A water-cooling sealed state monitoring system for a handheld end of an electromagnetic heating device

Abstract

The application discloses a kind of water-cooled sealing state monitoring systems of electromagnetic heating device handheld end, specifically related to electromagnetic heating device sealing monitoring and failure prediction field, including: acquisition module is used to collect the pressure value, flow value and temperature value corresponding to each sealing position of handheld end, and generates monitoring record according to acquisition time;Fragment generation module is used to calculate the change direction of adjacent records according to acquisition time sequence to the monitoring record of same sealing position, and the monitoring record of continuous consistency is merged into abnormal fragment, and output abnormal fragment sequence;Identity establishment module is used to compare abnormal fragment sequence of same sealing position according to appearance order section by section. By the monitoring record of each sealing position to generate abnormal fragment, damage identity and health state result, it is organized into continuous evolution chain to scattered fluctuation, so it is relatively improved to distinguish the ability of transient disturbance and irreversible damage sign under non-stop condition, and support early intervention.

Description

Technical Field

[0001] This invention relates to the field of sealing monitoring and fault prediction technology for electromagnetic heating devices, and more specifically, to a handheld water-cooled sealing status monitoring system for electromagnetic heating devices. Background Technology

[0002] In the handheld water-cooled sealing management of electromagnetic heating devices, existing technologies mostly focus on the discovery and shutdown after sealing failure. Common methods include collecting operating data such as water pressure, flow rate, temperature rise, leakage signals or alarm switches, and then combining threshold comparison, trend judgment or regular maintenance to judge the risk of loose joints, aging of sealing rings and leakage. However, in handheld electromagnetic heating operations, the handheld end is often in a state of high-frequency start-stop, alternating hot and cold, frequent changes in posture, and continuous operation without being able to stop at will. Moreover, the cooling circuit is close to the heat-generating part. It is required to complete online monitoring without disassembly, without stopping the machine, and without introducing additional operational burden. Under this condition, although the existing methods can record phenomena such as pressure fluctuations, flow fluctuations, and local temperature difference changes, it is difficult to distinguish whether these changes are due to instantaneous operating condition disturbances or because a certain sealing part has entered the irreversible damage evolution stage. In practice, the same part has multiple minor abnormalities but the conclusions are scattered, the repair is restored for a short time but then reappears, and there is no way to establish a continuous judgment basis between different batches of operation records. In the end, it can only be dealt with passively after the leakage expands or the fault takes shape, which is difficult to support the early intervention for true fault prediction and health management. Therefore, the problem is: how to continuously identify and correlate the abnormal evolution process of the water-cooled sealing parts under the condition that the handheld end of the electromagnetic heating device is running without stopping, to separate irreversible damage signs from transient disturbances, and to form a sealing health status judgment result that can be used for prediction and treatment. Summary of the Invention

[0003] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a handheld water-cooled sealing status monitoring system for an electromagnetic heating device. This system generates abnormal segments, merges damage identities, and determines the evolution status of each sealing component by analyzing the operational monitoring records of each sealing component. This results in the health status of the corresponding sealing component and provides early warning information, thereby solving the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a water-cooled sealing status monitoring system for the handheld end of an electromagnetic heating device, comprising: The data acquisition module is used to collect pressure, flow and temperature values ​​corresponding to each sealing part of the handheld terminal, and generate monitoring records according to the time of acquisition; The segment generation module is used to calculate the change direction of adjacent records of the same sealing part according to the acquisition time sequence, merge the monitoring records with continuous and consistent change directions into abnormal segments, and output the abnormal segment sequence. The identity establishment module is used to compare the abnormal segment sequences of the same sealing part in the order of appearance. When the change direction of the later abnormal segment is consistent with that of the previous abnormal segment and the same offset is still retained after the end, the latter two are classified into the same damage identity and the damage identity sequence is output. The evolution determination module is used to compare each abnormal segment contained in the same damage identity in the order of appearance. When the residual offset of the later abnormal segment is greater than the residual offset of the previous abnormal segment, the damage identity is determined to be in a state of continuous deterioration. When the residual offset increases continuously, the damage identity is determined to be in an irreversible damage state, and the health status result is output. The output module is used to generate a status record of the corresponding sealing part based on the health status result, and to output a warning message when the health status result is a state of continuous deterioration or irreversible damage.

[0005] In a preferred embodiment, the execution of the acquisition module includes: Based on the distribution of each sealing part of the handheld terminal in the water supply passage, water return passage and shell connection position, establish a one-to-one correspondence between each sealing part and the corresponding upstream pressure acquisition point, downstream pressure acquisition point, passage flow acquisition point and adjacent temperature acquisition point, and output a sealing part acquisition correspondence table. At each acquisition moment, the upstream pressure measurement, downstream pressure measurement, passage flow measurement and adjacent temperature measurement corresponding to the same sealing part are read synchronously according to the sealing part acquisition correspondence table. The difference between the upstream pressure measurement and the downstream pressure measurement is determined as the pressure value of the sealing part, the passage flow measurement is determined as the flow value of the sealing part, and the adjacent temperature measurement is determined as the temperature value of the sealing part. The acquisition value of the part is then output. The collected values ​​for each part are combined with the corresponding sealing part identifier and the collection time and written into the monitoring record, and the monitoring record sequence corresponding to each sealing part is output.

[0006] In a preferred embodiment, the execution of the fragment generation module includes: The monitoring records of the same sealing part are arranged in ascending order according to the acquisition time. The pressure change, flow rate change and temperature change between adjacent monitoring records are calculated. The pressure change sequence, flow rate change sequence and temperature change sequence are respectively subjected to three-term convolution smoothing and three-term median reconstruction. The reconstructed pressure change sign, flow rate change sign and temperature change sign form a direction vector. When the three signs of the direction vector are inconsistent, the sign combination corresponding to the sum of the absolute values ​​of the pressure residual, flow rate residual and temperature residual is taken as the change direction of the current monitoring record. The direction record sequence is output. Based on the directional recording sequence, a recording graph is established with each monitoring record as a node and adjacent monitoring records as directed edges. For each directed edge, the sum of the number of direction reversals, the number of pressure and flow sign mismatches, and the number of temperature timing mismatches is calculated as the edge cost. Dynamic programming is performed in the order of cumulative edge cost, number of monitoring records, and start acquisition time to segment the data. The resulting continuous node groups are determined as candidate segments, and the candidate segment sequence is output.

[0007] In a preferred embodiment, the execution of the fragment generation module further includes: For each candidate segment, calculate the cumulative pressure offset, cumulative flow offset, and cumulative temperature offset from the first monitoring record to the last monitoring record. Use the forward accumulation result and the backward substitution result as a check. If the cumulative pressure offset sign and the cumulative flow offset sign are the same in the forward accumulation result and the backward substitution result, retain the corresponding candidate segment as an abnormal segment. If the cumulative pressure offset sign or the cumulative flow offset sign is different, take the monitoring record corresponding to the sum of the edge costs of the two segments after splitting as the split point to reconstruct the candidate segment. Continue until the check of each reconstructed segment satisfies the requirement that the cumulative pressure offset sign and the cumulative flow offset sign are the same, and output the abnormal segment sequence.

[0008] In a preferred embodiment, the execution of the identity establishment module includes: Arrange the abnormal segment sequence of the same sealing part in the order of appearance, and calculate the direction consistency value, residual offset difference, duration difference and interval duration value between each subsequent abnormal segment and the previous abnormal segment in turn. When the direction consistency value is one, the residual offset difference and duration difference have the same sign and the interval duration value is not greater than the duration of the previous abnormal segment, candidate connection edges are established. When the direction consistency value is zero, the residual offset difference and duration difference have opposite signs or the interval duration value is greater than the duration of the previous abnormal segment, candidate connection edges are deleted. Output the candidate connection graph. Based on the candidate connection graph, the connection cost is calculated for each candidate connection edge, consisting of the absolute value of the residual offset difference, the absolute value of the duration difference, and the interval duration. Forward accumulation and backward substitution are performed in the order of the occurrence of abnormal segments. When there are multiple candidate connection edges in the same subsequent abnormal segment, the candidate connection edge with the smallest connection cost is selected as the retained connection edge. When the connection costs are the same, the candidate connection edge with the earlier end time of the previous abnormal segment is selected as the retained connection edge, and the remaining candidate connection edges are deleted. The connection chain is then output.

