A hydrogen purification adsorption tower pressure difference early warning method and system based on working condition self-adaptation

Abstract

The application discloses a hydrogen purification adsorption tower pressure difference early warning method and system based on working condition self-adaptation, belongs to the hydrogen purification technical field, and comprises the following steps: according to the pressure equalization stage identifier issued by the adsorption tower control system, curve interception is carried out on the pressure difference time sequence curve, a first curve pair is obtained, and each curve in the first curve pair is marked with a tower role identifier; according to the tower role identifier and the pressure equalization stage identifier, a single-tower reference curve is called, and a second curve pair is obtained; the residual error of the first curve pair and the second curve pair is extracted at each moment, and a residual error curve pair is obtained; the direction of the residual error curve pair is compared, an abnormal source discrimination result is obtained, and a pressure difference early warning result is generated according to the abnormal source discrimination result. The application realizes accurate positioning of the abnormal source of the pressure difference, and improves the accuracy and availability of the pressure difference early warning result.

Description

Technical Field

[0001] This invention relates to the field of hydrogen purification technology, specifically to a method and system for early warning of differential pressure in a hydrogen purification adsorption tower based on adaptive operating conditions. Background Technology

[0002] In the pressure swing adsorption (PSA) process of a hydrogen purification unit, there is a pressure coupling relationship between the two adsorption towers in the pressure equalization stage. Fluctuations in operating conditions such as inlet flow rate and gas composition will be synchronously transmitted to the two adsorption towers through this coupling relationship, causing the pressure difference curves of the two to shift synchronously. This shift is a normal linkage change caused by fluctuations in operating conditions, rather than a failure of the adsorption towers themselves.

[0003] Existing adsorption tower differential pressure early warning methods typically use fixed thresholds or adaptive thresholds based on historical data statistics. They only judge whether the differential pressure amplitude exceeds the threshold range, failing to distinguish whether the differential pressure deviation originates from the transmission of operating condition fluctuations or from the abnormal state of a single tower itself, and also failing to determine which adsorption tower the abnormal differential pressure specifically originates from.

[0004] For example, when the intake gas composition fluctuates, the pressure difference between the two adsorption towers coupled with the pressure equalization tower will shift synchronously. Existing methods only determine that the two adsorption towers are abnormal at the same time and trigger an early warning based on the pressure difference amplitude exceeding the limit. This leads to maintenance personnel performing unnecessary shutdowns to troubleshoot adsorption towers that have not actually failed, increasing maintenance costs and reducing the reliability of the pressure difference early warning results. Summary of the Invention

[0005] In view of the above-mentioned problems, the present invention provides a method and system for early warning of differential pressure in hydrogen purification adsorption tower based on operating condition adaptation.

[0006] Therefore, the technical problem solved by the present invention is that the existing adsorption tower differential pressure early warning method cannot distinguish between the synchronous differential pressure deviation of two towers caused by fluctuations in operating conditions and the differential pressure deviation caused by the abnormality of a single tower, nor can it determine the specific tower from which the differential pressure abnormality originates, resulting in low accuracy and poor usability of differential pressure early warning results.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a differential pressure early warning method for hydrogen purification adsorption tower based on adaptive operating conditions, comprising the following steps: according to the pressure equalization stage identifier issued by the adsorption tower control system, the differential pressure time series curve is truncated to obtain a first curve pair, and each curve in the first curve pair is labeled with a tower position role identifier; according to the tower position role identifier and the pressure equalization stage identifier, a single tower reference curve is retrieved to obtain a second curve pair; residuals are extracted from the first curve pair and the second curve pair at each time step to obtain a residual curve pair; the direction of the residual curve pair is compared to obtain an anomaly source discrimination result, and a differential pressure early warning result is generated based on the anomaly source discrimination result.

[0008] As a preferred embodiment of the hydrogen purification adsorption tower differential pressure early warning method based on operating condition adaptation described in this invention, the step of curve truncation includes: identifying the start time and end time of pressure equalization from the pressure equalization stage identifier; using the start time and end time of pressure equalization as interval boundaries, trunculating two curves in the differential pressure time series curve that are in a pressure equalization coupling relationship to obtain the first curve pair; and obtaining the tower position role identifier for each curve in the first curve pair based on the pressure change trend of the corresponding adsorption tower during this pressure equalization process.

[0009] As a preferred embodiment of the hydrogen purification adsorption tower differential pressure early warning method based on operating condition adaptation described in this invention, the step of retrieving the single-tower reference curve includes: retrieving differential pressure curves for multiple historical pressure equalization cycles under the corresponding role and stage from the operating condition morphology library according to the tower position role identifier and the pressure equalization stage identifier, to obtain a set of historical differential pressure curves; performing time-by-time averaging on each curve in the set of historical differential pressure curves to obtain the single-tower reference curve; and retrieving the corresponding single-tower reference curves for each of the two curves in the first curve pair to obtain the second curve pair.

[0010] As a preferred embodiment of the hydrogen purification adsorption tower differential pressure early warning method based on operating condition adaptation described in this invention, the step of extracting the residuals time-by-time includes: using the starting time of the equal pressure in the first curve pair and the second curve pair as the alignment reference point; performing time axis translation processing on each curve in the first curve pair and the corresponding single-tower reference curve in the second curve pair to obtain an aligned curve group; and performing time-by-time subtraction processing on the differential pressure amplitude corresponding to each curve in the aligned curve group at the same time to obtain the residual curve pair.

[0011] As a preferred embodiment of the hydrogen purification adsorption tower differential pressure early warning method based on operating condition adaptation described in this invention, the step of performing the direction comparison includes: performing time-by-time numerical accumulation processing on the two residual curves in each segment of the established residual segment sequence to obtain segmented residual accumulation value pairs; judging the positive and negative signs of each accumulation value in the segmented residual accumulation value pairs, and determining the positive and negative sign judgment results as the segment direction of the corresponding residual curve to obtain the first residual direction and the second residual direction; judging the consistency between the first residual direction and the second residual direction, including consistent segments in the synchronous residual segment, and including inconsistent segments in the independent residual segment.

