valve timing control and pressure wave suppression strategies to verify the switching completion status and lock the new operating mode. The calculation of flow buffer capacity requirements during the switching process includes: calculating the required buffer capacity based on the switching time and the current target flow rate, combined with the fluid loss coefficient; adjusting the value of the fluid loss coefficient based on the compressibility of the liquid in the pipeline; and planning the start-up timing of the standby pump in advance.

[0028] By monitoring parameter mutation rates in real time, graded early warning of faults is achieved, significantly improving the system's response sensitivity to sudden anomalies. The introduction of dynamic calculation of buffer capacity based on fluid compressibility and switching time accurately plans the start-up timing of the backup pump and valve sequence, effectively eliminating flow oscillations and pressure shocks during the switching process. Ultimately, it ensures a smooth transition of operating modes and guarantees the system's continuous and stable water supply capability under complex operating conditions.

[0029] The valve timing control and pressure wave suppression strategy includes: following the principle of opening before closing, opening the standby pump side valve first, and closing the main pump side valve after the pressure is balanced; calculating the pressure difference between the main pump outlet pressure and the standby pump outlet pressure in real time; reducing the valve opening rate when the pressure difference exceeds the preset pressure difference threshold to slow down the propagation speed of the pressure wave; and increasing the valve opening rate when the pressure difference does not exceed the preset pressure difference threshold to improve switching efficiency.

[0030] By using a pre-opening and post-closing mechanism and a dynamic pressure difference feedback mechanism, seamless connection of the dual-pump switching process is achieved, effectively eliminating water hammer effect and system pressure shock caused by asynchronous valve actions. The valve action rate is adaptively adjusted according to the real-time pressure difference, achieving the best balance between ensuring pipeline safety (suppressing pressure waves) and improving switching efficiency, significantly enhancing the system's operational stability and reliability during mode switching.

[0031] In the new operating mode, a timestamp-based sample data stream synchronization mechanism is built, segmented encryption and hash integrity verification are implemented, and adaptive bandwidth allocation and packet loss retransmission strategies are executed to verify the data consistency of the remote receiver and update the status log. By using timestamp synchronization and segmented encryption verification, high consistency and security of remote data during transmission are ensured, effectively preventing data tampering or loss. Combined with adaptive bandwidth allocation and packet loss retransmission strategies, the robustness and real-time performance of communication in weak network environments are significantly improved, providing a reliable data foundation for accurate updates of system status logs and ensuring the accuracy of remote monitoring and decision-making.

[0032] Based on the status log records, identify multidimensional abnormal characteristics and classify fault levels, perform local isolation and bypass redundancy activation, trigger self-diagnostic tests and component replacement verification to generate self-healing reports, aggregate full-cycle operation data to build health profiles, mine operation trends to predict potential failure risks, formulate dynamic preventive maintenance plans and provide feedback to optimize control parameters.

[0033] The system identifies multidimensional anomalies and classifies fault levels, including: continuously monitoring the pressure and current change rates of the main and standby pumps to detect sudden faults; calculating the integral result of the pressure mutation rate, classifying it as a Level 1 fault when the integral result exceeds the severe fault threshold; and marking it as a Level 2 fault and initiating early warning observation when the integral result is between the general fault threshold. Through multidimensional feature fusion and integral judgment mechanisms, the system achieves millisecond-level accurate identification and graded early warning of sudden faults, significantly reducing the false alarm rate. Combined with local isolation, bypass redundancy, and a self-diagnostic and self-healing closed loop, the system's survivability and continuous operational reliability are significantly improved during fault occurrences.

[0034] Among these, predicting potential failure risks by mining operational trends includes: determining the rate of equipment degradation by analyzing the slope of the health score over time; estimating the remaining lifespan by dividing the current health score and critical failure threshold by the rate of health score decline; and analyzing the periodicity of failure occurrences to identify whether failures are more likely to occur under certain temperature or humidity conditions.

[0035] By quantifying the slope of health degradation and dynamically estimating remaining lifespan, we have achieved a forward-looking and accurate early warning of equipment failure risks, effectively avoiding sudden downtime. Combined with the periodic failure correlation analysis of environmental factors (temperature / humidity), we have revealed potential operational causes, enabling maintenance plans to be upgraded from regular inspections to condition-based targeted interventions, significantly reducing unplanned downtime and optimizing the entire lifecycle maintenance cost.