[0009] In a preferred embodiment, the execution of the identity establishment module further includes: For each link segment in the link chain, calculate the number of times the direction is maintained, the number of times the residual offset increases, and the number of times the duration increases between the first and last anomalous segments. When the number of times the direction is maintained is equal to the number of links in the link segment and the sum of the number of times the residual offset increases and the number of times the duration increases is not zero, the anomalous segments contained in the link segment are classified into the same damage identity. When the sum of the number of times the residual offset increases and the number of times the duration increases is zero, the last anomalous segment in the link segment is taken as the end segment of the current damage identity, and the next anomalous segment is taken as the start segment of the next damage identity, and the damage identity sequence is output.

[0010] In a preferred embodiment, the execution of the evolution determination module includes: Arrange the abnormal segments containing the same damage identity in the order of appearance, calculate the residual offset difference, duration difference and segment interval duration between each subsequent abnormal segment and the previous abnormal segment in turn, and record a residual increment when the residual offset difference is positive and the segment interval duration is not greater than the duration of the previous abnormal segment, and record a continuous expansion when the duration difference is positive, and output the evolution count result.

[0011] In a preferred embodiment, the execution of the evolution determination module further includes: Based on the evolution count results, each abnormal segment of the same damage identity is accumulated segment by segment in the order of appearance. When the residual increase occurs once and the continuous expansion occurs once, the damage identity is determined to be in a state of continuous deterioration. When there are no abnormal segments with a negative residual offset difference between two adjacent residual increases and the residual increase occurs twice in a row, the damage identity is determined to be in an irreversible damage state, and the health status result is output. The health status results are combined with the corresponding damage identification, end-abnormal segment identification, residual increment count, and continuous expansion count and written into the health status record, and the health status result sequence corresponding to the same sealing part is output.

[0012] In a preferred embodiment, the execution of the output module includes: The health status results, damage identification, and abnormal fragment identification corresponding to the same sealing part are arranged into a state node sequence according to the order of appearance. For each adjacent state node, the state transition value is obtained by calculating whether the health status result has changed, the identity continuation value is obtained by calculating whether the damage identification is the same, and the fragment reproduction value is obtained by calculating whether the abnormal fragment identification of the next state node appears in the previous state node. The state transition value, identity continuation value, and fragment reproduction value are added together to obtain the edge cost. Then, forward accumulation and backward substitution are performed on the state node sequence, and the state node chain corresponding to the accumulated edge cost is taken to generate the state record. Based on the state record, the health status result, damage identification, abnormal fragment identification, and generation time corresponding to the final state node are extracted and concatenated in the order of the fields to obtain the original record text. A hash mapping is performed on the original record text to obtain the current record fingerprint. The current record fingerprint is concatenated with the record fingerprint of the previous state record to generate a record chain fingerprint. When the health status result of the final state node is a continuously deteriorating state, a continuous deterioration warning message containing the record chain fingerprint is generated. When the health status result of the final state node is an irreversible damage state, an irreversible damage warning message containing the record chain fingerprint is generated.

[0013] In a preferred embodiment, the execution of the output module further includes: The seal location identifier, damage identification identifier, abnormal fragment identifier, and current record fingerprint in the status record are compared item by item with the seal location identifier, damage identification identifier, abnormal fragment identifier, and current record fingerprint in the warning information. If all items are the same, the status record and warning information are output. If any item is different, the status node sequence is retrieved again using the damage identification identifier and abnormal fragment identifier in the warning information. The status record generation, current record fingerprint generation, and item-by-item comparison are repeated until the status record and the warning information are identical item by item.

[0014] The technical effects and advantages of this invention are as follows: By generating abnormal fragments, damage identification, and health status results from monitoring records of each sealing part, dispersed fluctuations are organized into a continuous evolution chain, thus relatively improving the ability to distinguish between transient disturbances and irreversible damage signs under non-stop conditions, and supporting early intervention. By establishing a one-to-one correspondence between the sealing parts and the pressure, flow and temperature acquisition points, and generating monitoring records at a unified acquisition time, the interference of inconsistent acquisition sources and time standards on subsequent judgment results is relatively reduced. By performing change direction generation, record image segmentation, and two-way verification on the monitoring records, candidate segments with unstable directions are eliminated or reconstructed, thus relatively improving the consistency of abnormal segment boundary identification and relatively reducing misjudgment caused by single-point fluctuations. By progressively accumulating and determining the residual offset and duration of abnormal segments within the same damage identity, and combining this with the consistency verification of status records and early warning information, the continuity and traceability of early warning output are relatively improved. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the system modules of the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Refer to the instruction manual appendix Figure 1 The present invention provides a handheld water-cooled sealing status monitoring system for an electromagnetic heating device, comprising: The data acquisition module is used to collect pressure, flow and temperature values ​​corresponding to each sealing part of the handheld terminal, and generate monitoring records according to the time of acquisition; In this specific embodiment, the acquisition module is used to organize the pressure, flow and temperature information corresponding to each sealing part of the handheld terminal into a continuously comparable monitoring record sequence, so that the subsequent abnormal segment generation, damage identification, evolution judgment and early warning output are all based on the same sealing part, the same acquisition time and the same field caliber. The acquisition module first determines the correspondence between each sealing part and the acquisition point according to the internal water cooling structure of the handheld terminal, then reads the measured value of each acquisition point under a unified time reference and calculates the acquisition value of the part, and finally writes the acquisition value of the part into the monitoring record sequence of the corresponding sealing part. Based on the distribution of each sealing component of the handheld terminal in the water supply path, return path, and shell connection position, a one-to-one correspondence is established between each sealing component and its corresponding upstream pressure sampling point, downstream pressure sampling point, flow rate sampling point, and adjacent temperature sampling point. The process of outputting the sealing component sampling correspondence table is as follows: Read the handheld terminal structural configuration data, which includes at least the sealing component identification, sealing component location, water supply path direction, return path direction, and shell connection position; using the flow direction of the coolant within the handheld terminal as the criterion for determining the upstream and downstream directions, for each sealing component, select the nearest pressure sampling point on the inlet side along the flow direction as the upstream pressure sampling point, select the nearest pressure sampling point on the outlet side along the flow direction as the downstream pressure sampling point, and select the flow rate sampling point on the corresponding cooling branch as the flow rate sampling point. The temperature acquisition point closest to the projected position of the sealing part's shell is selected as the adjacent temperature acquisition point. When the same acquisition point falls within the candidate range of two sealing parts, the distance from the acquisition point to the geometric center of the two sealing parts is calculated, and the sealing part with the shorter distance is selected as the corresponding object of the acquisition point. The other sealing part continues to select the next acquisition point along the corresponding direction. When a sealing part lacks any type of acquisition point, the sealing part identifier and the missing measurement point identifier are written into the sealing part acquisition correspondence table, and the sealing part is prohibited from entering the part acquisition value calculation of the current acquisition cycle. The resulting sealing part acquisition correspondence table includes at least the sealing part identifier, upstream pressure acquisition point identifier, downstream pressure acquisition point identifier, flow rate acquisition point identifier, and adjacent temperature acquisition point identifier, and is used for the measurement value reading process at the current acquisition time. At each acquisition moment, the upstream pressure measurement, downstream pressure measurement, flow rate measurement, and adjacent temperature measurement corresponding to the same sealing part are synchronously read according to the sealing part acquisition correspondence table. The difference between the upstream and downstream pressure measurements is determined as the pressure value of the sealing part, the flow rate measurement is determined as the flow rate value of the sealing part, and the adjacent temperature measurement is determined as the temperature value of the sealing part. The process of outputting the acquisition values ​​is as follows: the acquisition controller generates a unified time base according to a fixed acquisition cycle, and the moment corresponding to each unified time base is recorded as an acquisition moment; for each sealing part, the corresponding upstream pressure acquisition point, downstream pressure acquisition point, flow rate acquisition point, and adjacent temperature acquisition point are retrieved according to the sealing part acquisition correspondence table, and the most recent original measurement value before and after each acquisition point is read at the current acquisition moment; when a certain acquisition point has both before and after the current acquisition moment... When taking initial measurements, the initial measurement with the smaller time difference is taken as the measurement at the current acquisition time. When the time differences are the same, the initial measurement at the later time is taken as the measurement at the current acquisition time. Then, the upstream pressure measurement is subtracted from the downstream pressure measurement to obtain the pressure value, the flow rate measurement is taken from the flow rate measurement to obtain the flow rate value, and the adjacent temperature measurement is taken to obtain the temperature value. The pressure value, flow rate value, and temperature value are combined to form the location acquisition value. If any acquisition point has no initial measurement before or after the current acquisition time, the sealing location is marked as a missing measurement record at the current acquisition time, and no location acquisition value is generated. If any acquisition point returns an acquisition failure flag, the sealing location is marked as an invalid acquisition record at the current acquisition time, and the data at that acquisition time is not sent to subsequent change calculations. The output location acquisition value includes at least the sealing location identifier, acquisition time, pressure value, flow rate value, and temperature value, and is available for reading during the monitoring record writing process. The process of combining the collected values ​​of each sealing location with the corresponding sealing location identifier and collection time into the monitoring record, and outputting the monitoring record sequence corresponding to each sealing location, is as follows: For each valid location collected value, the sealing location identifier, collection time, collection cycle number, pressure value, flow rate value, and temperature value are written in a fixed field order to form a monitoring record, and this monitoring record is appended to the end of the monitoring record sequence of the corresponding sealing location; when the current collection time corresponds to a missing record, the sealing location identifier, collection time, and missing identifier are written into the monitoring record sequence of the corresponding sealing location, but the pressure value, flow rate value, and temperature value are not written, and the missing identifier is used as a sequence breakpoint in the subsequent abnormal segment generation; when the current collection time corresponds to an invalid collection record, the sealing location identifier, collection time, and invalid identifier are written into the monitoring record sequence of the corresponding sealing location, but the pressure value, flow rate value, and temperature value are not written, and the invalid collection record is prohibited from participating in the calculation of the change between adjacent monitoring records; thus, the monitoring record sequence corresponding to each sealing location is obtained, and each monitoring record sequence is stored in ascending order of collection time and used as the input sequence read by the segment generation module; Through the above processing, the acquisition module converts the scattered raw measurements into a unified monitoring record sequence for the sealed parts, clarifying the acquisition source, time alignment method, difference calculation direction, missing measurement handling method, and record writing structure for each sealed part. In practical applications: the water supply connector sealing ring, the return water connector sealing ring, and the shell end cap sealing seam are each designated as three sealed parts with independent identifiers. The acquisition module first determines the corresponding upstream pressure acquisition point, downstream pressure acquisition point, flow rate acquisition point, and adjacent temperature acquisition point for each sealed part. Then, it reads the corresponding measurements at the same acquisition time and calculates the corresponding pressure, flow rate, and temperature values. Subsequently, it writes the sealed part identifier, acquisition time, acquisition cycle number, and the three types of measurements into the corresponding monitoring record sequence. If the adjacent temperature acquisition point corresponding to the shell end cap sealing seam has no raw measurements before or after the current acquisition time, only the sealed part identifier, acquisition time, and missing measurement identifier are written for that acquisition time, and it is treated as a sequence breakpoint in subsequent processing.

