A remote fault diagnosis and early warning method for a chromatographic analyzer

Abstract

This invention relates to the field of fault early warning technology, and more specifically, to a remote fault diagnosis and early warning method for a chromatograph, comprising: acquiring operating data of the ion chromatograph, the operating data including a real-time pressure sequence of the high-pressure pump, a real-time torque signal sequence of the pump drive motor, and a baseline data sequence of the conductivity detector; and calculating a pressure energy fluctuation index in the overlapping area within each pumping cycle based on the pressure sequence and torque signal sequence, wherein the pressure energy fluctuation index is used to characterize the degree of fluid kinetic energy loss caused by a faulty one-way valve seal. This invention enables causal correlation analysis of the chromatograph's mechanical abnormality leading to a decrease in detection signal, provides early warning of one-way valve failures and offers tiered maintenance suggestions, significantly improving the intelligent remote maintenance level of precision analytical instruments.

Description

Technical Field

[0001] This invention relates to the field of fault early warning technology. More specifically, this invention relates to a remote fault diagnosis and early warning method for a chromatographic analyzer. Background Technology

[0002] Ion chromatography, as a sophisticated analytical instrument, plays a crucial role in water quality monitoring, chemical production, and food safety. The high-pressure pump, as the power source for the chromatograph, directly determines the accuracy and stability of the analytical results. In existing ion chromatography techniques, to achieve constant flow delivery and eliminate flow pulsation under high back pressure, high-pressure pumps generally employ a tandem dual-plunger structure and are designed with a dedicated "cross-overlap zone"—the phase interval between the main pump head and the compensation pump head simultaneously pushing the eluent—for pre-compression of the eluent.

[0003] In practical laboratory or industrial applications, ion chromatographs often need to operate continuously for extended periods. However, with increased operating time, contamination from minute particles in the eluent or mechanical wear of the check valve can easily lead to a failure to seal properly. In the early stages of the malfunction, this lack of sealing will only cause a small amount of liquid backflow, manifesting as a momentary collapse of the pressure curve at a specific phase.

[0004] Currently, the most significant technical challenge in this scenario is that traditional equipment condition monitoring methods struggle to sensitively detect such minute and short-lived pressure distortions. This often masks early, minor faults in check valves within normal system fluctuations. Simultaneously, random variations in the laboratory environment (such as temperature fluctuations caused by air conditioning and power supply instability) can cause baseline drift in conductivity detectors. Diagnostic systems frequently fail to accurately distinguish between environmental interference and actual mechanical component failures, leading to frequent false alarms or missed alarms. This not only makes it difficult for maintenance personnel to determine the optimal time for repairs but can also result in the failure to promptly identify potential problems, leading to damage to expensive chromatographic columns, sample spoilage, or even complete interruption of the experimental process. Summary of the Invention

[0005] This invention provides a remote fault diagnosis and early warning method for a chromatograph, aiming to solve the problem that traditional equipment condition monitoring methods in related technologies are unable to sensitively capture pressure distortions with small amplitudes and extremely short durations, causing early minor faults of check valves to be often masked by normal system fluctuations.

[0006] This invention provides a remote fault diagnosis and early warning method for a chromatograph, comprising: acquiring the operating data of the ion chromatograph, the operating data including the real-time pressure sequence of the high-pressure pump, the real-time torque signal sequence of the pump drive motor, and the baseline data sequence of the conductivity detector; calculating the pressure energy fluctuation index of the overlapping region in each pumping cycle based on the pressure sequence and torque signal sequence, the pressure energy fluctuation index being used to characterize the degree of fluid kinetic energy loss caused by poor sealing of the one-way valve; wherein, the overlapping region refers to the phase interval in which the main pump head and the compensation pump head are simultaneously in the liquid pushing stroke within a complete pumping cycle; improving the cumulative distance calculation function of DTW using the pressure energy fluctuation index, and calculating the physical weighted DTW final distance between the real-time pressure sequence and the healthy standard pressure template sequence, the physical weighted DTW final distance reflecting the overall difference between the currently acquired real-time pressure sequence and the healthy standard pressure template sequence; combining the physical weighted DTW final distance, the average DTW distance during historical normal operation, and the standard deviation of the baseline data sequence in the current pumping cycle to obtain the comprehensive fault probability value of the current ion chromatograph, and generating corresponding remote early warning information based on the magnitude of the comprehensive fault probability value. By integrating multi-source data such as high-pressure pump pressure, torque, and detector baseline, and introducing a pressure energy fluctuation index in the overlapping region of the pumping cycle to improve the DTW algorithm, this method can sensitively detect minute backflow and instantaneous pressure collapse caused by poor sealing of the one-way valve. In practical applications, this method can not only detect early minor faults in the one-way valve, but also effectively distinguish between actual mechanical wear and environmental fluctuation interference by the cross-correlation characteristics of mechanical state and signal anomalies. This significantly reduces the false alarm and missed detection rates, providing researchers with accurate references for maintenance timing and preventing column damage or sample waste.