[0036] On the other hand, this invention proposes a remote sampling control system for external pumps in clean areas based on dual-pump redundancy switching, such as... Figure 2 As shown, it includes: The data acquisition unit is used to acquire real-time multi-dimensional sensor data streams from the main pump and the standby pump. After low-pass filtering, it calculates the operating deviation value of each pump relative to the rated operating condition and the vibration amplitude normalization factor. The status assessment unit is used to construct a weighted fusion health score model based on the operating deviation value and vibration amplitude normalization factor, calculate the main pump health score and the standby pump health score, and compare the main pump health score and the standby pump health score with a preset dynamic threshold to determine the initial operating mode. The flow control unit is used to establish a mapping relationship between the terminal demand flow and the inlet pressure according to the initial operating mode, calculate the pressure gradient difference between the two pumps and the flow compensation requirement by combining the health score of the main pump and the health score of the standby pump, and generate and execute the health-based dual pump speed allocation command to monitor the instantaneous flow response. The redundant switching unit is used to monitor the mutation rate of the dual pump operating parameters and determine the fault level based on the instantaneous flow response, calculate the flow buffer capacity requirement during the switching process based on the determination result, and execute valve timing control and pressure wave suppression strategies to verify the switching completion status and lock the new operating mode. The data transmission unit is used to build a timestamp-based sample data stream synchronization mechanism in the new operating mode, implement segmented encryption and hash integrity verification, and execute adaptive bandwidth allocation and packet loss retransmission strategies to verify the data consistency of the remote receiving end and update the status log. The fault-tolerant recovery unit is used to identify multi-dimensional abnormal features and classify fault levels based on the status log records, perform local isolation and bypass redundancy activation, and trigger self-diagnostic tests and component replacement verification to generate a self-healing report. The record management unit is used to aggregate full-cycle operation data to build health records, identify operational trends to predict potential failure risks, formulate dynamic preventive maintenance plans, and provide feedback to optimize control parameters.

[0037] Furthermore, the above modules, during execution, are also used to implement other steps of the aforementioned method for remote sampling control of external pumps in clean areas based on dual-pump redundancy switching, as follows: Step 1: Initial state perception and dual-pump health quantification assessment; During the startup phase of a remote sampling system for external pumps in a cleanroom, the primary task is to establish a comprehensive understanding of the current physical state of the system, particularly by quantitatively assessing the health status of the main and standby pumps to provide data support for subsequent decision-making. This step is not a simple power-on / off check, but rather involves constructing a comprehensive health index based on real-time collected multi-dimensional sensor data. This index directly determines whether the system should prioritize the main pump or immediately switch to standby mode. This process relies on the previously deployed sensor network, calculating the instantaneous health score of each pump by reading pressure, vibration, and current signals, thereby establishing the initial operating baseline of the system. Specifically, this includes: 1.1 Acquire real-time multidimensional sensor data stream from the dual pumps; First, the system needs to synchronously acquire raw data streams from multi-source sensor arrays deployed on the main pump and standby pump. These data streams include the main pump outlet pressure. Standby pump outlet pressure Main pump motor phase current Standby pump motor phase current and the vibration acceleration of the main pump bearing and the vibration acceleration of the spare pump bearing The data acquisition frequency is set to... That is, collecting data per second Several data points were selected to ensure the capture of high-frequency transient fluctuations. To eliminate environmental noise interference, the acquired raw voltage signal V(t) was low-pass filtered to obtain the filtered signal. The calculation formula is as follows: ; Where N is the length of the sliding window. The sampling time interval, This represents the sensor reading after smoothing. This preprocessing step ensures that the data used for subsequent calculations is highly stable and representative, avoiding misjudgments caused by instantaneous spike pulses.

[0038] 1.2 Calculate the instantaneous operating deviation of each pump; After obtaining the filtered multidimensional data, the next step is to calculate the operating deviation of each pump relative to its rated operating condition at the current moment. For the main pump, its pressure deviation is defined. and current deviation For the standby pump, define its pressure deviation. and current deviation These deviation values ​​reflect the degree to which the pump's actual performance under the current load deviates from the standard operating point.

[0039] The specific calculation method is to subtract the rated reference value from the measured value, for example, the main pump pressure deviation. ,in The rated outlet pressure is determined by the system design. Simultaneously, the vibration amplitude normalization factor is calculated based on vibration data. ,in This represents the upper limit of the safety threshold. Calculating these deviation values ​​provides a clear indication of whether the pump body exhibits abnormal signs such as overload, blockage, or mechanical wear, offering quantitative input for comprehensive scoring.

[0040] 1.3 Construct a health score model based on weighted fusion; Based on the aforementioned deviation values ​​and normalization factors, a comprehensive health rating model is constructed to quantify the overall health status of the main pump and standby pump. This model employs a weighted linear fusion strategy, using pressure stability, current stability, and vibration level as three core dimensions. Main pump health rating. The calculation formula is defined as follows: in, The weighting coefficients for pressure, current, and vibration dimensions are respectively, satisfying... Similarly, the standby pump health score The same formula structure is used for calculation. The weighting coefficients are adjusted according to the different requirements of the clean area for sampling stability. Usually, the pressure dimension has the highest weight because pressure fluctuations directly affect the stability of the sampling flow rate. Through this formula, the system transforms complex physical quantities into a dimensionless value between 0 and 1. The closer the value is to 1, the healthier the pump is; the lower the value, the greater the potential risk of failure.