[0018] The segment generation module is used to calculate the change direction of adjacent records of the same sealing part according to the acquisition time sequence, merge the monitoring records with continuous and consistent change directions into abnormal segments, and output the abnormal segment sequence. In this specific embodiment, the fragment generation module is used to convert the monitoring record sequence corresponding to the same sealing part into an abnormal fragment sequence that can be used for damage identification. Its function is to first eliminate the interference of single-point fluctuations and sampling jitter on the determination of the direction of change, then determine the candidate fragment boundaries based on the correlation between pressure changes, flow rate changes, and temperature changes at the same sealing part during continuous acquisition, and finally eliminate candidate fragments with unstable directions through bidirectional verification, so that the abnormal fragments entering the subsequent damage identification process have consistent directional attributes and continuous offset attributes. The fragment generation module takes the monitoring record sequence of the same sealing part as input, first generates a directional record sequence, then constructs a candidate fragment sequence based on the directional record sequence, and finally performs verification and reconstruction on the candidate fragment sequence to output an abnormal fragment sequence. This implementation process includes the following steps: Monitoring records for the same sealing location are arranged in ascending order of acquisition time. The pressure, flow, and temperature changes between adjacent records are calculated. The pressure, flow, and temperature change sequences are then subjected to three-term convolution smoothing and three-term median reconstruction. A direction vector is formed using the reconstructed signs of these three changes. When the three signs of the direction vector are inconsistent, the sign combination corresponding to the sum of the absolute values ​​of the pressure, flow, and temperature residuals is taken as the direction of change for the current monitoring record. The process of outputting the direction record sequence is as follows: The monitoring record sequence corresponding to the same sealing location is read. Records with invalid markers are first removed. Then, the location of the missing measurement marker is used as a breakpoint, and the monitoring records on both sides of the breakpoint are retained in independent continuous subsequences. Each continuous subsequence is arranged in ascending order of acquisition time. The pressure, flow, and temperature changes between adjacent monitoring records are calculated sequentially by subtracting the corresponding value of the previous monitoring record from the corresponding value of the next monitoring record, thus forming pressure change sequences, flow change sequences, and temperature change sequences. Then, a three-term convolution smoothing process is performed on each change sequence in the order of acquisition time. The convolution window consists of the current term, the previous term, and the next term. The convolution coefficients use a fixed 1, 2, 1 ratio and are normalized according to the four-term sum. If the first term is missing a previous term, it is supplemented by the first term itself; if the last term is missing a next term, it is supplemented by the last term itself, resulting in a smoothed sequence. Next, a three-term median reconstruction is performed on each smoothed sequence. The median window consists of the current smoothed value, the previous smoothed value, and the next smoothed value. The first and last positions are still filled using the self-supplementation method, resulting in reconstructed pressure change sequences, reconstructed flow change sequences, and reconstructed temperature change sequences. Subsequently, at each acquisition moment, the reconstructed pressure change symbol, reconstructed flow rate change symbol, and reconstructed temperature change symbol corresponding to that moment are extracted to form a direction vector. When the three symbols are the same, the common symbol is directly determined as the change direction of the current monitoring record. When the three symbols are inconsistent, three candidate symbol combinations for that acquisition moment are enumerated. The three candidate symbol combinations are respectively taken from the reconstructed pressure change symbol, reconstructed flow rate change symbol, and reconstructed temperature change symbol as the unified change direction, and the absolute values ​​of the pressure residual, flow rate residual, and temperature residual are calculated respectively. Among them, the absolute value of the pressure residual is the current reconstructed... The absolute difference between the pressure change after reconstruction and the pressure change after determining the direction according to the candidate symbols is calculated. The absolute value of the flow residual is the absolute difference between the current reconstructed flow change and the flow change after determining the direction according to the candidate symbols. The absolute value of the temperature residual is the absolute difference between the current reconstructed temperature change and the temperature change after determining the direction according to the candidate symbols. These three values ​​are then added together to obtain the sum of residuals. The candidate symbol combination with the smallest sum of residuals is taken as the change direction of the current monitoring record. When the sum of residuals corresponding to two candidate symbol combinations is the same, the candidate symbol combination with the same change direction as the previous monitoring record is taken first. If the previous monitoring record does not exist, the candidate symbol combination corresponding to the pressure change symbol is taken. This generates a direction recording sequence, which includes at least the seal location identifier, acquisition time, and direction of change. The sequence is written in the order of acquisition time for the candidate segment generation process to read. If a continuous subsequence contains only one monitoring record, the monitoring record does not generate a direction of change and is directly written as a pending single-point record into the single-point record set, and does not participate in the current round of candidate segment generation. Based on the directional recording sequence, a recording graph is constructed with each monitoring record as a node and adjacent monitoring records as directed edges. For each directed edge, the sum of the number of direction reversals, the number of pressure / flow sign mismatches, and the number of temperature / time mismatches is calculated as the edge cost. Dynamic programming is performed in the order of cumulative edge cost, number of monitoring records, and start acquisition time to segment the data. The resulting continuous node groups are determined as candidate segments. The process of outputting the candidate segment sequence is as follows: Read the directional recording sequence corresponding to the same sealing part, take the monitoring record corresponding to each directional record as a node, and establish a directed edge only between the preceding and following nodes adjacent to the acquisition time to form the recording graph; for each directed edge, first compare the change direction of the preceding node with the change direction of the following node. If the directions are the same, the number of direction reversals is recorded as zero; if the directions are different, the number of direction reversals is recorded as one; then compare the sign of the pressure change and the flow change corresponding to the following node. The system first identifies the symbols for pressure and flow rate. If the symbols are the same, the number of pressure and flow rate symbol mismatches is recorded as zero; if the symbols are different, the number of pressure and flow rate symbol mismatches is recorded as one. Then, it determines the number of temperature and time-series mismatches. The rule is: if the symbol of the temperature change corresponding to the next node is the same as the symbol of the pressure change corresponding to the next node, or if the symbol of the temperature change corresponding to the next node is the same as the symbol of the pressure change corresponding to the previous node, the number of temperature and time-series mismatches is recorded as zero; otherwise, it is recorded as one. The number of direction reversals, the number of pressure and flow rate symbol mismatches, and the number of temperature and time-series mismatches are then added together to obtain the edge cost of the directed edge. Next, dynamic programming is performed in the order of data acquisition time. For each node, the cumulative edge cost, the number of nodes, and the starting acquisition time are calculated. The cumulative edge cost is obtained by adding the cumulative edge cost of the predecessor node to the edge cost of the current directed edge. The number of nodes is obtained by adding one to the number of nodes of the predecessor node. The starting acquisition time is inherited from the predecessor node. When multiple predecessor paths exist for the same node, the path with the smaller cumulative edge cost is selected first. If the cumulative edge costs are the same, the path with more nodes is selected. If the number of nodes is the same, the path with the earlier start time is selected. If the start times are the same, the path with the earlier end time is selected. This determines the optimal predecessor path for the node. When a node does not have a predecessor path that satisfies the continuous connection condition, the node is used as the starting point of a new path. After calculating the optimal path for all nodes, starting from the end node, the optimal predecessor path is used to backtrack and extract continuous node groups from the record graph. Each continuous node group constitutes a candidate segment, and the acquisition times of the first and last monitoring records in the node group are written into the candidate segment sequence. If a candidate segment contains only one monitoring record, the candidate segment is marked as a single-point candidate segment and retained in the candidate segment sequence for subsequent verification to determine whether to delete it. For each candidate segment, calculate the cumulative pressure offset, cumulative flow offset, and cumulative temperature offset from the first monitoring record to the last monitoring record. Use the forward accumulation result and the backward substitution result as a checksum. If the cumulative pressure offset sign and the cumulative flow offset sign are the same in the forward accumulation result and the backward substitution result, retain the corresponding candidate segment as an abnormal segment. If the cumulative pressure offset sign or the cumulative flow offset sign is different, take the monitoring record corresponding to the sum of the edge costs of the two segments after splitting as the split point to reconstruct the candidate segment. Continue until the checksum of each reconstructed segment satisfies the requirement that the cumulative pressure offset sign and the cumulative flow offset sign are the same. The process of outputting the abnormal segment sequence is as follows: Read the candidate segment sequence corresponding to the same sealing part. For each candidate segment, extract its first monitoring record, last monitoring record, and all monitoring records within the segment. Accumulate the pressure change, flow change, and temperature change line by line in the order from the first monitoring record to the last monitoring record to obtain the cumulative pressure offset, cumulative flow offset, and cumulative temperature offset in the forward accumulation result. Then, calculate the cumulative pressure offset, cumulative flow offset, and cumulative temperature offset line by line in the order from the last monitoring record to the first monitoring record. The same set of changes is replaced sequentially in reverse order to obtain the cumulative pressure offset, cumulative flow offset, and cumulative temperature offset in the backward replacement result. The forward accumulation result and the backward replacement result together constitute the verification of the candidate segment. If the cumulative pressure offset sign and the cumulative flow offset sign are the same in the forward accumulation result and the backward replacement result, the candidate segment is retained as an abnormal segment, and the sealing part identifier, segment start acquisition time, segment end acquisition time, change direction, cumulative pressure offset, cumulative flow offset, and cumulative temperature offset of the abnormal segment are written into the abnormal segment sequence. If the cumulative pressure offset sign or the cumulative flow offset sign is different in the forward accumulation result and the backward replacement result, each monitoring record in the candidate segment except for the first and last monitoring records is used as a splitting point for trial splitting. The sum of the edge costs of the first segment and the sum of the edge costs of the second segment are calculated after splitting, and the monitoring record with the smallest sum of the two edge costs is taken as the splitting point. When the sum of the two edge costs corresponding to multiple monitoring records is the same, the monitoring record with the earlier acquisition time is taken as the splitting point. The original candidate segments are split into front-end candidate segments and back-end candidate segments according to the determined split points. Then, forward accumulation, backward substitution, and sign verification are repeatedly performed on the front-end candidate segments and back-end candidate segments respectively until all the reconstructed segments meet the condition that the cumulative pressure offset sign and the cumulative flow offset sign of the forward accumulation result and the backward substitution result are the same. When only a single monitoring record remains after the candidate segment is split and still does not meet the verification conditions, the segment corresponding to the single monitoring record is deleted and not written into the abnormal segment sequence. The abnormal segment sequence output in this way is stored in ascending order according to the segment start acquisition time and is used as the input sequence of the identity establishment module. Through the above processing, the fragment generation module transforms the monitoring record sequence corresponding to the same sealing part into an abnormal fragment sequence with clear boundaries, consistent direction, and continuous offset. This completes the calculation of change direction values, temperature following relationships, candidate fragment formation rules, reconstruction and splitting rules, and termination boundaries, enabling the subsequent damage identification process to be based on stable fragments rather than the original fluctuation records. In practical applications: for the monitoring record sequence corresponding to the water supply connector sealing ring, the pressure change, flow rate change, and temperature change are first calculated according to the acquisition time. Then, the influence of pressure fluctuations at a single acquisition time is eliminated through three-term convolution smoothing and three-term median reconstruction. If the pressure change, flow rate change, and temperature change at a certain acquisition time are positive, then they are respectively marked with a positive sign. Positive and negative values ​​are used as candidate unified change directions to calculate the residual sum, and the candidate unified change direction with the smallest residual sum is taken as the change direction at that acquisition time. Then, a record graph is built and the edge cost of each directed edge is calculated to obtain a candidate segment covering six monitoring records. When performing bidirectional verification on the candidate segment, if the pressure cumulative offset sign of the forward accumulation result is positive and the pressure cumulative offset sign of the backward substitution result is negative, then the sum of the edge costs of the two segments after splitting is calculated sequentially using the middle monitoring record as the splitting point. The splitting point corresponding to the sum of the edge costs of the two segments is selected to split the original candidate segment into two reconstructed segments until the two reconstructed segments satisfy the same pressure cumulative offset sign and the same flow cumulative offset sign. Then, the two reconstructed segments are written into the abnormal segment sequence.