[0007] Furthermore, the pressure energy fluctuation index of the overlapping area within each pumping cycle is calculated, including the positive correlation between the pressure energy fluctuation index and the absolute value of the derivative of the pressure with respect to the motor rotation angle within the overlapping area and the product of the torque at the corresponding sampling point within the overlapping area. By establishing the coupling relationship between the pressure energy fluctuation index, the pressure derivative, and the motor torque, microscopic mechanical wear is successfully quantified into a measurable physical index. During pumping, when a minor leak occurs, the algorithm can compensate for the pressure loss through the physical logic of drastic changes in motor torque, amplifying the expressiveness of fault characteristics and making weak signals that were originally masked in a single pressure sequence clearly identifiable in the energy dimension.

[0008] Furthermore, the method for obtaining corresponding sampling points within the overlapping area includes: for any pumping cycle, logically filtering the real-time acquired data stream based on a preset cam mechanical angle range, selecting sampling points where the real-time rotation angle fed back by the motor enters the aforementioned angle range, all of which are taken as corresponding sampling points within the overlapping area. Using the cam mechanical angle range for logical filtering can accurately pinpoint the sensitive phase interval where the main pump head and the compensation pump head simultaneously push liquid, i.e., the overlapping area. This ensures that the data source for fault diagnosis comes from the window most sensitive to the sealing performance of the one-way valve, eliminating invalid data interference from non-critical phase segments, and improving the targeting and computational efficiency of remote diagnosis.

[0009] Furthermore, the cumulative distance calculation function of DTW is improved using the pressure energy fluctuation index, including: introducing the pressure energy fluctuation index into the distance matching formula, wherein the cumulative distance and the pressure energy fluctuation index are positively correlated. Introducing the pressure energy fluctuation index into the cumulative distance calculation of DTW allows the algorithm to dynamically adjust the local path penalty weights during the matching process. In application, when a sequence contains key fault information, the algorithm automatically increases its matching penalty, thereby significantly widening the gap between the fault state and the healthy state in the final distance, greatly improving the algorithm's ability to identify early potential problems.

[0010] Furthermore, the physical weighted DTW final distance is calculated as follows: for the physical weighted DTW final distance of the real-time pressure sequence and the health standard pressure template sequence, the cumulative distance is iteratively calculated until the end of both sequences. The final value of the last element of the matrix is ​​then taken as the physical weighted DTW final distance characterizing the overall difference between the two sequences. By iteratively calculating until the end of the sequence and taking the last element value, a globally optimal solution reflecting the overall difference between the real-time operating state and the health benchmark is provided. This allows the diagnostic system to macroscopically assess the overall degradation trend of the pump system, providing a reliable quantitative indicator for the long-term stability assessment of the instrument.

[0011] Furthermore, the method for obtaining the health standard pressure template sequence is as follows: With the instrument in brand-new condition or in a calibrated and fault-free operating state, real-time pressure sequence data within its standard pumping cycle is collected in its entirety, serving as the health standard pressure template sequence. By collecting real-time data from the instrument in its brand-new or calibrated fault-free state as a template, a personalized health record is established for each instrument. Since the mechanical tolerances and operating backgrounds of each pump are slightly different, this dynamic acquisition method ensures that subsequent fault diagnosis is based on a comparison with the instrument's optimal condition, eliminating the impact of individual differences between instruments on diagnostic accuracy.

[0012] Furthermore, the formula for calculating the cumulative distance is: In the formula, Indicates the pressure sequence before Each sampling point and the health standard pressure template sequence before The cumulative distance of each sampling point; The first in the real-time acquired pressure sequence One sampling point; The first in the health standard stress template sequence One sampling point; In order to be with the first The pressure energy fluctuation index corresponding to each sampling point. This is determined through a mathematical formula... The product term logic effectively amplifies minute physical distortions. In actual monitoring, even if the change in absolute pressure difference is not significant, as long as the energy fluctuation index increases, the physically weighted DTW distance will deviate significantly from the normal range, ensuring that the system can issue a warning at the very early stage of one-way valve contamination.