[0041] 1.4. Determine the initial operating mode and compare it with the health threshold; The calculated health score is compared with a preset dynamic threshold to determine the system's initial operating mode. A lower health threshold is set. If the main pump health score And greater than the standby pump health score If the system confirms the "main pump priority" mode, the main pump will start working as the sole execution unit, and the standby pump will be in hot standby mode. Conversely, if... or If the system fails to meet the minimum requirements, it will directly enter a dual-pump coordinated or standby priority mode. In this mode, it does not rely on a single pump but dynamically adjusts the output ratio of the two pumps based on their scores. This decision logic ensures that the system can select the optimal configuration at startup, avoiding the risk of sampling failure due to blindly activating the main pump. Simultaneously, this decision result will serve as the input condition for subsequent flow distribution control, guiding the opening and closing actions of the valves.

[0042] Step 2: Adaptive dynamic traffic allocation based on pressure feedback; After establishing the initial operating mode, the system enters the core flow control phase. The goal of this phase is to adjust the output of the dual pumps in real time based on changes in demand at the cleanroom's end, ensuring that the sampled flow rate not only meets the total volume requirements but also maintains extremely high stability. Since there may be slight positive or negative pressure fluctuations within the cleanroom, simply relying on a fixed speed cannot guarantee a constant flow rate; therefore, an adaptive adjustment mechanism based on pressure feedback must be introduced. This step uses the health score calculated in the previous step as the adjustment boundary, combined with the real-time pressure difference, to precisely allocate the speed commands of the main pump and the standby pump through a mathematical model, achieving a seamless and smooth flow transition. Specifically, this includes: 2.1 Establish the mapping relationship between end-point demand flow and inlet pressure; To achieve precise control, it is first necessary to establish the target flow rate at the end of the clean area. With system inlet pressure The dynamic mapping relationship between them. In actual operation, as dust accumulates in the cleanroom filter or the temperature changes, the inlet resistance changes, causing flow rate fluctuations at the same pump speed. The system fits a correction curve using historical operating data, expressed as... ,in This represents the overall pump speed parameter. To simplify real-time calculations, a linear approximation model is used, expressing the target flow rate as a function of inlet pressure plus a linear term for the pump speed: ; in, For the set system back pressure, The pressure sensitivity coefficient, This is the speed gain coefficient. This represents the average equivalent rotational speed of the current dual pumps. This formula reveals the physical law that flow rate is affected by both pressure and rotational speed; when the inlet pressure... When it rises, if you want to maintain If the speed remains unchanged, the rotational speed must be increased accordingly. This step provides a theoretical basis for subsequent flow allocation, enabling the control system to predict the impact of pressure changes on flow.

[0043] 2.2 Calculate the pressure gradient difference and flow compensation requirements of the two pumps; After determining the target flow rate, the system needs to calculate the pressure gradient difference between the main pump and the standby pump to identify which side bears the main load and calculate the flow compensation requirement accordingly. The pressure difference at the outlets of the main pump and the standby pump is defined as... .like A significantly non-zero value indicates a pressure imbalance between the two pumps, which may lead to uneven flow distribution. In this case, the system calculates the required flow compensation. The calculation formula is as follows: ; in, For flow regulation gain coefficient, This represents the effective flow resistance of the connecting pipeline. Based on fluid mechanics principles, this formula shows that the greater the pressure difference, the greater the unexpected flow deviation, requiring correction through a compensation mechanism. The sign of the compensation determines whether to increase the main pump speed, the standby pump speed, or both simultaneously. This is achieved by introducing a flow resistance parameter. The system takes into account the influence of pipeline geometry on fluid flow, making the compensation calculation closer to physical reality and avoiding errors caused by idealized assumptions.

[0044] 2.3 Generate dual-pump speed allocation instructions based on health status; Based on the traffic compensation requirements, combined with the health score obtained in step one. and This generates specific dual-pump speed allocation commands. A non-linear allocation strategy is used here; the pump with the higher health score receives a larger proportion of the speed increment. The target speed increment for the main pump... and the target speed increment of the standby pump The following formulas are given respectively: ; ; in, This is the speed conversion factor, used to convert flow rate units to speed units. This allocation mechanism uses a health score as a weight; when a pump experiences a slight abnormality (a slight decrease in health), the system automatically reduces its speed load, allowing another healthy pump to take on more of the load, thereby extending equipment life and preventing the failure from escalating.