[0019] The identity establishment module is used to compare the abnormal segment sequences of the same sealing part in the order of appearance. When the change direction of the later abnormal segment is consistent with that of the previous abnormal segment and the same offset is still retained after the end, the latter two are classified into the same damage identity and the damage identity sequence is output. In this specific embodiment, the identity establishment module is used to further organize abnormal fragments appearing sequentially under the same sealing location into damage identities with continuous evolutionary relationships. This allows subsequent evolutionary judgments to be made based on the continuation of the same damage process across multiple operating cycles, rather than isolated abnormal fragments. The identity establishment module first establishes candidate connection edges based on the directional relationship, residual offset relationship, duration relationship, and occurrence interval relationship between abnormal fragments. Then, it retains the unique preceding connection edge corresponding to each subsequent abnormal fragment from the candidate connection edges to form a connection chain. Finally, it merges the continuous connection chain segments in the connection chain and outputs the damage identity sequence. This implementation process includes the following steps: The abnormal segment sequence of the same sealing part is arranged in order of appearance. The direction consistency value, residual offset difference, duration difference, and interval duration value between each subsequent abnormal segment and the preceding abnormal segment are calculated sequentially. Candidate connection edges are established when the direction consistency value is one, the residual offset difference and duration difference have the same sign, and the interval duration value is not greater than the duration of the preceding abnormal segment. Candidate connection edges are deleted when the direction consistency value is zero, the residual offset difference and duration difference have opposite signs, or the interval duration value is greater than the duration of the preceding abnormal segment. The process of outputting the candidate connection graph is as follows: Read the abnormal segments corresponding to the same sealing part. The sequence is arranged in ascending order of the start time of the abnormal segment. The duration of the abnormal segment is the difference between the end time of the abnormal segment and the start time. The residual offset is the difference between the corresponding value of the last monitoring record and the corresponding value of the first monitoring record of the abnormal segment along the direction of change of the abnormal segment. The interval is the difference between the start time of the next abnormal segment and the end time of the previous abnormal segment. Then, two adjacent abnormal segments are grouped together, and the direction of change of the next abnormal segment is compared with the direction of change of the previous abnormal segment. If the directions are the same, the value of consistent direction is recorded as one; if the directions are different, the value of consistent direction is recorded as zero. Next, calculate the residual offset difference and duration difference using the method of "subtracting the corresponding value of the previous abnormal segment from the corresponding value of the subsequent abnormal segment," and determine the sign relationship between the residual offset difference and duration difference. If both are positive or both are negative, they are recorded as having the same sign; if one is positive and the other is negative, they are recorded as having opposite signs; if either difference is zero, zero is taken as having the same sign as the other difference. Then, compare the interval duration value with the duration of the previous abnormal segment. If the interval duration value is not greater than the duration of the previous abnormal segment, the interval continuity is retained; if the interval duration value is greater than the duration of the previous abnormal segment, the interval continuity is deleted. When the direction is consistent (value is one), the residual offset difference and duration difference have the same sign, and the interval... When the duration value is not greater than the duration of the previous abnormal segment, a candidate connection edge is established between the previous and subsequent abnormal segments, and the identifier of the previous abnormal segment, the identifier of the subsequent abnormal segment, the direction consistency value, the residual offset difference value, the duration difference value, and the interval duration value are written into the candidate connection graph; when the direction consistency value is zero, or the residual offset difference value and the duration difference value have opposite signs, or the interval duration value is greater than the duration of the previous abnormal segment, no candidate connection edge is established; if there is only one abnormal segment in the same sealing part, no candidate connection edge is generated, and the abnormal segment is marked as an independent segment and directly written into the independent segment set for subsequent damage identification merging process to read. Based on the candidate connection graph, the connection cost for each candidate connection edge is calculated, consisting of the absolute value of the residual offset difference, the absolute value of the duration difference, and the interval duration. Forward accumulation and backward substitution are performed in the order of the occurrence of the abnormal segments. When multiple candidate connection edges exist for the same subsequent abnormal segment, the candidate connection edge with the smallest connection cost is selected as the retained connection edge. When the connection costs are the same, the candidate connection edge with the earlier end time of the previous abnormal segment is selected as the retained connection edge, and the remaining candidate connection edges are deleted. The process of outputting the connection chain is as follows: Read all candidate connection edges in the candidate connection graph; for each candidate connection edge, calculate the absolute value of the residual offset difference, the absolute value of the duration difference, and the interval duration, and add them together to obtain the connection cost of the candidate connection edge; then perform forward accumulation in ascending order of the start time of the subsequent abnormal segment. For each subsequent abnormal segment, all candidate connection edges pointing to that segment are read. The corresponding connection cost is added to the cumulative connection cost of the preceding abnormal segment to obtain the cumulative connection cost of the current path. When there are multiple candidate connection edges for the same subsequent abnormal segment, the candidate connection edge with the smaller cumulative connection cost of the current path is selected as the reserved connection edge. If the cumulative connection costs of the current path are the same, the candidate connection edge with the earlier end time of the preceding abnormal segment is selected as the reserved connection edge. If the end time of the preceding abnormal segment is still the same, the candidate connection edge with the earlier start time of the preceding abnormal segment is selected as the reserved connection edge. After completing the forward accumulation, the backward replacement is performed starting from the last abnormal segment with a reserved connection edge. The preceding abnormal segments are traced back sequentially according to the reserved connection edges, and the connection relationship of the abnormal segments obtained by the continuous backtracking is written into the connection chain. In this process, each connection chain retains only one reserved connection edge pointing to the same subsequent abnormal segment, and all other candidate connection edges are deleted. If a subsequent abnormal segment does not have a candidate connection edge, then the subsequent abnormal segment is written into the connection chain as the starting point of the new connection chain. If an abnormal segment has both a predecessor reserved connection edge and a successor reserved connection edge, then the abnormal segment is retained as an intermediate segment in the connection chain to preserve its bidirectional connection relationship. For each link segment in the link chain, the number of times the direction is maintained, the number of times the residual offset increases, and the number of times the duration increases are calculated sequentially between the first and last anomalous segments. When the number of times the direction is maintained equals the number of connected edges in the link segment and the sum of the number of times the residual offset increases and the number of times the duration increases is not zero, the anomalous segments contained in that link segment are classified into the same damage identity. When the sum of the number of times the residual offset increases and the number of times the duration increases is zero, the last anomalous segment in the link segment is taken as the end segment of the current damage identity, and the next anomalous segment is taken as the start segment of the next damage identity. The process of outputting the damage identity sequence is as follows: read the link chain and retain the links that are continuously connected end to end. A set of abnormal segments formed by edges is defined as a connection chain segment. The first abnormal segment in a connection chain segment is the one without a preceding retained connection edge, and the last abnormal segment is the one without a successor retained connection edge. The number of connection edges is the number of retained connection edges within the connection chain segment. Then, for each connection chain segment, adjacent abnormal segments are compared one by one in the order of their appearance. If the change direction of the preceding and following abnormal segments is the same, the direction retention count is incremented by one. When the residual offset of the following abnormal segment is greater than the residual offset of the preceding abnormal segment, the residual offset increment count is incremented by one. When the duration of the following abnormal segment is greater than the duration of the preceding abnormal segment, the duration increment count is incremented by one. After all comparisons are completed, if the number of direction maintenance counts equals the number of connected edges in the connected chain segment, and the sum of the residual offset increment count and the duration increment count is not zero, then all abnormal segments within the connected chain segment are classified into the same damage identity, and a damage identity identifier is generated according to the start acquisition time of the first abnormal segment of the connected chain segment. The damage identity identifier, the first abnormal segment identifier, the last abnormal segment identifier, and the identifiers of the included abnormal segments are written into the damage identity sequence in sequence. If the sum of the residual offset increment count and the duration increment count is zero, then the last abnormal segment in the connected chain segment is taken as the end segment of the current damage identity, and the... After connecting the segments, the first unassigned abnormal segment that appears immediately in the order of appearance is taken as the starting segment of the next damage identity. Two damage identity identifiers are generated and written into the damage identity sequence. If a connecting segment contains only one abnormal segment, the number of connecting edges of the connecting segment is recorded as zero, the number of direction retention times is recorded as zero, and the single abnormal segment is directly assigned to an independent damage identity. The abnormal segment is written into the damage identity sequence as both the first and last abnormal segment of the independent damage identity. The damage identity sequence output in this way is stored in ascending order according to the starting acquisition time of the first abnormal segment of each damage identity and is used as the input sequence of the evolution determination module. Through the above processing, the identity establishment module merges multiple scattered abnormal fragments under the same sealed part into damage identities with continuous evolution relationship, and clarifies the value and formation boundary of direction consistency value, residual offset difference value, duration difference value, interval duration value, connection cost, connection chain segment and independent damage identity, so as to avoid the subsequent evolution judgment directly facing isolated fragments and causing the judgment object to drift. In practical applications: In the sequence of abnormal segments corresponding to the same sealing part, if the first abnormal segment and the second abnormal segment change in the same direction, the residual offset of the second abnormal segment is higher than that of the first abnormal segment, the duration of the second abnormal segment is longer than that of the first abnormal segment, and the interval between the start acquisition time of the second abnormal segment and the end acquisition time of the first abnormal segment is not greater than the duration of the first abnormal segment, then a candidate connection edge is established between them. If there is another candidate connection edge from the first abnormal segment between the second abnormal segment and the third abnormal segment, then the connection cost of the two candidate connection edges is calculated respectively, and the candidate connection edge with the smaller connection cost is retained to form a connection chain. In this connection chain segment, if the number of times the direction is maintained is equal to the number of connection edges, and the sum of the number of times the residual offset increases and the number of times the duration increases is not zero, then the three abnormal segments in this connection chain segment are classified into the same damage identity, and this damage identity is written into the damage identity sequence for the evolution determination module to continue reading.