[0013] Furthermore, the first The method for determining the pressure energy fluctuation index corresponding to each sampling point includes: firstly, identifying the... The pumping cycle to which each sampling point belongs and its position within the cycle, if the first sampling point... The sampling point is located within the overlapping area of ​​this pumping cycle. The pressure energy fluctuation index corresponding to each sampling point is the pressure energy fluctuation index for that pumping cycle; if the first sampling point... If the sampling point is located in the non-overlapping region, then the... The pressure energy fluctuation index corresponding to each sampling point is 0. By identifying the period and location of the sampling point, the E value is determined, enabling differentiated processing of overlapping and non-overlapping areas. This method allows the diagnostic logic to focus on the critical timing sequence where faults frequently occur, while maintaining zero weight in non-overlapping areas. This protects key features from being smoothed out and effectively suppresses noise interference from normal pumping fluctuations.

[0014] Furthermore, when the overall failure probability value is greater than or equal to the upper limit of a set interval, it is determined to be a serious failure, and an early warning is sent to the remote control system. The set interval ranges from [value missing]. By setting specific upper limits for probability intervals, quantitative calculation results are transformed into qualitative diagnostic strategies. In laboratory management scenarios, this provides clear decision support for maintenance personnel, such as prompting the cleaning of check valves in sub-healthy states and automatically shutting down pumps for protection in the event of serious faults, thereby achieving an intelligent closed-loop operation and maintenance system from fault detection to automatic handling.

[0015] Furthermore, after acquiring the operating data of the ion chromatograph, the process also includes smoothing the acquired real-time pressure sequence, real-time torque signal sequence, and baseline data sequence using a median filtering algorithm. Pre-processing the raw sensor signals using the median filtering algorithm effectively filters out high-frequency electromagnetic noise commonly found in laboratory environments. This provides better consistent and higher signal-to-noise ratio baseline data for subsequent feature extraction and DTW matching, avoiding algorithmic misjudgments caused by spike noise and ensuring the robustness of the fault warning system.

[0016] Beneficial effects: Addressing the pain points of difficulty in detecting early faults in one-way valves and the susceptibility to false alarms due to environmental interference, this method firstly constructs the most mechanically sensitive window—the high-pressure pump cross-overlap area—and establishes a pressure energy fluctuation index by coupling the pressure change rate with motor torque, transforming microscopic mechanical wear into significant energy characteristics. Secondly, this index improves the Dynamic Time Warping (DTW) algorithm, enabling physically weighted sequence alignment and effectively amplifying minute physical distortions. Finally, by combining the baseline noise of the conductivity detector to establish a comprehensive fault probability value, this method achieves causal correlation analysis of the chromatograph's transition from abnormal mechanical condition to decreased detection signal, enabling early warning of one-way valve faults and providing tiered maintenance recommendations, significantly improving the intelligent remote maintenance level of precision analytical instruments. Attached Figure Description

[0017] Figure 1 This is a schematic flowchart illustrating an early warning method according to an embodiment of the present invention;

[0018] Figure 2 This is a schematic diagram illustrating the comparison between the physical weighted DTW cumulative distance and the conventional DTW cumulative distance as a function of sampling points, according to an embodiment of the present invention. Detailed Implementation

[0019] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0020] like Figure 1 As shown, S101: Data acquisition and preprocessing.

[0021] In this embodiment, the method first acquires the real-time operating data streams of multiple sensors inside the ion chromatograph via a data interface. These data form the basis for diagnostic analysis. Specifically, the acquired data includes: the real-time pressure sequence of the high-pressure pump; the real-time torque signal sequence of the pump drive motor, which is directly fed back by the pump driver and reflects the load required for the motor to maintain constant current output; and the baseline data sequence of the conductivity detector, characterizing the background conductivity value of the eluent as it passes through the detection cell.

[0022] It is understandable that the original sensor signals inevitably contain high-frequency electromagnetic noise. To improve the signal-to-noise ratio, this embodiment uses a median filtering algorithm to smooth the acquired real-time pressure sequence, real-time torque signal sequence, and baseline data sequence. Subsequently, the preprocessed real-time pressure sequence is precisely segmented based on the high-pressure pump's working cycle, ensuring that each data segment strictly corresponds to a complete pumping cycle, i.e., containing one complete suction and discharge stroke. Specifically, the sequence segmentation in this step can be achieved as follows: When no external hardware signal is involved, since the high-pressure pump generates inherent pressure changes at the moment of alternating suction and discharge, the system uses a peak detection algorithm or a zero-crossing detection algorithm to analyze the time-domain waveform of the real-time pressure sequence itself, identifying the most significant trough point (such as the lowest pressure point generated during the suction stroke) or peak point within each pumping cycle, and using the sequence between two adjacent characteristic trough points or peak points as a complete pumping cycle.