[0045] 2.4 Execute speed commands and monitor instantaneous flow response; After generating the speed command, the system sends a control signal to the drive unit to adjust the output frequency of the inverters for the main pump and the standby pump, thereby changing the actual pump speed. Subsequently, the system immediately enters the monitoring phase, acquiring the new outlet pressure. and and flow meter readings This is to verify the effect of command execution. Instantaneous flow response error. Calculated using the following formula: ; like Less than the preset tolerance range If the error exceeds the acceptable range, the flow allocation is considered successful, and the system continues to maintain the current control strategy. If the error exceeds the acceptable range, a fine-tuning loop is triggered to recalculate the pressure gradient difference and compensation requirements. This closed-loop control process ensures that the system can quickly converge to the target flow state regardless of changes in external disturbances. By monitoring the instantaneous response in real time, the system can promptly detect hysteresis or drift of the actuators, ensuring the continuity of the sampling process.

[0046] Step 3: Redundancy switching trigger condition determination and smooth transition control; After completing dynamic flow allocation, the system enters the critical redundancy management phase. Despite the adoption of an adaptive allocation strategy, a sharp decline in single-pump performance may still occur under extreme conditions. In this case, it is essential to determine whether pump switching is necessary based on strict triggering conditions, ensuring zero interruption of sampled flow during the switching process. The core of this step lies in designing a sensitive and reliable switching logic that avoids both frequent switching due to minor fluctuations causing system oscillations and slow response leading to sampling failure. Specifically, this includes: 3.1 Monitor the sudden change rate of operating parameters of the dual pumps and determine the fault level; First, the system continuously monitors key operating parameters of the main pump and standby pump, especially the pressure change rate. and rate of change of current To detect sudden failures. Define the main pump pressure mutation rate. Standby pump pressure mutation rate When the rate of change of a certain parameter exceeds the preset critical slope At that time, the system determined that the pump might have experienced a sudden failure. Fault level. Based on mutation duration and exceeding the limit The comprehensive evaluation and calculation formula is as follows: ; Where sgn is the sign function, and the integral result reflects the severity of the fault. If Exceeding the critical fault threshold If the fault is within the normal fault threshold, it is classified as a Level 1 fault, and a switching procedure must be initiated immediately; if the fault falls within the normal fault threshold, it is marked as a Level 2 fault, and an early warning observation is initiated. By introducing a time integral term, the system effectively filters out transient noise interference, focusing only on persistent abnormal trends, thus improving the accuracy of fault determination.

[0047] 3.2 Calculate the traffic buffer capacity requirements during the handover process; Once a switchover is determined to be necessary, the system must pre-calculate the required flow buffer capacity during the switchover period to prevent flow vacuum caused by pump shutdown. Buffer capacity Depends on switching time and current target traffic Considering fluid inertia and pipe volume, the actual required buffer space should be slightly larger than the theoretical value. The calculation formula is as follows: ; in, This is the fluid loss coefficient, used to compensate for leakage and turbulence losses during valve closing and opening. The system adjusts based on the compressibility of the liquid in the pipeline. The value of this coefficient is relatively large for gas sampling and relatively small for liquid sampling. By calculating the buffer capacity, the system can plan the start-up timing of the standby pump in advance, ensuring that the new pump has built up sufficient pressure reserve before the old pump completely stops, thus achieving a seamless transition.

[0048] 3.3 Implement valve timing control and pressure wave suppression strategies; After confirming sufficient buffer capacity, the system enters the valve timing control phase. To avoid significant water hammer or pressure surges during switching, valve actions must follow specific timing logic. (Main pump outlet valve) and standby pump outlet valve The operation sequence follows the "open first, close later" principle, meaning the standby pump side valve is opened first, and the main pump side valve is closed only after pressure balance is achieved. Pressure wave suppression strength. This is reflected by adjusting the rate of change of valve opening. The preset pressure difference threshold is set to 15 kPa, and its control law is as follows: ; in, For valve opening, This represents the valve response speed coefficient. The formula indicates that when the pressure difference between the two pumps is large, the valve opening rate should be slowed down to reduce the propagation speed of the pressure wave; when the pressure difference is small, the action can be accelerated to improve efficiency. Through this dynamic adjustment, the system can minimize pressure fluctuations during switching, protecting precision instruments in the clean area from impact.

[0049] 3.4 Verify the switch completion status and lock the new operating mode; The final step is to verify that the switchover was successful and lock the new operating mode. The system checks whether the outlet pressure of the standby pump (now the main pump) has stabilized. Nearby, and traffic Has it been restored to Within the specified range. If the conditions are met, the system will mark the original main pump as offline maintenance status and upgrade the standby pump to online primary status, while simultaneously updating the role identifier in the health score matrix. The switching completion check flag. Defined as: ; in, and These are the allowable deviations for pressure and flow rate, respectively. Only when... Only when the conditions are met will the system officially end the switchover process and enter the next stage of stable operation monitoring. If the conditions are not met, the system will attempt a second switchover or issue an alarm prompting manual intervention.