[0020] The evolution determination module is used to compare each abnormal segment contained in the same damage identity in the order of appearance. When the residual offset of the later abnormal segment is greater than the residual offset of the previous abnormal segment, the damage identity is determined to be in a state of continuous deterioration. When the residual offset increases continuously, the damage identity is determined to be in an irreversible damage state, and the health status result is output. In this specific embodiment, the evolution determination module is used to continuously determine the sequential changes of abnormal segments within the same damage identity. It further classifies the established damage identities into three categories: non-deteriorated, continuously deteriorating, and irreversible damage. This ensures that the output module reads not just a single abnormal segment, but a damage identity with a clear evolutionary stage. The evolution determination module first counts the residual offset, duration, and segment interval relationships between adjacent abnormal segments within the same damage identity segment by segment. Then, based on the counting results, it determines the health status result. Finally, it writes the determination result back as a health status record and organizes it into a health status result sequence. This implementation process includes the following steps: The abnormal segments contained in the same damage identity are arranged in the order of appearance. The residual offset difference, duration difference and segment interval duration between each subsequent abnormal segment and the previous abnormal segment are calculated in turn. When the residual offset difference is positive and the segment interval duration is not greater than the duration of the previous abnormal segment, a residual increment is recorded. When the duration difference is positive, a continuous expansion is recorded. The process of outputting the evolution count result is as follows: read the abnormal segment sequence corresponding to the same damage identity and arrange them in ascending order according to the start time of the abnormal segment acquisition. The residual offset difference is calculated as "the residual offset of the next abnormal segment minus the residual offset of the previous abnormal segment", the duration difference is calculated as "the duration of the next abnormal segment minus the duration of the previous abnormal segment", and the segment interval duration is calculated as "the start acquisition time of the next abnormal segment minus the end acquisition time of the previous abnormal segment". Then, adjacent abnormal segments are compared sequentially. When the residual offset difference is positive and the segment interval duration is not greater than the duration of the previous abnormal segment, a residual increment is recorded. When the duration difference is positive, a continuous expansion is recorded. The corresponding previous abnormal segment identifier, next abnormal segment identifier, residual offset difference, duration difference, segment interval duration, residual increment identifier, and continuous expansion identifier are written into the evolution count result. When the same damage identity contains only one abnormal segment, adjacent comparisons are not performed; the residual increment count and continuous expansion count are directly recorded as zero and written into the evolution count result. Based on the evolution count results, each abnormal segment of the same damage identity is cumulatively accumulated in the order of appearance. When a residual increase occurs once and a continuous expansion occurs once, the damage identity is determined to be in a state of continuous deterioration. When there are no abnormal segments with a negative residual offset difference between two adjacent residual increases and the residual increase occurs twice consecutively, the damage identity is determined to be in an irreversible damage state. The process of outputting the health status result is as follows: read the evolution count result corresponding to the same damage identity, and cumulatively accumulate the residual increase marker and the continuous expansion marker in the order of appearance of the abnormal segments to obtain the number of residual increases and the number of continuous expansions corresponding to the current damage identity; when the number of residual increases is greater than or equal to one and the number of continuous expansions is... If the number is greater than or equal to one, the damage is classified as a continuously deteriorating state. Based on this, the system continues to check for consecutive occurrences of residual increment markers. If there are two consecutive residual increments, and the residual offset difference between adjacent abnormal segments from the next abnormal segment corresponding to the previous residual increment to the previous abnormal segment corresponding to the next residual increment is not negative, then the damage is classified as an irreversible damage state. If neither of these conditions is met, the damage is classified as a non-deteriorating state. When the damage is initially classified as continuously deteriorating and then becomes irreversible, the irreversible damage state overrides the continuously deteriorating state, and the final classification result is output as the health status result of the damage. The process of combining the health status result with the corresponding damage identity identifier, the last abnormal segment identifier, the residual increment count, and the continuous expansion count into the health status record and outputting the health status result sequence corresponding to the same sealed part is as follows: Read the health status result and its corresponding damage identity identifier, extract the last abnormal segment in the damage identity arranged in the order of appearance as the last abnormal segment, and count the residual increment count and the continuous expansion count corresponding to the damage identity; then write the sealed part identifier, damage identity identifier, last abnormal segment identifier, health status result, residual increment count, and continuous expansion count in a fixed field order to generate a health status record, and append this health status record to the end of the health status result sequence of the corresponding sealed part; when there are multiple damage identities for the same sealed part, write multiple health status records in ascending order according to the starting time of the first abnormal segment of each damage identity to form the health status result sequence corresponding to the sealed part; when a certain damage identity does not generate a valid evolution count result, still write the damage identity identifier and the last abnormal segment identifier, record the health status result as undeteriorated, and record the residual increment count and the continuous expansion count as zero; Through the above processing, the evolution determination module transforms the changing relationships of each abnormal segment within the damaged identity into a health status result that can be directly used for output and early warning. It clarifies the value selection and connection methods of residual offset difference, duration difference, segment interval duration, residual increment count, continuous expansion count, and state coverage relationship, thereby relatively avoiding the problem of unclear health status determination objects. In practical applications: If a certain damage identity of the same sealing part contains three abnormal segments, and the residual offset difference between the second abnormal segment and the first abnormal segment is positive and the segment interval is not greater than the duration of the first abnormal segment, and the duration difference is also positive, then one residual increase and one continuous expansion are recorded; if the residual offset difference between the third abnormal segment and the second abnormal segment is again positive, and there are no abnormal segments with negative residual offset differences between the second residual increase and the first residual increase, then the damage identity is determined to be an irreversible damage state, and the corresponding damage identity identifier, the third abnormal segment identifier, the number of residual increases and the number of continuous expansions are written into the health status record, and then written into the health status result sequence corresponding to the sealing part.