[0023] S102: Construct the pressure energy fluctuation index.

[0024] In the tandem dual-plunger pump flow path system used in ion chromatography, to overcome the physical compressibility of the eluent under high back pressure and eliminate flow pulsation, the mechanical profile of the pump body cam is designed with a specific motion sequence. Specifically, in each complete pumping cycle, when the main pump head completes liquid aspiration and prepares to discharge liquid to the high-pressure system, the liquid in the pump chamber must first be pre-compressed to raise its pressure from atmospheric pressure to the system operating pressure. During this pre-compression period, to maintain a constant total system flow rate, the compensating pump head must continue to push liquid. Therefore, mechanically, there must exist a phase interval where the main pump head and the compensating pump head are simultaneously in the pushing stroke, i.e., the aforementioned cross-over overlap region. This region appears consistently in each pumping cycle and is a key guarantee for the system to achieve stable, pulsation-free liquid delivery. It is also the most sensitive window reflecting the sealing performance of the one-way valve and the accuracy of the mechanical transmission. According to fluid dynamics theory, when a pump's check valve fails to seal properly due to microparticle contamination, a small amount of eluent backflow will occur in the overlapping area, causing a unique, instantaneous collapse distortion in the pressure curve. This distortion is altered by traditional Euclidean distance or... The algorithm is not sensitive to physical distortions with small amplitude but drastic slope changes.

[0025] It should be noted that the data set of the overlapping area in each pumping cycle is determined as follows: In this embodiment, the total number of sampling points in the overlapping area is not a fixed constant, but is determined by the data acquisition frequency and the current rotational speed of the motor. Specifically, the system uses a preset cam mechanical angle range (e.g., Logical filtering is performed on the real-time acquired data stream: within a complete pumping cycle, all sampling points whose real-time rotation angle fed back by the motor falls within the above-mentioned angle range are included in the cross-overlap area dataset.

[0026] Therefore, this embodiment constructs a pressure energy fluctuation index. Its construction is based on quantifying the physical distortion caused by the poor sealing of the one-way valve, coupling the mechanical motion state of the pump (reflected by the motor torque) with the fluid pressure state, thereby transforming microscopic mechanical wear into measurable energy fluctuation characteristics. The calculation formula is as follows: In the formula: The pressure energy fluctuation index represents the area of ​​overlap between pump heads during the pumping cycle. Indicates the first At each sampling point, pressure For motor rotation angle The instantaneous derivative of reflects the rate of change of pressure per unit mechanical stroke; Indicates the first Real-time torque signal at each sampling point This represents the total number of sampling points contained in the overlapping area dataset.

[0027] As can be seen from this formula, in the event of a one-way valve leak, the pressure drops sharply due to backflow, leading to... The absolute value of the pressure loss increases abnormally; simultaneously, to compensate for this pressure loss, the motor needs to instantaneously increase its output torque, leading to... A dramatic jump also occurred. The combined effect of these two factors led to… The value increased significantly. Therefore, It can sensitively detect the physical process of fluid kinetic energy loss caused by pump mechanical wear.

[0028] S103: Construct the physical weighted DTW cumulative distance.

[0029] Traditional DTW algorithms use fixed weights for all data points when calculating the similarity between real-time acquired pressure sequences and health standard pressure template sequences. This makes it difficult for the algorithm to distinguish between ordinary noise and specific waveform distortions caused by mechanical wear. Specifically, real-time pressure sequence data within a complete standard pumping cycle is acquired when the instrument is in brand-new condition or has been verified to be fault-free, and this data serves as the health standard pressure template sequence.

[0030] This embodiment is... The core of the algorithm—the cumulative distance calculation function—is improved. Its construction is based on dynamically adjusting the local path penalty weight in the DTW matching process using the pressure energy fluctuation index extracted in step S102. In other words, if a certain segment of the pressure sequence... A higher value indicates that the sequence contains more critical mechanical fault information, and the algorithm should automatically increase the path penalty for matching to that sequence. The improved cumulative distance function is constructed. as follows: In the formula: Indicates the pressure sequence before Each sampling point and the health standard pressure template sequence before The cumulative distance of each sampling point; The first in the real-time acquired pressure sequence One sampling point; The first in the health standard stress template sequence One sampling point; In order to be with the first The pressure energy fluctuation index corresponding to each sampling point.