[0050] Step 4: Multi-node data fusion, transmission, and integrity verification; After completing the redundant switching and flow stabilization control of the dual pumps, the system enters the data acquisition and transmission phase. The core task of this phase is to convert the physical sample information acquired at the cleanroom sampling end into digital signals and transmit them securely and completely to the remote analysis center via the network channel. Because cleanrooms typically have strict electromagnetic shielding requirements or special network architectures, signal attenuation, packet loss, and potential data tampering risks must be overcome during data transmission. This step utilizes the flow and pressure benchmarks established in the previous steps to perform spatiotemporal alignment of the transmitted data, ensuring that each sample data accurately corresponds to the system state at the time of acquisition, thereby providing a reliable basis for subsequent analysis. Specifically, this includes: 4.1 Construct a timestamp-based sample data stream synchronization mechanism; First, the system needs to establish a unified time base to synchronize the sensor readings of the main pump, standby pump, and sampling probe. Each data packet carries a high-precision timestamp. This timestamp is generated by a high-frequency crystal oscillator within the system and periodically calibrated against an external standard time source. To eliminate latency jitter caused by network transmission, a logical arrival time for data packets is defined. With actual physical arrival time Deviation between The corrected timestamp is calculated using the following formula. : ; in, The summation term represents the cumulative network latency statistics, corresponding to the number of historical data packets used for calibration. This synchronization mechanism ensures that when the main pump switches (as described in step 3), the flow rate change curve and the corresponding sample concentration data perfectly overlap on the time axis, avoiding data analysis errors caused by time misalignment. The synchronized data stream becomes the basic unit for subsequent transmission.

[0051] 4.2 Implement segmented encryption and hash integrity verification; Considering the sensitivity and integrity requirements of cleanroom data, the system segments the synchronized data stream before transmission and applies encryption and verification protection. The data stream is divided into several segments of length [missing information]. For each data segment, calculate the cyclic redundancy check code. and hash digest The encryption process uses a symmetric key algorithm to generate ciphertext blocks. Its generating formula can be simplified as follows: ; in, Indicates the encryption function. This is the original data segment. A dynamically generated random key based on the current timestamp. This represents the XOR operation. Simultaneously, it calculates the integrity fingerprint of the entire data packet. : ; In the formula, || represents the string concatenation operation. By introducing dynamic keys and timestamps, even if an attacker intercepts part of the ciphertext, they cannot forge valid data without possessing the real-time key. The integrity fingerprint, as the digital fingerprint of the data packet, will be compared at the receiving end. Any failed comparison will trigger a retransmission mechanism to ensure the absolute authenticity of the transmitted data.

[0052] 4.3 Implement adaptive bandwidth allocation and packet loss retransmission strategies; During network transmission, the communication environment in the clean area may be affected by interference, leading to bandwidth fluctuations or packet loss. The system dynamically adjusts the transmission rate based on the current network conditions and implements an intelligent retransmission strategy. The currently available bandwidth is defined as... The system estimates the retransmission request rate to be... If packet loss rate is detected Exceeding the threshold This automatically reduces the size of transmitted packets and increases the check frequency. The specific adaptive adjustment coefficient... The calculation is as follows: ; in, The equivalent network pipeline resistance coefficient, Here, is the packet loss penalty factor, and e is a natural constant. This formula indicates that when network congestion or severe packet loss occurs, the system will automatically reduce the data transmission rate, prioritizing the delivery of critical control commands (such as switching signals) while temporarily compressing the uploading of non-real-time historical data.

[0053] 4.4 Verify the consistency of data at the remote receiving end and update the status log; The final step is to confirm that the remote analysis center has successfully received and parsed the data. The receiving end decrypts, deduplicates, and verifies the integrity of the received data packets. If the sequence number matches and is unique, it is marked as confirmed. The system records a successful transmission event and updates the transmission counter in the local status log. and error count Successful transmission flag Defined as: ; like The system will initiate a local cache queue, waiting for the next network window to open before retransmitting, until a remote confirmation signal is received. Simultaneously, the metadata of this transmission (including transmission duration, number of lost packets, and encryption strength) will be written into an immutable blockchain-style log for subsequent auditing.