[0021] The output module is used to generate a status record of the corresponding sealing part based on the health status result, and output a warning message when the health status result is a state of continuous deterioration or irreversible damage. In this specific embodiment, the output module is used to further organize the health status results obtained under the same sealing location into verifiable and traceable status records, and generate corresponding warning information when the warning conditions are met, so that the monitoring chain extends from internal judgment results to external output results. The output module first constructs a status node sequence based on the health status results, damage identification, and abnormal segment identification, and then filters out the status node chain for output. Then, it generates status records and record chain fingerprints based on the status node chain. Finally, it verifies the status records and warning information item by item to ensure that the output content is consistent. This implementation process includes the following steps: The process of generating state records by arranging the health status results, damage identification, and abnormal fragment identification corresponding to the same sealing part into a state node sequence according to their order of appearance is as follows: For each adjacent state node, the state transition value is calculated based on whether the health status result has changed; the status continuation value is calculated based on whether the damage identification is the same; and the fragment reproduction value is calculated based on whether the abnormal fragment identification of the subsequent state node appears in the preceding state node. The edge cost is obtained by adding the state transition value, status continuation value, and fragment reproduction value. Then, forward accumulation and backward substitution are performed on the state node sequence, and the state record is generated by taking the state node chain corresponding to the accumulated edge cost. The process is as follows: Read the health status result sequence corresponding to the same sealing part, and increment the state node sequence according to the end time of the last abnormal fragment corresponding to each health status record. The state nodes are sequentially arranged, and the health status result, damage identification, and last abnormal segment identification of each health status record are extracted to form a state node, thus forming a state node sequence. Then, each adjacent state node is compared in turn. If the health status results of the two adjacent state nodes are different, the state transition value is recorded as 1; if they are the same, the state transition value is recorded as 0. If the damage identification of the two adjacent state nodes is the same, the identity continuation value is recorded as 1; if they are different, the identity continuation value is recorded as 0. If there is an abnormal segment identification that is the same as the next state node among all the state nodes before the previous state node, the segment reproduction value is recorded as 1; if not, the segment reproduction value is recorded as 0. Finally, the state transition value, identity continuation value, and segment reproduction value are added together to obtain the edge cost between the adjacent state nodes. Then, forward accumulation is performed according to the order in which state nodes appear. For each state node, the cumulative edge cost from the first state node to the current state node is calculated. If there are multiple paths to reach it, the path with the smaller cumulative edge cost is selected. If the cumulative edge costs are the same, the path with more state nodes is selected. If the number of state nodes is the same, the path with the earlier end time of the last abnormal segment corresponding to the starting state node is selected. After the forward accumulation is completed, backward substitution is performed starting from the last state node. The state node chain is obtained by backtracking in reverse order according to the retained paths. The sealed part identifier, damage identity identifier, abnormal segment identifier, health status result and generation time are written into the state record according to the fixed field order. When there is only one state node for the same sealed part, the state record is directly generated with that state node and the edge cost is not calculated. Based on the state record, the health status result, damage identification, abnormal fragment identification, and generation time corresponding to the last state node are extracted. These are then concatenated in field order to obtain the original record text. A hash mapping is performed on the original record text to obtain the current record fingerprint. The current record fingerprint is then concatenated with the record fingerprint of the previous state record to generate a record chain fingerprint. When the health status result of the last state node is in a continuously deteriorating state, a continuous deterioration warning message containing the record chain fingerprint is generated. When the health status result of the last state node is in an irreversible damage state, an irreversible damage warning message containing the record chain fingerprint is generated. The process is as follows: Read the currently generated state record, extract the health status result, damage identification, abnormal fragment identification, and generation time corresponding to the last state node. The record text is generated by concatenating the identity identifier, abnormal fragment identifier, and generation time in the order of sealed part identifier, damage identity identifier, abnormal fragment identifier, health status result, and generation time, with fixed separators inserted between each field. Then, a hash mapping is performed on the record text to obtain the current record fingerprint. Next, the most recently generated status record before the current status record is retrieved in reverse order of generation time from the historical status records of the same sealed part. When the previous status record is found, the record fingerprint in the previous status record is read and concatenated with the beginning and end of the current record fingerprint to generate a record chain fingerprint. When the previous status record is not found, the current record fingerprint is used as the record chain fingerprint. Next, the health status result corresponding to the final state node is determined. When the health status result is a continuously deteriorating state, the sealed part identifier, damage identification identifier, abnormal fragment identifier, current record fingerprint, record chain fingerprint, and continuously deteriorating state identifier are written in a fixed field order to generate a continuously deteriorating warning message. When the health status result is an irreversible damage state, the sealed part identifier, damage identification identifier, abnormal fragment identifier, current record fingerprint, record chain fingerprint, and irreversible damage state identifier are written in a fixed field order to generate an irreversible damage warning message. When the health status result of the final state node is a non-deteriorated state, only the status record is output, and no warning message is generated. The process of comparing the sealed part identifier, damage identification identifier, abnormal fragment identifier, and current record fingerprint in the status record with the sealed part identifier, damage identification identifier, abnormal fragment identifier, and current record fingerprint in the warning information item by item, and outputting the status record and warning information when all items are the same, and re-searching the status node sequence using the damage identification identifier and abnormal fragment identifier in the warning information when any item is different, and repeating the process of generating the status record, generating the current record fingerprint, and comparing item by item until the status record and the warning information are identical item by item, is as follows: read the status record and the warning information, extract the sealed part identifier, damage identification identifier, abnormal fragment identifier, and current record fingerprint in both in sequence, and compare them item by item according to the same field order; when all four items are the same, write the status record to the status record output queue and write the warning information to the warning output queue. When any field is different, the damage identification and abnormal fragment identification in the warning information are used as search keys to re-search the state node corresponding to the damage identification and abnormal fragment identification in the state node sequence corresponding to the same sealed part. If multiple state nodes are found, the state node with the later generation time is taken as the target state node. If no state node is found, the current warning information is deleted and only the state record is retained. After the target state node is found, the state node chain generation, state record generation, current record fingerprint generation and item-by-item comparison are re-executed until the regenerated state record is identical to the corresponding field in the warning information item by item, and then the output is executed. When the state record does not correspond to the warning information, no content is written to the warning output queue. Through the above processing, the output module converts the health status results into status records and early warning information with chain fingerprint verification relationships. This clarifies the basis for generating the status node chain, the source of the record fingerprint and record chain fingerprint, the composition of the early warning information, and the consistency verification method between the status records and early warning information, avoiding a disconnect between the output content and the internal judgment results. In practical applications: if three consecutive health status records are generated for a certain sealed area, and the health status result corresponding to the third health status record is an irreversible damage state, the output module first constructs a status node sequence and generates a status node chain based on the three health status records. Then, it concatenates the sealing area identifier, damage identification identifier, abnormal fragment identifier, health status result, and generation time from the third status record to form the original record text, calculates the current record fingerprint, and concatenates it end-to-end with the record fingerprint from the second status record to form a record chain fingerprint. Subsequently, it generates irreversible damage early warning information containing this record chain fingerprint. If the abnormal fragment identifier in the early warning information is inconsistent with the abnormal fragment identifier in the status record, the corresponding status node is retrieved again and the status record is reconstructed until the early warning information and the status record are identical item by item before outputting.