[0012] As a preferred embodiment of the hydrogen purification adsorption tower differential pressure early warning method based on operating condition adaptation described in this invention, the step of establishing the residual segment sequence includes: detecting extreme points of each residual curve in the residual curve pair to obtain a set of extreme point locations; using each extreme point in the extreme point location set as the segment boundary, segmenting each residual curve in the residual curve pair to obtain the residual segment sequence.

[0013] As a preferred embodiment of the hydrogen purification adsorption tower differential pressure early warning method based on operating condition adaptation described in this invention, the step of obtaining the anomaly source discrimination result includes: determining the time interval corresponding to the synchronous residual segment as the conduction offset interval; performing amplitude accumulation processing on the two residual curves in the independent residual segment to obtain residual magnitude pairs; comparing the two magnitudes in the residual magnitude pairs, and determining the tower number to which the residual curve corresponding to the magnitude with the largest value belongs as the anomaly source tower; determining the time interval corresponding to the independent residual segment as the independent offset interval of this tower, and associating it with the anomaly source tower to form the anomaly source discrimination result.

[0014] As a preferred embodiment of the hydrogen purification adsorption tower differential pressure early warning method based on operating condition adaptation described in this invention, the step of generating the differential pressure early warning result includes: determining the early warning level based on the number of independent offset intervals of the tower and their corresponding durations; associating and encapsulating the early warning level, the tower of the abnormal source, and the independent offset intervals of the tower to obtain the differential pressure early warning result; and sending the differential pressure early warning result to the adsorption tower control system.

[0015] As a preferred embodiment of the hydrogen purification adsorption tower differential pressure early warning method based on operating condition adaptation described in this invention, the method further includes: feeding back the residual curve pair corresponding to the conduction offset interval to the operating condition pattern library; updating the single tower reference curve corresponding to the tower position role identifier in the operating condition pattern library to obtain the updated operating condition pattern library.

[0016] This invention provides a differential pressure early warning system for hydrogen purification adsorption towers based on adaptive operating conditions.

[0017] To solve the above technical problems, the present invention provides the following technical solution: a hydrogen purification adsorption tower differential pressure early warning system based on operating condition adaptation, comprising: a curve truncation and labeling module, used to truncate the differential pressure time series curve according to the pressure equalization stage identifier issued by the adsorption tower control system, to obtain a first curve pair, and to label the tower position role identifier on each curve in the first curve pair.

[0018] The baseline curve retrieval module is used to retrieve the single-tower baseline curve based on the tower position role identifier and the pressure equalization stage identifier to obtain the second curve pair;

[0019] The residual extraction module is used to extract the residuals of the first curve pair and the second curve pair at each time step to obtain the residual curve pair.

[0020] The anomaly detection and early warning module is used to compare the directions of the residual curve pairs to obtain the anomaly source detection result, and generate a differential pressure early warning result based on the anomaly source detection result.

[0021] The beneficial effects of this invention are as follows: By retrieving the single-tower reference curve based on the tower position role identifier and the pressure equalization stage identifier, the residual curve pair is obtained by extracting the residuals of the first curve pair and the retrieved second curve pair at each time step. Then, the direction of the residual curve pair is compared: the segments with the same residual direction are included in the synchronous residual segment and determined as the conduction offset interval, and the segments with inconsistent residual directions are included in the independent residual segment and determined as the independent offset interval of this tower. This distinguishes the synchronous offset of the two towers caused by the fluctuation of the operating conditions from the independent offset caused by the fault of a single tower body, and avoids misjudging the normal linkage change corresponding to the conduction offset interval as an adsorption tower abnormality.

[0022] Meanwhile, based on the comparison of the residual magnitudes of the two residual curves in the independent residual segment, this invention identifies the tower corresponding to the one with the larger residual magnitude as the source of the anomaly, and associates the independent offset interval of this tower with the source of the anomaly to form an anomaly source discrimination result, thereby achieving accurate positioning of the source of pressure difference anomalies and improving the accuracy and usability of pressure difference early warning results. Attached Figure Description

[0023] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1The above is a flowchart of an overall method for early warning of differential pressure in a hydrogen purification adsorption tower based on adaptive operating conditions, provided as an embodiment of the present invention.

[0025] Figure 2 This is a schematic diagram illustrating an application scenario of a differential pressure early warning method for a hydrogen purification adsorption tower based on adaptive operating conditions, provided in one embodiment of the present invention. Detailed Implementation

[0026] To make the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0027] Example 1, referring to Figure 1 This is one embodiment of the present invention, which provides a method for early warning of differential pressure in a hydrogen purification adsorption tower based on adaptive operating conditions, comprising the following steps:

[0028] Based on the pressure equalization stage indicator issued by the adsorption tower control system, the pressure difference time series curve is truncated to obtain the first curve pair, and the tower position role indicator is marked on each curve in the first curve pair.

[0029] Based on the tower location role identifier and the pressure equalization stage identifier, retrieve the single tower reference curve to obtain the second curve pair;

[0030] Time-by-time residual extraction is performed on the first curve pair and the second curve pair to obtain residual curve pairs;

[0031] The residual curves are compared in direction to obtain the anomaly source identification result, and a differential pressure warning result is generated based on the anomaly source identification result.

[0032] Existing differential pressure warning methods for adsorption towers cannot distinguish between differential pressure deviations caused by fluctuations in operating conditions between two towers and differential pressure deviations caused by anomalies in a single tower. They also cannot determine the specific tower from which the differential pressure abnormality originates, resulting in low accuracy and poor usability of differential pressure warning results.