[0054] Step 5: Multi-level fault tolerance and self-healing recovery under abnormal operating conditions; Despite the robust monitoring and switching mechanisms designed for the system, abnormal situations beyond the norm may still occur in the complex cleanroom environment, such as sensor failure, network outages, or simultaneous performance degradation of dual pumps. In such cases, the system must possess multi-level fault tolerance capabilities, capable of identifying fault levels and executing corresponding self-healing procedures. This step builds upon the data transmission in step four, utilizing real-time feedback information to dynamically assess the system's current survivability and adopt differentiated response strategies for different levels of failure. This ensures that even under extreme conditions, the system can maintain minimum functionality or safely shut down, preventing secondary disasters. Specifically, this includes: 5.1 Identify multidimensional anomaly characteristics and classify fault levels; The system first performs correlation analysis on data from all sensors to identify multidimensional anomalies. These features include sensor reading drift, communication link interruptions, and persistent deviations in traffic from target values. Through fuzzy logic, anomalies are categorized into three levels: Level 1 Warning, Level 2 Fault, and Level 3 Catastrophic. Level 1 Warning indicates a minor exceedance of a single parameter, not affecting overall operation; Level 2 Fault indicates persistent anomalies in critical parameters, requiring local isolation; Level 3 Catastrophic indicates loss of core system functions, requiring immediate shutdown for protection. Fault Levels The calculation is based on the weighted anomaly index. : ; Where M is the total number of monitoring parameters, The importance weight of parameter j, These are measured values. This is the nominal value. This is the standard deviation of the parameter. When... Different fault levels correspond to different ranges. For example, if If the risk is high, it is classified as a Level 3 disaster. This quantitative classification method enables the system to objectively assess current risks and avoid overreaction or underreaction.

[0055] 5.2 Perform local isolation and bypass redundancy activation; Once a level 2 or 3 fault is identified, the system immediately implements a local isolation strategy, disconnecting the faulty component from the main control loop to prevent the fault from spreading. For sensor faults, the system automatically switches to the average value or historical trend value of nearby sensors as a replacement; for pump body faults, a bypass redundancy loop is activated. The logic for bypass activation depends on the buffer capacity calculated in step three. To ensure uninterrupted traffic during switching. Bypass traffic. The control formula is: ; Here, Faulted represents the set of components that have failed. The system provides instantaneous traffic to the faulty component. By subtracting the traffic from the faulty component from the total demand, the system calculates the additional load that the remaining healthy components need to handle. If the remaining components cannot handle the full load, the system will gradually reduce the traffic demand of non-critical loads according to a preset priority list, prioritizing the smooth operation of the core sampling channel.

[0056] 5.3 Trigger self-diagnostic testing and component replacement verification; After isolating the faulty component, the system automatically initiates a self-diagnostic test program to perform in-depth testing on suspected faulty components. This includes no-load testing of the standby pump and zero-point calibration checks of sensors. Self-diagnostic results... This is used to determine whether a backup component can be promoted to the primary component. The replacement verification process is completed by comparing the performance metrics of the old and new components. If the health score of the new component is... satisfy If so, component replacement is allowed. The replacement decision logic is as follows: ; If a replacement is decided upon, the system automatically adjusts the routing table and control parameters to incorporate the new component into the control loop; if the decision is made to maintain the status quo, the system continues to operate in degraded mode and increases the monitoring frequency.

[0057] 5.4 Generate a self-healing report and notify the remote management terminal; Regardless of whether the self-healing is successful or not, the system generates a detailed self-healing report, including the time of the fault, the fault type, the measures taken, the final status, and the recommended maintenance plan. The report content is sent to the remote management terminal through an encrypted channel to ensure that administrators can grasp the on-site situation in real time. The integrity verification of the report also follows the hash mechanism in step four. After the report is generated, the system enters observation mode, collecting data at a higher frequency to continuously monitor system stability until it is confirmed that the fault has been completely eliminated or the system has returned to normal. The generation of the self-healing report marks the end of this anomaly handling process and also provides valuable data accumulation for subsequent preventive maintenance.

[0058] Step Six: Full Lifecycle Status Archiving and Predictive Maintenance Planning After completing sampling control, data transmission, and anomaly handling, the system enters the full lifecycle status management and optimization phase. This phase shifts its focus from real-time control actions to in-depth mining and analysis of historical data, aiming to uncover potential patterns, optimize future operating strategies, and plan preventative maintenance. By integrating all operating data, fault records, and switchover logs generated in steps one through five, the system constructs a complete equipment health profile. This not only helps extend equipment lifespan but also allows for proactive risk prediction through trend analysis, enabling a shift from reactive maintenance to proactive prevention. Specifically, this includes: 6.1 Aggregate full-cycle operational data and build health records; The system first aggregates all operational data from a past period (e.g., one month or one quarter). This data includes the dual pump speed curves, pressure fluctuation graphs, switching counts, flow deviation records, and all fault alarm information. Through data cleaning and standardization, a structured health record database is constructed. Key indicators in the record include cumulative operating time. Average health score Fault density The health records are stored in a time-series database format, supporting efficient querying and analysis.

[0059] 6.2. Identify operational trends and predict potential failure risks; Based on the constructed health records, the system uses statistical methods to mine operational trends and predict potential equipment failure risks. By analyzing the slope of the health score H(t) over time, the rate of equipment degradation can be determined. If the health score shows a linear downward trend, the remaining useful life (RUL) can be estimated using the following formula: ; in, For the current health status, This is the critical failure threshold. The rate of health decline. This predictive model does not consider sudden failures, only gradual wear and tear.