[0022] Working principle: This solution first synchronously collects pressure, flow, and temperature values ​​corresponding to each sealing part of the handheld terminal, forming monitoring records arranged according to the collection time. Then, it generates abnormal segments based on the changing direction of adjacent monitoring records to separate continuous offsets from single-point fluctuations. Subsequently, related abnormal segments are merged into the same damage identity to describe the continuity of the same sealing problem at different times. On this basis, the residual offset and duration changes of each abnormal segment within the same damage identity are compared to determine whether the damage identity is in a state of continuous deterioration or irreversible damage. Finally, the determination results are generated into a status record, and warning information is output when a state of continuous deterioration or irreversible damage occurs. The links between each step are clear, with the output of the previous step serving as the input of the next step, transforming the raw collected data layer by layer into sealing health results that can be directly used for maintenance decisions. In practical applications, when the handheld end of the electromagnetic heating device is continuously operating, new monitoring records will be continuously generated at the sealing parts of the water supply connector, the water return connector, and the housing joint. When a certain sealing part exhibits increased pressure difference, abnormal flow rate changes, and temperature response changes at multiple consecutive sampling times, the system first identifies this record as an abnormal segment. If the sealing part subsequently exhibits abnormal segments with the same direction and residual offset, the system will classify these abnormal segments into the same damage type and then determine whether the damage type continues to worsen. When the judgment result reaches a state of continuous deterioration, the system outputs a corresponding warning, indicating that the sealing part has entered a stage requiring treatment. When the judgment result further reaches an irreversible damage state, the system outputs a higher-level warning to guide maintenance or replacement, thereby avoiding passive handling after obvious leakage.

[0023] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A handheld water-cooled sealing status monitoring system for an electromagnetic heating device, characterized in that, include: The data acquisition module is used to collect pressure, flow and temperature values ​​corresponding to each sealing part of the handheld terminal, and generate monitoring records according to the time of acquisition; The segment generation module is used to calculate the change direction of adjacent records of the same sealing part according to the acquisition time sequence, merge the monitoring records with continuous and consistent change directions into abnormal segments, and output the abnormal segment sequence. The identity establishment module is used to compare the abnormal segment sequences of the same sealing part in the order of appearance. When the change direction of the later abnormal segment is consistent with that of the previous abnormal segment and the same offset is still retained after the end, the latter two are classified into the same damage identity and the damage identity sequence is output. The evolution determination module is used to compare each abnormal segment contained in the same damage identity in the order of appearance. When the residual offset of the later abnormal segment is greater than the residual offset of the previous abnormal segment, the damage identity is determined to be in a state of continuous deterioration. When the residual offset increases continuously, the damage identity is determined to be in an irreversible damage state, and the health status result is output. The output module is used to generate a status record of the corresponding sealing part based on the health status result, and to output a warning message when the health status result is a state of continuous deterioration or irreversible damage.