[0031] It should be noted that, compared with the first The method for obtaining the pressure energy fluctuation weight corresponding to the sampling point is as follows: First, identify the first... The pumping cycle to which each sampling point belongs and its position within the cycle; if the first sampling point... If each sampling point is located within the overlapping region of that period, then The value is taken as the pressure energy fluctuation index calculated for this period in step S102; if the first If each sampling point is located in a non-overlapping region, then The value is 0.

[0032] This formula shows that when the instrument experiences slight leakage or the one-way valve becomes contaminated, although the absolute pressure... Compared with standard value The difference The changes may not be significant, but It will increase significantly. (Through the product term) The logic is that this tiny physical distortion is effectively amplified, thus making the final distance of the physically weighted DTW increase. It can deviate significantly from the normal range, enabling the algorithm to detect anomalies in the early stages of a fault.

[0033] S104: Construct the overall failure probability value of the current ion chromatograph.

[0034] The ultimate performance characteristic of a chromatograph is the stability of its conductivity baseline. Diagnostic logic cannot solely rely on the mechanical condition of the pump; a causal relationship must be established between abnormal mechanical conditions and decreased detection performance. If... The increase in distance is indeed caused by the pressure pulsation of the pump, which will inevitably lead to noise at the synchronization frequency in the conductivity detector.

[0035] This embodiment constructs a comprehensive fault probability value. Its construction is based on the fact that by constructing the cross-correlation characteristics between mechanical anomalies and signal anomalies, false alarms due to baseline drift caused by fluctuations in the laboratory environment, such as temperature and power instability, which are unrelated to the pump, can be effectively eliminated. The calculation formula is as follows: In the formula: This represents the overall failure probability value of the current ion chromatograph; The final physical weighted DTW distance between the currently acquired real-time pressure sequence and the health standard pressure template sequence; The average of the final physical weighted DTW distance over historical normal operating cycles serves as a healthy benchmark; To determine the baseline data sequence output by the conductivity detector within the same pumping cycle time window. The standard deviation represents the level of baseline noise; The baseline data sequence within the same pumping cycle time window. The mean; For example, a very small constant Its function is to prevent the denominator from being zero and to ensure the stability of the calculation.

[0036] The aforementioned physical weighted DTW final distance This is the globally optimal solution obtained by recursively traversing the cumulative distance function constructed in step S103. Specifically, if the total length of the currently collected pressure sequence is N and the total length of the health standard pressure template sequence is M, the algorithm iteratively calculates the cumulative distance using the above formula. Until the end of both sequences, finally take the value of the last element of the matrix, that is, let The physical weighted DTW final distance serves as a representation of the overall degree of difference between these two complete sequences.

[0037] Due to the introduction of the hyperbolic tangent function This indicator exhibits good saturation characteristics. When mechanical fluctuations and electrical conductance noise increase simultaneously, the indicator rapidly approaches its maximum value. This indicates a high probability of a real fault occurring. If there is only baseline drift and the final physically weighted DTW distance is normal, the system will determine it as environmental interference rather than an instrument hardware failure, thus significantly improving diagnostic accuracy.

[0038] S105: Perform remote early warning based on the magnitude of the comprehensive fault probability value.

[0039] In this embodiment, the system calculates the comprehensive failure probability value based on step S104. The system classifies the health status of instruments and automatically generates corresponding remote early warning and maintenance strategies.

[0040] like The system determines that the consumable is in a sub-healthy state. At this time, it will automatically push a level-two warning message to the remote client, suggesting cleaning the one-way valve and attaching a link to relevant operation instructions. The system has identified a serious malfunction, which may have resulted in issues such as a broken plunger rod or complete failure of the sealing ring. To protect the core components of the instrument, the system will immediately issue a protection command to stop the high-pressure pump via remote control and automatically generate a maintenance task order containing the fault diagnostic code and the required spare parts number, which will then be sent to the maintenance engineer's management system.

[0041] like Figure 2 As shown in the figure, the cumulative distance calculated by the two algorithms varies with the sampling point during one pumping cycle of the ion chromatography pump. The overlapping area is specifically marked in the figure. It can be seen that under the method of this invention, due to the introduction of the pressure energy fluctuation index as a penalty weight, the slope of the cumulative distance in this area is significantly increased, resulting in a significant improvement in the final physically weighted DTW distance compared to the traditional method, thus enabling more sensitive detection of minute pressure distortions.