[0060] 6.3 Develop dynamic preventative maintenance plans and spare parts scheduling; Based on the forecast results, the system automatically generates a dynamic preventative maintenance plan. The plan includes the suggested maintenance time, a list of parts to be replaced, required tools, and the expected downtime window. The priority of the maintenance plan depends on the predicted failure risk and the urgency of current production tasks. If a pump is predicted to fail within the next 7 days and is currently operating at high load, the system will recommend immediate downtime maintenance; conversely, if it is operating at low load, it can be postponed. The spare parts scheduling strategy is based on inventory levels and predicted demand, calculated using the following formula: ; in, To predict demand, This represents the current inventory level. This serves as a safety stock buffer. Through this formula, the system ensures sufficient supplies of spare parts when needed, while avoiding financial waste caused by excessive inventory. Once the maintenance plan is generated, it is automatically sent to the relevant management departments and simultaneously updated to the equipment management calendar.

[0061] 6.4. Feedback and optimization of control parameters and updates to the system knowledge base; The lessons learned from this full-cycle operation are fed back into the control system, enabling self-evolution. The system fine-tunes control parameters (such as weighting coefficients) based on the deviation between actual operating results and predicted outcomes. and gain coefficient This allows the control strategy to better align with the actual characteristics of the current equipment. Simultaneously, typical failure cases, effective emergency response measures, and optimized parameter combinations are stored in the system knowledge base. The knowledge base is updated using incremental learning, continuously expanding the case library to improve response speed and accuracy when facing similar scenarios in the future.

[0062] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A remote sampling control method for an external pump in a clean area based on dual-pump redundancy switching, characterized in that, include: The real-time multi-dimensional sensor data streams of the main pump and the standby pump are acquired, and after low-pass filtering, the operating deviation value and vibration amplitude normalization factor of each pump relative to the rated operating condition are calculated. Based on the operating deviation value and vibration amplitude normalization factor, a weighted fusion health score model is constructed to calculate the health score of the main pump and the health score of the standby pump. The health scores of the main pump and the standby pump are then compared with a preset dynamic threshold to determine the initial operating mode. Based on the initial operating mode, establish the mapping relationship between the terminal demand flow and the inlet pressure. Combine the health scores of the main pump and the standby pump to calculate the pressure gradient difference and flow compensation requirement of the two pumps. Generate and execute the health-based dual pump speed allocation command to monitor the instantaneous flow response. Based on the instantaneous flow response, monitor the mutation rate of dual pump operating parameters and determine the fault level. Calculate the flow buffer capacity requirement during the switching process based on the determination result, and execute valve timing control and pressure wave suppression strategies to verify the switching completion status and lock the new operating mode. In the new operating mode, a timestamp-based sample data stream synchronization mechanism is constructed, segmented encryption and hash integrity verification are implemented, and adaptive bandwidth allocation and packet loss retransmission strategies are executed to verify the data consistency of the remote receiver and update the status log. Based on the status log records, multidimensional abnormal features are identified and fault levels are classified. Local isolation and bypass redundancy activation are performed, self-diagnostic tests and component replacement verification are triggered to generate self-healing reports, full-cycle operation data is aggregated to build health profiles, operation trends are explored to predict potential failure risks, dynamic preventive maintenance plans are formulated, and control parameters are optimized.

2. The method for remote sampling control of external pumps in clean areas based on dual-pump redundancy switching according to claim 1, characterized in that, The acquisition of real-time multi-dimensional sensor data streams from the main pump and the standby pump includes: Raw data streams are synchronously acquired from a multi-source sensor array deployed on the main pump and the standby pump. The raw data streams include the main pump outlet pressure, the standby pump outlet pressure, the main pump motor phase current, the standby pump motor phase current, the main pump bearing vibration acceleration, and the standby pump bearing vibration acceleration. The acquired raw voltage signal is smoothed by using the sliding window length and sampling time interval to eliminate environmental noise interference and obtain the filtered sensor reading.

3. The method for remote sampling control of external pumps in clean areas based on dual-pump redundancy switching according to claim 1, characterized in that, The calculation of the operating deviation value and vibration amplitude normalization factor of each pump relative to the rated operating condition includes: The pressure deviation is obtained by subtracting the rated reference pressure from the measured main pump outlet pressure and the standby pump outlet pressure. The current deviation is obtained by subtracting the rated reference current from the measured main pump motor phase current and the standby pump motor phase current. The vibration amplitude normalization factor is obtained by dividing the measured vibration acceleration of the main pump bearing and the standby pump bearing by the upper limit of the safety threshold. The pressure deviation, current deviation, and vibration amplitude normalization factor directly reflect whether the pump body is overloaded, blocked, or mechanically worn.