2. The handheld water-cooled sealing status monitoring system for an electromagnetic heating device according to claim 1, characterized in that: The execution of the acquisition module includes: Based on the distribution of each sealing part of the handheld terminal in the water supply passage, water return passage and shell connection position, establish a one-to-one correspondence between each sealing part and the corresponding upstream pressure acquisition point, downstream pressure acquisition point, passage flow acquisition point and adjacent temperature acquisition point, and output a sealing part acquisition correspondence table. At each acquisition moment, the upstream pressure measurement, downstream pressure measurement, passage flow measurement and adjacent temperature measurement corresponding to the same sealing part are read synchronously according to the sealing part acquisition correspondence table. The difference between the upstream pressure measurement and the downstream pressure measurement is determined as the pressure value of the sealing part, the passage flow measurement is determined as the flow value of the sealing part, and the adjacent temperature measurement is determined as the temperature value of the sealing part. The acquisition value of the part is then output. The collected values ​​for each part are combined with the corresponding sealing part identifier and the collection time and written into the monitoring record, and the monitoring record sequence corresponding to each sealing part is output.

3. The handheld water-cooled sealing status monitoring system for an electromagnetic heating device according to claim 2, characterized in that: The execution of the fragment generation module includes: The monitoring records of the same sealing part are arranged in ascending order according to the acquisition time. The pressure change, flow rate change and temperature change between adjacent monitoring records are calculated. The pressure change sequence, flow rate change sequence and temperature change sequence are respectively subjected to three-term convolution smoothing and three-term median reconstruction. The reconstructed pressure change sign, flow rate change sign and temperature change sign form a direction vector. When the three signs of the direction vector are inconsistent, the sign combination corresponding to the sum of the absolute values ​​of the pressure residual, flow rate residual and temperature residual is taken as the change direction of the current monitoring record. The direction record sequence is output. Based on the directional recording sequence, a recording graph is established with each monitoring record as a node and adjacent monitoring records as directed edges. For each directed edge, the sum of the number of direction reversals, the number of pressure and flow sign mismatches, and the number of temperature timing mismatches is calculated as the edge cost. Dynamic programming is performed in the order of cumulative edge cost, number of monitoring records, and start acquisition time to segment the data. The resulting continuous node groups are determined as candidate segments, and the candidate segment sequence is output.

4. The handheld water-cooled sealing status monitoring system for an electromagnetic heating device according to claim 3, characterized in that: The execution of the fragment generation module also includes: For each candidate segment, calculate the cumulative pressure offset, cumulative flow offset, and cumulative temperature offset from the first monitoring record to the last monitoring record. Use the forward accumulation result and the backward substitution result as a check. If the cumulative pressure offset sign and the cumulative flow offset sign are the same in the forward accumulation result and the backward substitution result, retain the corresponding candidate segment as an abnormal segment. If the cumulative pressure offset sign or the cumulative flow offset sign is different, take the monitoring record corresponding to the sum of the edge costs of the two segments after splitting as the split point to reconstruct the candidate segment. Continue until the check of each reconstructed segment satisfies the requirement that the cumulative pressure offset sign and the cumulative flow offset sign are the same, and output the abnormal segment sequence.

5. A handheld water-cooled sealing status monitoring system for an electromagnetic heating device according to claim 4, characterized in that: The execution of the identity establishment module includes: Arrange the abnormal segment sequence of the same sealing part in the order of appearance, and calculate the direction consistency value, residual offset difference, duration difference and interval duration value between each subsequent abnormal segment and the previous abnormal segment in turn. When the direction consistency value is one, the residual offset difference and duration difference have the same sign and the interval duration value is not greater than the duration of the previous abnormal segment, candidate connection edges are established. When the direction consistency value is zero, the residual offset difference and duration difference have opposite signs or the interval duration value is greater than the duration of the previous abnormal segment, candidate connection edges are deleted. Output the candidate connection graph. Based on the candidate connection graph, the connection cost is calculated for each candidate connection edge, consisting of the absolute value of the residual offset difference, the absolute value of the duration difference, and the interval duration. Forward accumulation and backward substitution are performed in the order of the occurrence of abnormal segments. When there are multiple candidate connection edges in the same subsequent abnormal segment, the candidate connection edge with the smallest connection cost is selected as the retained connection edge. When the connection costs are the same, the candidate connection edge with the earlier end time of the previous abnormal segment is selected as the retained connection edge, and the remaining candidate connection edges are deleted. The connection chain is then output.

6. The handheld water-cooled sealing status monitoring system for an electromagnetic heating device according to claim 5, characterized in that: The execution of the identity establishment module also includes: For each link segment in the link chain, calculate the number of times the direction is maintained, the number of times the residual offset increases, and the number of times the duration increases between the first and last anomalous segments. When the number of times the direction is maintained is equal to the number of links in the link segment and the sum of the number of times the residual offset increases and the number of times the duration increases is not zero, the anomalous segments contained in the link segment are classified into the same damage identity. When the sum of the number of times the residual offset increases and the number of times the duration increases is zero, the last anomalous segment in the link segment is taken as the end segment of the current damage identity, and the next anomalous segment is taken as the start segment of the next damage identity, and the damage identity sequence is output.

7. A handheld water-cooled sealing status monitoring system for an electromagnetic heating device according to claim 6, characterized in that: The execution of the evolution determination module includes: Arrange the abnormal segments containing the same damage identity in the order of appearance, calculate the residual offset difference, duration difference and segment interval duration between each subsequent abnormal segment and the previous abnormal segment in turn, and record a residual increment when the residual offset difference is positive and the segment interval duration is not greater than the duration of the previous abnormal segment, and record a continuous expansion when the duration difference is positive, and output the evolution count result.

8. A handheld water-cooled sealing status monitoring system for an electromagnetic heating device according to claim 7, characterized in that: The execution of the evolution determination module also includes: Based on the evolution count results, each abnormal segment of the same damage identity is accumulated segment by segment in the order of appearance. When the residual increase occurs once and the continuous expansion occurs once, the damage identity is determined to be in a state of continuous deterioration. When there are no abnormal segments with a negative residual offset difference between two adjacent residual increases and the residual increase occurs twice in a row, the damage identity is determined to be in an irreversible damage state, and the health status result is output. The health status results are combined with the corresponding damage identification, end-abnormal segment identification, residual increment count, and continuous expansion count and written into the health status record, and the health status result sequence corresponding to the same sealing part is output.

9. A handheld water-cooled sealing status monitoring system for an electromagnetic heating device according to claim 8, characterized in that: The execution of the output module includes: The health status results, damage identification, and abnormal fragment identification corresponding to the same sealing part are arranged into a state node sequence according to the order of appearance. For each adjacent state node, the state transition value is obtained by calculating whether the health status result has changed, the identity continuation value is obtained by calculating whether the damage identification is the same, and the fragment reproduction value is obtained by calculating whether the abnormal fragment identification of the next state node appears in the previous state node. The state transition value, identity continuation value, and fragment reproduction value are added together to obtain the edge cost. Then, forward accumulation and backward substitution are performed on the state node sequence, and the state node chain corresponding to the accumulated edge cost is taken to generate the state record. Based on the state record, the health status result, damage identification, abnormal fragment identification, and generation time corresponding to the final state node are extracted and concatenated in the order of the fields to obtain the original record text. A hash mapping is performed on the original record text to obtain the current record fingerprint. The current record fingerprint is concatenated with the record fingerprint of the previous state record to generate a record chain fingerprint. When the health status result of the final state node is a continuously deteriorating state, a continuous deterioration warning message containing the record chain fingerprint is generated. When the health status result of the final state node is an irreversible damage state, an irreversible damage warning message containing the record chain fingerprint is generated.

10. A handheld water-cooled sealing status monitoring system for an electromagnetic heating device according to claim 9, characterized in that: The execution of the output module also includes: The seal location identifier, damage identification identifier, abnormal fragment identifier, and current record fingerprint in the status record are compared item by item with the seal location identifier, damage identification identifier, abnormal fragment identifier, and current record fingerprint in the warning information. If all items are the same, the status record and warning information are output. If any item is different, the status node sequence is retrieved again using the damage identification identifier and abnormal fragment identifier in the warning information. The status record generation, current record fingerprint generation, and item-by-item comparison are repeated until the status record and the warning information are identical item by item.

Images