[0033] By retrieving the single-tower baseline curve based on the tower position role identifier and the pressure equalization stage identifier, the residual curve pair is obtained by extracting the residuals of the first curve pair and the retrieved second curve pair at each time step. Then, the direction of the residual curve pair is compared: the segments with the same residual direction are included in the synchronous residual segment and determined as the transmission offset interval, and the segments with inconsistent residual directions are included in the independent residual segment and determined as the independent offset interval of this tower. In this way, the synchronous offset of the two towers caused by the fluctuation of the operating condition is distinguished from the independent offset caused by the fault of a single tower body, and the normal linkage change corresponding to the transmission offset interval is misjudged as an adsorption tower anomaly.

[0034] Meanwhile, based on the comparison of the residual magnitudes of the two residual curves in the independent residual segment, this invention identifies the tower corresponding to the one with the larger residual magnitude as the source of the anomaly, and associates the independent offset interval of this tower with the source of the anomaly to form an anomaly source discrimination result, thereby achieving accurate positioning of the source of pressure difference anomalies and improving the accuracy and usability of pressure difference early warning results.

[0035] Example 2, refer to Figures 1-2 As an embodiment of the present invention, based on the previous embodiment, a method for early warning of differential pressure in a hydrogen purification adsorption tower based on adaptive operating conditions is provided, comprising the following steps:

[0036] S1. Based on the pressure equalization stage identifier issued by the adsorption tower control system, the pressure difference time series curve is truncated to obtain the first curve pair, and the tower position role identifier is marked on each curve in the first curve pair.

[0037] In this embodiment, the hydrogen purification device includes multiple adsorption towers connected in parallel, numbered 1# to 8# respectively. Each adsorption tower is equipped with a differential pressure detection device, which is used to collect the differential pressure signal of the adsorption tower according to a preset sampling period and upload the collected data to the database of the monitoring system. A continuously updated differential pressure time series curve is formed for each adsorption tower.

[0038] The adsorption tower control system controls the operation of valves such as the inlet valve and pressure equalization valve of each adsorption tower according to the preset process sequence table, so that each adsorption tower switches between process stages such as adsorption, pressure equalization and regeneration. When the control system executes a certain pressure equalization process, it generates and issues a pressure equalization stage identifier. This identifier carries at least the stage number to which the current pressure equalization belongs (e.g., E1), the tower numbers of the two adsorption towers participating in the current pressure equalization, and information that can characterize the start and end time of the current pressure equalization.

[0039] Specifically, the steps for truncation of the curve include S1.1 to S1.3:

[0040] S1.1 Identify the start time and end time of equalization from the equalization stage identifier.

[0041] The information representing the start and end times carried in the pressure equalization stage identifier can be the timestamp field directly generated and written into the identifier by the control system at the moment when the pressure equalization valve opens and closes, or it can be the moment when the switching signal triggered by the opening and closing actions of the pressure equalization valve is received by the processing entity.

[0042] The processing unit parses the equalization stage identifier, reads the field recording the start time to obtain the equalization start time, and reads the field recording the end time to obtain the equalization end time.

[0043] For example, the equalization stage identifier issued by the control system is as follows: stage number E1, participating tower number 2# / 5#, start time 10:03:15.000, end time 10:03:47.500.

[0044] After the processing entity parses the identifier, it identifies the start time of this equalization as 10:03:15.000 and the end time as 10:03:47.500.

[0045] S1.2. Using the start time and end time of the equalization as the interval boundary, extract two curves that are in the equalization coupling relationship from the pressure difference time series curve to obtain the first curve pair.

[0046] The pressure equalization stage identifier also includes the tower numbers of the two adsorption towers participating in the pressure equalization. These two adsorption towers are connected to the pressure equalization valve through the pressure equalization pipeline. During this pressure equalization process, the pressure of one adsorption tower is transmitted to the other adsorption tower through the pressure equalization valve. The pressure difference changes of the two towers are coupled and correspond to each other in time, that is, they are in a pressure equalization coupling relationship.

[0047] Based on the participating tower number recorded in the pressure equalization stage identifier, the processing entity retrieves the corresponding differential pressure time series curves of the two adsorption towers from the database. Using the pressure equalization start time and pressure equalization end time identified in step S1.1 as the interval boundary, the processing entity extracts curve segments within the time interval from the two differential pressure time series curves. The extracted two curve segments are then combined to obtain the first curve pair.

[0048] In this embodiment, based on the participating tower numbers 2# and 5# recorded in the pressure equalization stage identifier, the processing entity retrieves the pressure difference time series curves corresponding to adsorption tower 2# and adsorption tower 5# from the database respectively; with 10:03:15.000 as the starting boundary and 10:03:47.500 as the ending boundary, curve segments within this time interval are extracted from the pressure difference time series curves of adsorption tower 2# and adsorption tower 5# respectively, to obtain the first curve pair composed of the curve segments of adsorption tower 2# and adsorption tower 5#.

[0049] S1.3. Based on the pressure change trend of the corresponding adsorption tower during the pressure equalization process, obtain the tower position role identifier for each curve in the first curve.

[0050] The processing unit performs pressure change trend analysis on each curve in the first curve pair. This analysis can be done by comparing the values ​​at the start and end points of the curve segment, or by performing a linear fit on the curve segment and taking the sign of the slope of the fitted line.

[0051] If the curve segment shows a downward trend, that is, the pressure of the corresponding adsorption tower changes from high to low during this pressure equalization process, then the adsorption tower corresponding to the curve is marked as a pressure equalization and reduction tower.

[0052] If the curve segment shows an upward trend, that is, the pressure of the corresponding adsorption tower changes from low to high during this pressure equalization process, then the adsorption tower corresponding to the curve is marked as the pressure equalization tower role, thus obtaining the tower position role identifier corresponding to each curve in the first curve pair.