4. The method for remote sampling control of external pumps in clean areas based on dual-pump redundancy switching according to claim 1, characterized in that, The construction of the weighted fusion health score model includes: Pressure stability, current stability, and vibration level are taken as three core dimensions, and weights are set for pressure dimension, current dimension, and vibration dimension, and the sum of the three is equal to 1. Using the pressure deviation, current deviation, vibration amplitude normalization factor, and weighting coefficient, the health score of the main pump and the health score of the standby pump are calculated respectively. The calculated health score is converted into a dimensionless value between 0 and 1. The closer the value is to 1, the healthier the pump is.

5. The method for remote sampling control of external pumps in clean areas based on dual-pump redundancy switching according to claim 1, characterized in that, The determination of the initial operating mode includes: If the health score of the main pump is greater than or equal to the preset lower health threshold and greater than the health score of the standby pump, then the main pump priority mode is confirmed. If the health score of the main pump is less than the preset lower health threshold or the health score of the standby pump is greater than the health score of the main pump, then the dual-pump collaborative mode or standby priority mode will be entered. The judgment results guide the opening and closing actions of the valves and establish the initial operating benchmark of the system.

6. The method for remote sampling control of external pumps in clean areas based on dual-pump redundancy switching according to claim 1, characterized in that, The traffic buffer capacity requirements during the calculation switching process include: The required buffer capacity is calculated based on the switching time and the current target flow rate, combined with the fluid loss coefficient; the value of the fluid loss coefficient is adjusted according to the compressibility of the liquid in the pipeline, and the start-up timing of the standby pump is planned in advance.

7. The method for remote sampling control of external pumps in clean areas based on dual-pump redundancy switching according to claim 1, characterized in that, The valve timing control and pressure wave suppression strategy includes: Following the principle of opening before closing, first open the valve on the standby pump side, and then close the valve on the main pump side after the pressure is balanced. Calculate the pressure difference between the main pump outlet pressure and the standby pump outlet pressure in real time; When the pressure difference exceeds a preset pressure difference threshold, the valve opening change rate is reduced to slow down the propagation speed of the pressure wave. When the pressure difference does not exceed the preset pressure difference threshold, the valve opening change rate is increased to improve switching efficiency.

8. The method for remote sampling control of external pumps in clean areas based on dual-pump redundancy switching according to claim 1, characterized in that, The identification of multidimensional abnormal features and the classification of fault levels include: Continuously monitor the pressure and current change rates of the main pump and standby pump to detect sudden failures; The integral result of the pressure mutation rate is calculated, and when the integral result exceeds the severe fault threshold, it is judged as a level one fault; When the integral result is between the general fault threshold, it is marked as a level 2 fault and an early warning observation is initiated.

9. The method for remote sampling control of external pumps in clean areas based on dual-pump redundancy switching according to claim 1, characterized in that, The predicted failure risks of the excavation operation trend include: The rate of equipment degradation is determined by analyzing the slope of the health score over time. Estimate remaining lifespan by dividing the current health status and critical failure threshold by the rate of health decline. Analyze the periodicity of failure occurrences to identify whether failures are more likely to occur under certain temperature or humidity conditions.

10. A remote sampling control system for an external cleanroom pump based on dual-pump redundancy switching for implementing the method as described in any one of claims 1-9, characterized in that, include: The data acquisition unit is used to acquire real-time multi-dimensional sensor data streams from the main pump and the standby pump. After low-pass filtering, it calculates the operating deviation value of each pump relative to the rated operating condition and the vibration amplitude normalization factor. The status assessment unit is used to construct a weighted fusion health score model based on the operating deviation value and vibration amplitude normalization factor, calculate the main pump health score and the standby pump health score, and compare the main pump health score and the standby pump health score with a preset dynamic threshold to determine the initial operating mode. The flow control unit is used to establish a mapping relationship between the terminal demand flow and the inlet pressure according to the initial operating mode, calculate the pressure gradient difference between the two pumps and the flow compensation requirement by combining the health score of the main pump and the health score of the standby pump, and generate and execute the health-based dual pump speed allocation command to monitor the instantaneous flow response. The redundant switching unit is used to monitor the mutation rate of the dual pump operating parameters and determine the fault level based on the instantaneous flow response, calculate the flow buffer capacity requirement during the switching process based on the determination result, and execute valve timing control and pressure wave suppression strategies to verify the switching completion status and lock the new operating mode. The data transmission unit is used to build a timestamp-based sample data stream synchronization mechanism in the new operating mode, implement segmented encryption and hash integrity verification, and execute adaptive bandwidth allocation and packet loss retransmission strategies to verify the data consistency of the remote receiving end and update the status log. The fault-tolerant recovery unit is used to identify multi-dimensional abnormal features and classify fault levels based on the status log records, perform local isolation and bypass redundancy activation, and trigger self-diagnostic tests and component replacement verification to generate a self-healing report. The record management unit is used to aggregate full-cycle operation data to build health records, identify operational trends to predict potential failure risks, formulate dynamic preventive maintenance plans, and provide feedback to optimize control parameters.

Images