[0042] The embodiments described above are merely examples of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention.

Claims

1. A remote fault diagnosis and early warning method for a chromatographic analyzer, characterized in that, include: Acquire the operating data of the ion chromatograph, including the real-time pressure sequence of the high-pressure pump, the real-time torque signal sequence of the pump drive motor, and the baseline data sequence of the conductivity detector; Based on the pressure sequence and torque signal sequence, the pressure energy fluctuation index of the overlapping area in each pumping cycle is calculated. The pressure energy fluctuation index is positively correlated with the absolute value of the derivative of the pressure with respect to the motor rotation angle in the overlapping area and the product of the torque at the corresponding sampling point in the overlapping area. The pressure energy fluctuation index is used to characterize the degree of fluid kinetic energy loss caused by poor sealing of a check valve; The overlapping area refers to the phase interval in which the main pump head and the compensation pump head are simultaneously in the liquid pushing stroke within a complete pumping cycle. The cumulative distance calculation function of DTW is improved by utilizing the pressure energy fluctuation index. The pressure energy fluctuation index is introduced into the distance matching formula, and the cumulative distance and the pressure energy fluctuation index are positively correlated. The cumulative distance is: ； Indicates the pressure sequence before Each sampling point and the health standard pressure template sequence before The cumulative distance of each sampling point; The first in the real-time acquired pressure sequence One sampling point; The first in the health standard stress template sequence One sampling point; In order to be with the first The pressure energy fluctuation index corresponding to each sampling point; No. The method for determining the pressure energy fluctuation index corresponding to each sampling point includes: First, identify the first The pumping cycle to which each sampling point belongs and its position within the cycle, if the first sampling point... The sampling point is located within the overlapping area of ​​this pumping cycle. The pressure energy fluctuation index corresponding to each sampling point is the pressure energy fluctuation index for that pumping cycle; if the first sampling point... If the sampling point is located in the non-overlapping region, then the... The pressure energy fluctuation index corresponding to each sampling point is 0; The physical weighted DTW final distance between the real-time stress sequence and the healthy standard stress template sequence was calculated. The physical weighted DTW final distance reflects the overall degree of difference between the currently acquired real-time stress sequence and the healthy standard stress template sequence. By combining the physical weighted DTW final distance, the average DTW distance during historical normal operation, and the standard deviation of the baseline data sequence within the current pumping cycle, the comprehensive failure probability value of the current ion chromatograph is obtained, and corresponding remote early warning information is generated based on the magnitude of the comprehensive failure probability value.

2. The remote fault diagnosis and early warning method for a chromatographic analyzer according to claim 1, characterized in that, The method for obtaining corresponding sampling points within the overlapping region includes: For any pumping cycle, the real-time data stream is logically filtered according to the preset cam mechanical angle range, and the sampling points where the real-time rotation angle fed back by the motor enters the above angle range are selected as the corresponding sampling points in the cross-overlap area.

3. The remote fault diagnosis and early warning method for a chromatographic analyzer according to claim 1, characterized in that, The method for calculating the final distance of the Physically Weighted DTW is as follows: For the physical weighted DTW final distance between the real-time stress sequence and the health standard stress template sequence, the cumulative distance is calculated iteratively until the end of the two sequences. Finally, the value of the last element of the matrix is ​​taken as the physical weighted DTW final distance that characterizes the overall difference between the two sequences.

4. The remote fault diagnosis and early warning method for a chromatograph according to claim 1, characterized in that, The method for obtaining the health standard stress template sequence is as follows: When the instrument is in brand new condition or has been verified to be fault-free, real-time pressure sequence data within its standard pumping cycle is collected in its entirety as a health standard pressure template sequence.

5. The remote fault diagnosis and early warning method for a chromatograph according to claim 1, characterized in that, When the overall failure probability value is greater than or equal to the upper limit of the set interval, it is determined to be a serious failure, and an early warning is sent to the remote control system. The set interval ranges from [value missing]. .

6. The remote fault diagnosis and early warning method for a chromatographic analyzer according to claim 1, characterized in that, After acquiring the operating data of the ion chromatograph, the process also includes smoothing the acquired real-time pressure sequence, real-time torque signal sequence, and baseline data sequence using a median filtering algorithm.

Images