[0053] In this embodiment, the trend of the curve segment of adsorption tower #2 in the first curve is determined. The pressure difference at the starting point is 0.82 MPa and the pressure difference at the ending point is 0.48 MPa. The value changes from high to low, showing a downward trend. Therefore, the tower position role of adsorption tower #2 in this pressure equalization process is identified as pressure equalization tower reduction. The trend of the curve segment of adsorption tower #5 is determined. The pressure difference at the starting point is 0.15 MPa and the pressure difference at the ending point is 0.48 MPa. The value changes from low to high, showing an upward trend. Therefore, the tower position role of adsorption tower #5 in this pressure equalization process is identified as pressure equalization tower elevation.

[0054] At this point, the two curves in the first curve are labeled with the corresponding tower positions and roles: tower #2 is a pressure equalization and reduction tower, and tower #5 is a pressure equalization and increase tower.

[0055] S2. Based on the tower position role identifier and the pressure equalization stage identifier, retrieve the single tower reference curve to obtain the second curve pair.

[0056] The steps for retrieving the single-tower reference curve include S2.1 to S2.3:

[0057] S2.1 Based on the tower position role identifier and the pressure equalization stage identifier, retrieve the pressure difference curves of multiple historical pressure equalization cycles for the corresponding role and the corresponding stage from the working condition form library to obtain a set of historical pressure difference curves.

[0058] For each curve in the first curve pair, the processing unit uses the tower number of the adsorption tower corresponding to the curve, the tower position role identifier obtained from the annotation, and the stage number carried in the current pressure equalization stage identifier as search conditions to perform a matching search in the operating condition database. It reads multiple historical differential pressure curve records that meet the conditions of tower number, role, and stage number, and combines the multiple historical differential pressure curves obtained from the search to obtain the set of historical differential pressure curves corresponding to the adsorption tower.

[0059] For the curve segment of adsorption tower #2 in the first curve, its tower position role is identified as pressure equalization tower, and the current pressure equalization stage number is E1. The main processing body uses "tower number 2#, role pressure equalization tower, stage number E1" as the search criteria to retrieve 20 differential pressure curves collected from the operating condition database for the previous 20 historical pressure equalization cycles of this tower under this role and this stage. These 20 curves are combined to obtain the set of historical differential pressure curves corresponding to adsorption tower #2.

[0060] Similarly, for the curve segment of adsorption tower #5, the historical differential pressure curve set corresponding to the tower can be retrieved by using "tower number 5#, role of equalizing pressure tower, stage number E1" as search criteria.

[0061] S2.2. The historical differential pressure curve set is averaged time-by-time to obtain the single-tower reference curve.

[0062] The processing unit uses the average pressure start time corresponding to each curve in the historical differential pressure curve set as the zero point of time, and aligns the curves according to the same sampling interval. After alignment, for each same relative time, the differential pressure value corresponding to each curve in the historical differential pressure curve set at that relative time is read, and the arithmetic mean of the multiple differential pressure values ​​is performed to obtain the average differential pressure value at that relative time. The average differential pressure values ​​corresponding to each relative time are connected in chronological order to obtain the single tower reference curve.

[0063] For example, the historical pressure difference curve set corresponding to adsorption tower #2 contains 20 curves, each represented by 65 sampling points with a sampling interval of 0.5s. After aligning the average pressure start time of each curve as the zero point, the processing unit takes the relative time as 16.0s and reads the pressure difference values ​​of the 20 curves at this relative time, which are 0.63MPa, 0.65MPa, 0.61MPa, etc. The arithmetic mean is then applied to obtain the average pressure difference value at this relative time as 0.64MPa.

[0064] By averaging the 65 relative time points in this way, the 65 average pressure difference values ​​were connected in chronological order to obtain the single-tower reference curve corresponding to adsorption tower #2.

[0065] S2.3. Retrieve the corresponding single-tower reference curves for each of the two curves in the first curve pair to obtain the second curve pair.

[0066] Specifically, for each curve in the first curve pair, the processing unit reads the single-tower reference curve calculated for that tower number in step S2.2 according to its corresponding tower number, and combines the single-tower reference curves retrieved for each of the two curves in the first curve pair to obtain the second curve pair.

[0067] The first curve pair consists of curve segments from adsorption tower #2 and adsorption tower #5. The processing unit reads the single-tower reference curves corresponding to adsorption tower #2 and adsorption tower #5 respectively, and combines these two single-tower reference curves to obtain the second curve pair, which corresponds one-to-one with the curve segments from adsorption tower #2 and adsorption tower #5 in the first curve pair.

[0068] S3. Perform time-by-time residual extraction on the first curve pair and the second curve pair to obtain the residual curve pair.

[0069] Specifically, the steps for performing the time-by-time residual extraction include S3.1 to S3.3:

[0070] S3.1. The starting time of the equal pressure between the first curve pair and the second curve pair is taken as the alignment reference point.

[0071] The first curve contains the actual start time of the equalization process, which is the equalization start time identified in S1.1.

[0072] The second curve carries the relative time zero point on each single-tower reference curve in S2.2, which is the basis for the time-by-time averaging process. This zero point represents the starting time of the averaging in the historical averaging cycle.

[0073] The processing unit reads the relative zero point of the starting time of the equalization of the first curve and the corresponding time of the second curve, and determines the two as the alignment reference points for subsequent time axis translation processing.

[0074] In this embodiment, the pressure equalization start time corresponding to the curve segment of adsorption tower #2 in the first curve alignment is 10:03:15.000; the relative time zero point corresponding to the single-tower reference curve of adsorption tower #2 in the second curve alignment is 0.0s. The processing unit determines 10:03:15.000 as the alignment reference point for the curve segment of adsorption tower #2 and 0.0s as the alignment reference point for the single-tower reference curve of adsorption tower #2; the alignment reference points for the curve segment of adsorption tower #5 and its corresponding single-tower reference curve are determined in the same way.

[0075] S3.2. Perform time axis translation processing on each curve in the first curve pair and the corresponding single-tower reference curve in the second curve pair to obtain an aligned curve group.

[0076] The processing unit uses the determined alignment reference point as a reference point and shifts the time axis of each curve in the first curve alignment, so that the time coordinate corresponding to the start time of the equalization is transformed to a relative time of 0.0s. That is, the relative time corresponding to the data points on the curve that were originally collected t seconds after the start time of the equalization is recorded as t after the shift.

[0077] For each single-tower reference curve in the second curve, since its relative time zero point already represents the start time of pressure equalization, its original time axis remains unchanged.

[0078] After processing, the corresponding curves in the first curve pair and the second curve pair are comparable under the same relative time coordinates. The first curve pair after translation is combined with the second curve pair that remains unchanged to obtain the aligned curve group.

[0079] For example, in the curve segment of adsorption tower #2, the pressure difference value collected 16.0s after 10:03:15.000, i.e. at 10:03:31.000, is 0.63MPa. After time axis shifting, the relative time corresponding to this data point is changed from 10:03:31.000 to 16.0s.

[0080] The average pressure difference at 16.0 s relative to the reference curve of adsorption tower #2 is 0.64 MPa, and its time axis remains unchanged. After translation processing, the curve segment of adsorption tower #2 corresponds to the data of the reference curve of adsorption tower #2 at 16.0 s relative to the reference curve of adsorption tower #2. The curve segment of adsorption tower #5 is aligned with the reference curve of adsorption tower #5 in the same way. The processing unit combines the first pair of curves after the above translation processing with the corresponding second pair of curves to obtain the aligned curve group.

[0081] S3.3. Subtract the differential pressure amplitude corresponding to each curve in the alignment curve group at the same time moment by moment to obtain the residual curve pair.

[0082] Specifically, the processing unit traverses the common relative time coordinates of each pair of corresponding curves in the alignment curve group (i.e., the curve segment of the same adsorption tower and its single tower reference curve).

[0083] At each relative moment, the differential pressure amplitude corresponding to the curve segment at that moment is subtracted from the differential pressure amplitude corresponding to the single-tower reference curve at that moment to obtain the residual value at that moment;

[0084] Connect the residual values ​​corresponding to each relative time point in chronological order to obtain a residual curve corresponding to the adsorption tower.

[0085] Perform the above subtraction process on each pair of curves in the aligned curve group, and the resulting combination of two residual curves is the residual curve pair.

[0086] Furthermore, at a relative time of 16.0s, the pressure difference amplitude of the curve segment of adsorption tower #2 is 0.63MPa, while the pressure difference amplitude of its single-tower reference curve at the same relative time is 0.64MPa. Subtracting the two, we obtain the residual value of adsorption tower #2 at that time as -0.01MPa.

[0087] By subtracting all the sampling points at the relative time coordinates in this way, the residual curve corresponding to adsorption tower #2 is obtained.

[0088] Similarly, by subtracting the curve segment of adsorption tower #5 from its single-tower reference curve at each time step, the residual curve corresponding to adsorption tower #5 is obtained. The processing unit then combines the residual curves of adsorption tower #2 and #5 to obtain a residual curve pair.

[0089] S4. Perform directional comparison on the residual curve pair to obtain the anomaly source identification result, and generate differential pressure warning result based on the anomaly source identification result.

[0090] Specifically, the steps for performing the direction comparison include S4.1 to S4.3:

[0091] S4.1. Perform time-by-time numerical accumulation on the two residual curves in each segment of the established residual segment sequence to obtain the segmented residual accumulation value pairs.

[0092] For each segment in the residual segment sequence, the processing unit reads all the residual values ​​contained in the residual curve segments of adsorption tower #2 and adsorption tower #5 within that segment. The residual values ​​in each curve segment are accumulated one by one in time sequence to obtain the accumulated values ​​corresponding to the two residual curves under that segment. The two accumulated values ​​are combined to obtain the segment residual accumulated value pair corresponding to that segment. The above accumulation process is performed on each segment in the residual segment sequence.

[0093] For segment 1 (0.0s~8.0s), the processing unit accumulates the residual values ​​at each time step within the residual curve segment of adsorption tower #2, and the accumulated value is -0.15MPa.

[0094] The residual values ​​at each time step within the residual curve segment of adsorption tower #5 were accumulated, and the accumulated value was -0.12 MPa.

[0095] Combining these two accumulated values, we obtain the segmented residual accumulated value pair corresponding to segment one as (-0.15MPa, -0.12MPa).

[0096] For segment two (8.0s~20.0s), the cumulative value of adsorption tower #2 is obtained as +0.45MPa and the cumulative value of adsorption tower #5 is -0.08MPa. The cumulative residual value of the segment is (+0.45MPa, -0.08MPa).

[0097] For segment three [20.0s~32.0s], the cumulative value of adsorption tower #2 is +0.05MPa and the cumulative value of adsorption tower #5 is +0.06MPa. The cumulative residual value pair of segments is (+0.05MPa, +0.06MPa).

[0098] The steps for establishing the residual segmented sequence include B1 to B2:

[0099] B1. Perform extreme point detection on each residual curve in the residual curve pair to obtain the set of extreme point locations.

[0100] The processing unit performs extreme point detection on the residual curves of adsorption tower #2 and adsorption tower #5 respectively. The detection method is to compare the residual values ​​of adjacent sampling points point by point.

[0101] If the residual value of a sampling point is greater than the residual values ​​of the previous and next sampling points at the same time, then the point is determined to be a maximum point; if the residual value of a sampling point is less than the residual values ​​of the previous and next sampling points at the same time, then the point is determined to be a minimum point.

[0102] The relative time positions of all detected maxima and minima are summarized to obtain the set of extreme point positions corresponding to the residual curve. The above detection is performed on the two residual curves in the residual curve pair, and the extreme point position sets of the two curves are merged to obtain the total extreme point position set.

[0103] Extreme point detection was performed on the residual curve of adsorption tower #2, and a minimum point was detected at a relative time of 8.0 s, and a maximum point was detected at a relative time of 20.0 s. Extreme point detection was also performed on the residual curve of adsorption tower #5, and a maximum point was detected at a relative time of 8.0 s, and a maximum point was detected at a relative time of 20.0 s. The detection results were combined to obtain the set of extreme point locations {8.0 s, 20.0 s}.

[0104] B2. Using each extreme point in the set of extreme point locations as the segmentation boundary, segment the residual curves in the residual curve pair to obtain the residual segment sequence.

[0105] Arrange the extreme point locations in the extreme point location set, along with the starting relative time of 0.0s and the ending relative time of this equalization process, in chronological order, and determine the interval between two adjacent time points as a segmented interval;

[0106] Based on the segmented intervals obtained, the two residual curves in the residual curve pair are cut off respectively, so that each residual curve is divided into several segments. The curve segments in the same segment interval of the two residual curves correspond to each other, thus obtaining the residual segment sequence.

[0107] In this embodiment, the set of extreme point locations is {8.0s, 20.0s}. Adding the starting relative time of 0.0s and the ending relative time of 32.0s, the segment boundaries are obtained by arranging them in chronological order: 0.0s, 8.0s, 20.0s, and 32.0s. This divides the time into three segment intervals: segment one [0.0s~8.0s], segment two [8.0s~20.0s], and segment three [20.0s~32.0s].

[0108] Based on the above segmented intervals, the processing unit extracts the residual curves of adsorption tower #2 and adsorption tower #5 respectively, resulting in a residual segment sequence consisting of three segments and two corresponding curve segments within each segment.

[0109] S4.2. Determine the sign of each accumulated value in the segmented residual accumulated value pair, and determine the sign determination result as the segmented direction of the corresponding residual curve to obtain the first residual direction and the second residual direction.

[0110] For each segment of the cumulative residual value pair, the processing unit determines the sign of the two cumulative values: if the cumulative value is greater than 0, the result is positive; if the cumulative value is less than 0, the result is negative.

[0111] The sign determination result of the cumulative value corresponding to the residual curve of adsorption tower #2 is determined as the first residual direction of this segment, and the sign determination result of the cumulative value corresponding to the residual curve of adsorption tower #5 is determined as the second residual direction of this segment.

[0112] For example, the cumulative residual value pair corresponding to segment one is (-0.15MPa, -0.12MPa), and the first residual direction is determined to be negative and the second residual direction is negative; the cumulative residual value pair corresponding to segment two is (+0.45MPa, -0.08MPa), and the first residual direction is determined to be positive and the second residual direction is negative; the cumulative residual value pair corresponding to segment three is (+0.05MPa, +0.06MPa), and the first residual direction is determined to be positive and the second residual direction is positive.

[0113] S4.3. Perform consistency judgment on the first residual direction and the second residual direction, include consistent segments in the synchronous residual segment, and include inconsistent segments in the independent residual segment.

[0114] For each segment, the processing entity compares whether the first residual direction and the second residual direction corresponding to that segment are the same: if they are the same, the segment is included in the synchronous residual segment; if they are different, the segment is included in the independent residual segment.

[0115] For segment 1, both the first and second residuals are negative, so they are consistent and segment 1 is included in the synchronous residual segment. For segment 2, both the first and second residuals are positive, so they are inconsistent and segment 2 is included in the independent residual segment. For segment 3, both the first and second residuals are positive, so segment 3 is included in the synchronous residual segment. Based on the above criteria, the synchronous residual segment includes segment 1 and segment 3, and the independent residual segment includes segment 2.

[0116] The steps for obtaining the anomaly source determination result include S4.4~S4.7:

[0117] S4.4. Determine the time interval corresponding to the synchronous residual segment as the conduction offset interval.

[0118] The processing unit converts the relative time intervals corresponding to the divided synchronous residual segments into the actual clock time intervals corresponding to this voltage equalization process, and determines the converted time intervals as the conduction offset intervals.

[0119] For example, the synchronous residual segmentation includes segment one (relative time 0.0s~8.0s) and segment three (relative time 20.0s~32.0s). Based on the equalization start time of 10:03:15.000, the conduction offset intervals are 10:03:15.000~10:03:23.000 and 10:03:35.000~10:03:47.000.

[0120] S4.5. Perform amplitude accumulation processing on the two residual curves in the independent residual segment to obtain residual magnitude pairs.

[0121] Specifically, the processing unit performs a step-by-step summation of the magnitude (i.e., absolute value) of the residual curve segments of adsorption tower #2 and adsorption tower #5 within each independent residual segment, obtaining the residual magnitude corresponding to that curve. The residual magnitude corresponding to the residual curve of adsorption tower #2 and the residual magnitude corresponding to the residual curve of adsorption tower #5 are then combined to obtain a residual magnitude pair.

[0122] If the independent residual segment is segment two (relative time 8.0s~20.0s), the processing body will sum up the amplitude of the residual values ​​at each time point within the segment of the residual curve of adsorption tower #2, and obtain a residual magnitude of 0.52MPa.

[0123] The amplitudes of the residual values ​​at each time point within the residual curve segment of adsorption tower #5 were summed one by one to obtain a residual magnitude of 0.10 MPa; the two residual magnitudes were combined to obtain a residual magnitude pair of (0.52 MPa, 0.10 MPa).

[0124] S4.6. Compare the values ​​of the two magnitudes of the residual magnitude pair, and determine the tower number to which the residual curve corresponding to the magnitude with the largest value belongs as the tower from which the anomaly originates.

[0125] The residual magnitude pairs are (0.52MPa, 0.10MPa), where 0.52MPa corresponds to the residual curve of adsorption tower #2 and 0.10MPa corresponds to the residual curve of adsorption tower #5. The comparison shows that 0.52MPa has the largest value, so tower number #2, which corresponds to 0.52MPa, is identified as the tower from which the anomaly originated.

[0126] S4.7. Determine the time interval corresponding to the independent residual segment as the independent offset interval of this tower, and associate it with the tower of the anomaly source to form the anomaly source discrimination result.

[0127] The processing unit converts the relative time intervals corresponding to the independent residual segments into actual clock time intervals and determines them as independent offset intervals of this tower.

[0128] The independent offset interval of this tower is associated with the anomaly source tower determined in step S4.6, and together with the conduction offset interval obtained in step S4.4, they are combined to form the anomaly source discrimination result.

[0129] The relative time interval corresponding to the independent residual segment (segment two) is 8.0s~20.0s, which is converted into the actual clock time interval of 10:03:23.000~10:03:35.000, and is determined to be the independent offset interval of this tower. This independent offset interval of this tower is associated with the anomaly source tower #2, and the result is "independent offset interval of this tower 10:03:23.000~10:03:35.000, anomaly source tower #2". Together with the conduction offset intervals 10:03:15.000~10:03:23.000 and 10:03:35.000~10:03:47.000, they are combined to form the anomaly source identification result.

[0130] The steps for generating the differential pressure warning result include S4.8~S4.10:

[0131] S4.8. Determine the warning level based on the number of independent offset intervals of this tower and their corresponding duration.

[0132] The system processes the number of independent offset intervals of the tower in the statistical anomaly source identification results, calculates the duration of each independent offset interval, and compares the number and duration of independent offset intervals of the tower with preset quantity and duration thresholds according to the pre-set level classification rules to determine the corresponding warning level.

[0133] In one specific implementation, the pre-defined hierarchy classification rules are as follows:

[0134] If the number of independent offset intervals of this tower is 1 and the duration is less than 10 seconds, it is judged as a level 3 warning; if the number of independent offset intervals of this tower is 1 and the duration is between 10 seconds and 20 seconds, it is judged as a level 2 warning.

[0135] If the number of independent offset intervals of this tower reaches two or more, or the duration of a single independent offset interval of this tower exceeds 20 seconds, it is determined to be a Level 1 warning. In the anomaly source identification results of this time, the number of independent offset intervals of this tower is 1, with a corresponding duration of 12.0 seconds from 10:03:23.000 to 10:03:35.000, which is between 10 seconds and 20 seconds. Therefore, the warning level of this time is determined to be a Level 2 warning.

[0136] S4.9. Associate and encapsulate the warning level, the tower of the abnormal source, and the independent offset interval of this tower to obtain the differential pressure warning result.

[0137] The processing unit associates and encapsulates the determined warning level, the tower from which the anomaly originates, and the independent offset range of the tower to form a structured differential pressure warning result.

[0138] For example, the processing entity sets the warning level to Level II, the source of the anomaly to Adsorption Tower #2, and the independent offset range of this tower to 10:03:23.000~10:03:35.000. Then, it associates and encapsulates these three information to obtain the differential pressure warning result as {Warning Level: Level II; Source of Anomaly: Tower #2; Independent Offset Range of This Tower: 10:03:23.000~10:03:35.000; Corresponding Pressure Equalization Stage: E1}.

[0139] S4.10. Send the differential pressure warning result to the adsorption tower control system.

[0140] The processing unit sends the differential pressure warning result to the adsorption tower control system through the communication interface. After receiving the result, the adsorption tower control system displays the prompt information in the form of a level-two warning on the monitoring interface for operation and maintenance personnel to refer to and troubleshoot.

[0141] It is important to know that steps A1 to A2 are also included:

[0142] A1. Feed back the residual curve pair corresponding to the transmission offset interval to the working condition database.

[0143] In subsequent steps, after determining the transmission offset interval based on the direction comparison results of the residual curve pairs, the processing entity feeds back the residual curve pairs within the time period corresponding to the transmission offset interval, along with the original differential pressure curves corresponding to the first curve pair, to the operating condition database as new historical data to be merged.

[0144] A2. Update the single tower reference curve corresponding to the tower position role identifier in the working condition form library to obtain the updated working condition form library.

[0145] The processing entity merges the current differential pressure curve obtained in step A1 into the historical differential pressure curve set under the corresponding tower number, corresponding tower position role identifier, and corresponding stage number. The time-by-time averaging process in step S2.2 is re-executed on the merged historical differential pressure curve set to obtain the updated single-tower reference curve. The updated single-tower reference curve is then used to replace the original single-tower reference curve in the operating condition form library to obtain the updated operating condition form library.

[0146] Example 3 is an embodiment of the present invention, which provides a differential pressure early warning system for a hydrogen purification adsorption tower based on adaptive operating conditions, comprising:

[0147] The curve extraction and annotation module is used to extract the pressure difference time series curve based on the pressure equalization stage identifier issued by the adsorption tower control system, obtain the first curve pair, and annotate the tower position role identifier on each curve in the first curve pair.

[0148] The baseline curve retrieval module is used to retrieve the single-tower baseline curve based on the tower position role identifier and the pressure equalization stage identifier to obtain the second curve pair;

[0149] The residual extraction module is used to extract the residuals of the first curve pair and the second curve pair at each time step to obtain the residual curve pair.

[0150] The anomaly detection and early warning module is used to compare the directions of the residual curve pairs to obtain the anomaly source detection result, and generate a differential pressure early warning result based on the anomaly source detection result.

[0151] This embodiment also provides a computer device applicable to a condition-adaptive hydrogen purification adsorption tower differential pressure early warning method, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the condition-adaptive hydrogen purification adsorption tower differential pressure early warning method proposed in the above embodiment.

[0152] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements a differential pressure early warning method for hydrogen purification adsorption tower based on adaptive operating conditions as proposed in the above embodiment.

[0153] The storage medium proposed in this embodiment and the method for early warning of differential pressure in a hydrogen purification adsorption tower based on operating condition adaptation proposed in the above embodiment belong to the same inventive concept. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0154] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0155] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for early warning of differential pressure in a hydrogen purification adsorption tower based on adaptive operating conditions, characterized in that, Includes the following steps: Based on the pressure equalization stage indicator issued by the adsorption tower control system, the pressure difference time series curve is truncated to obtain the first curve pair, and the tower position role indicator is marked on each curve in the first curve pair. Based on the tower location role identifier and the pressure equalization stage identifier, retrieve the single tower reference curve to obtain the second curve pair; Time-by-time residual extraction is performed on the first curve pair and the second curve pair to obtain residual curve pairs; The residual curves are compared in direction to obtain the anomaly source identification result, and a differential pressure warning result is generated based on the anomaly source identification result.

2. The method for early warning of differential pressure in a hydrogen purification adsorption tower based on adaptive operating conditions as described in claim 1, characterized in that, The steps for performing the curve truncation include: Identify the start and end times of the equalization phase from the equalization phase identifiers; Using the pressure equalization start time and the pressure equalization end time as the interval boundary, two curves that are in pressure equalization coupling relationship are extracted from the pressure difference time series curve to obtain the first curve pair; Based on the pressure change trend of the corresponding adsorption tower during the pressure equalization process, the role identifier of the tower is obtained for each curve in the first curve pair.

3. The method for early warning of differential pressure in a hydrogen purification adsorption tower based on adaptive operating conditions as described in claim 2, characterized in that, The steps for retrieving the single-tower reference curve include: Based on the tower position role identifier and the pressure equalization stage identifier, the pressure difference curves of multiple historical pressure equalization cycles under the corresponding role and the corresponding stage are retrieved from the working condition form library to obtain a set of historical pressure difference curves. The single-tower reference curve is obtained by averaging each curve in the set of historical differential pressure curves over time. The corresponding single-tower reference curve is retrieved from each of the two curves in the first curve pair to obtain the second curve pair.

4. The method for early warning of differential pressure in a hydrogen purification adsorption tower based on adaptive operating conditions as described in claim 3, characterized in that, The steps for performing the time-by-time residual extraction include: The starting point of the average pressure between the first curve pair and the second curve pair is used as the alignment reference point; The time axis is shifted between each curve in the first curve pair and the corresponding single-tower reference curve in the second curve pair to obtain an aligned curve group. The differential pressure amplitudes corresponding to each curve in the alignment curve group at the same time are subtracted time by time to obtain the residual curve pair.

5. The method for early warning of differential pressure in a hydrogen purification adsorption tower based on adaptive operating conditions as described in claim 4, characterized in that, The steps for performing the orientation comparison include: The two residual curves in each segment of the established residual segment sequence are accumulated time-by-time to obtain the segmented residual accumulation value pairs. The sign of each accumulated value in the segmented residual accumulated value pair is determined, and the sign determination result is determined as the segmentation direction of the corresponding residual curve to obtain the first residual direction and the second residual direction; Consistency is determined between the first residual direction and the second residual direction. Consistent segments are included in the synchronous residual segments, while inconsistent segments are included in the independent residual segments.

6. The method for early warning of differential pressure in a hydrogen purification adsorption tower based on adaptive operating conditions as described in claim 5, characterized in that, The steps for constructing a residual segmented sequence include: Extreme point detection is performed on each residual curve in the residual curve pair to obtain the set of extreme point locations; Using each extreme point in the set of extreme point locations as the segment boundary, the residual curves in the residual curve pair are segmented to obtain the residual segment sequence.

7. The method for early warning of differential pressure in a hydrogen purification adsorption tower based on adaptive operating conditions as described in claim 6, characterized in that, The steps to obtain the anomaly source determination result include: The time interval corresponding to the synchronous residual segment is determined as the conduction offset interval; The amplitude of each of the two residual curves in the independent residual segment is accumulated to obtain a residual magnitude pair; The two magnitudes of the residual magnitude are compared numerically, and the tower number to which the residual curve corresponding to the magnitude with the largest value belongs is determined as the tower from which the anomaly originates. The time interval corresponding to the independent residual segment is determined as the independent offset interval of this tower, and associated with the tower of the anomaly source to form the anomaly source discrimination result.

8. The method for early warning of differential pressure in a hydrogen purification adsorption tower based on adaptive operating conditions as described in claim 7, characterized in that, The steps for generating the differential pressure early warning result include: The warning level is determined based on the number of independent offset intervals of this tower and their corresponding duration. The differential pressure warning result is obtained by associating and encapsulating the warning level, the tower of the abnormal source, and the independent offset range of the tower. The differential pressure warning result is sent to the adsorption tower control system.

9. The method for early warning of differential pressure in a hydrogen purification adsorption tower based on adaptive operating conditions as described in claim 8, characterized in that, Also includes: The residual curves corresponding to the transmission offset interval are fed back to the working condition database. The single-tower reference curve corresponding to the tower position role identifier in the working condition form library is updated to obtain the updated working condition form library.

10. A differential pressure early warning system for a hydrogen purification adsorption tower based on adaptive operating conditions, employing the differential pressure early warning method for a hydrogen purification adsorption tower based on adaptive operating conditions as described in any one of claims 1 to 9, characterized in that, include: The curve extraction and annotation module is used to extract the pressure difference time series curve based on the pressure equalization stage identifier issued by the adsorption tower control system, obtain the first curve pair, and annotate the tower position role identifier on each curve in the first curve pair. The baseline curve retrieval module is used to retrieve the single-tower baseline curve based on the tower position role identifier and the pressure equalization stage identifier to obtain the second curve pair; The residual extraction module is used to extract the residuals of the first curve pair and the second curve pair at each time step to obtain the residual curve pair. The anomaly detection and early warning module is used to compare the directions of the residual curve pairs to obtain the anomaly source detection result, and generate a differential pressure early warning result based on the anomaly source detection result.

Images