A method and system for monitoring the operational state of a wastewater treatment plant

Abstract

The present application relates to the technical field of sewage treatment, in particular to a sewage treatment facility operation state monitoring method and system. The present application combines the collection time, the receiving time and the point quality for state frame generation, and performs quality prerequisite inspection before operation state determination, so that the data with non-comparable time or insufficient effective samples remain as untrusted data, and the control, electrical and process evidence in the same determination have consistent time boundaries, thereby reducing the situation that data loss or clock drift is misjudged as equipment failure, and the quality label of the state frame is also retained with the state event, so that the operation state received by the center platform has the data availability basis when the state is formed, and the determination stability under the fluctuating conditions of field sampling is improved.

Description

Technical Field

[0001] This invention belongs to the field of wastewater treatment technology, specifically relating to a method and system for monitoring the operational status of wastewater treatment facilities. Background Technology

[0002] In wastewater treatment facilities, booster pumps, return pumps, and blowers are driven by controller commands, and their operation causes corresponding changes in current, water level, flow rate, or dissolved oxygen. Existing centralized monitoring methods typically upload controller commands, electrical feedback, and process sensor data separately, with a central platform then assessing individual points. However, the clocks, sampling periods, and network latency of different data sources are not consistent. Consequently, commands, current, and process responses from the same equipment action can easily be categorized into different timeframes, leading to contradictory phenomena such as equipment starting without feedback or operating without commands.

[0003] Edge computing can push data processing down to the facility site, shortening the path from data acquisition to judgment. However, due to the existence of breakpoints, out-of-bounds errors, and clock drift in field data, applying single-point threshold rules only at the edge can lead to data gaps being mistaken for equipment failures, and momentary fluctuations near the threshold causing alarms to be repeatedly generated and restored. Because it is not first determined whether the data is sufficient to support operational status judgment, fault judgment and data quality anomalies can be confused.

[0004] To address the above problems, this invention proposes a method and system for monitoring the operational status of wastewater treatment facilities. Summary of the Invention

[0005] To address the aforementioned problems in the existing technology, the present invention aims to provide a method and system for monitoring the operational status of wastewater treatment facilities.

[0006] To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows: A method for monitoring the operational status of a wastewater treatment facility, comprising: Control intent, electrical response, and process response data are collected from field data sources and converted into unified point records. The unified point records are then aligned by time windows and marked with quality to generate status frames with quality markings. Evidence feature groups are constructed from state frames, quality-priority checks are performed, and state determination and migration are carried out based on evidence consistency comparison after the proportion of effective samples reaches a preset threshold. The running state and evidence summary are encapsulated into state events, and after local persistence, they are uploaded to the central platform. After being deduplicated and received by the central platform, a confirmation is returned, and breakpoint retransmission is achieved after the link is interrupted; configuration packets are sent through the downlink channel, version switching is completed at the boundary of the status frame, and rollback is performed when the rollback conditions are met.

[0007] Preferably, control intent, electrical response, and process response data are collected from field data sources and converted into unified point records, including: Collect control intent, electrical response, and process response data. Control intent refers to the start / stop status or frequency requirements issued by the controller to the equipment. Electrical response refers to the motor action represented by the operating feedback and current. Process response refers to the changes in water level, flow rate, or dissolved oxygen after the equipment is activated. Boolean states, integer register values, or floating-point sensor values ​​are converted into unified point records. The unified point records are assigned by facility identifiers and equipment identifiers, and carry the point type, acquisition time, and reception time. Records from the same facility, the same equipment, the same time window, and the same configuration version are generated into status frames after quality inspection.

[0008] Preferably, the uniform point records are aligned according to a time window and marked with quality to generate a status frame with quality marking, including: The start and end boundaries of the window are defined by the local clock, and the records are assigned to the corresponding windows according to the acquisition time. The alignment tolerance is used to limit the maximum acquisition time difference of records within the same time slice. When the time difference is less than or equal to the alignment tolerance, the records are assigned to the same slice. When the time difference is greater than the alignment tolerance, the records are retained and marked as unaligned. The status frame retains the window boundaries and alignment values, and carries the number of samples to be sampled, the number of valid samples, and the maximum time deviation. The defect code indicates the reason for missing, out-of-bounds, duplicate, or time misalignment. Records with inconsistent attribution are kept in the isolation queue. When the configuration versions are inconsistent, the version that has been in effect at the start of the status frame shall prevail, and the newer version shall be retained until the next time window boundary.

[0009] Preferably, an evidence feature group is constructed from the state frame, a quality-priority test is performed, and state determination and transition are carried out based on evidence consistency comparison after the proportion of effective samples reaches a preset threshold, including: The evidence feature group is intermediate data that binds the control intent, electrical response, and process response of the equipment with their respective quality conclusions. Once the facility identifier, equipment identifier, time window, and configuration version are all consistent, it enters the operation status determination. First, the effective sample ratio of the necessary points is compared. If any necessary point is lower than the first validity threshold, the data is output as unreliable and the corresponding defect code is output, but no equipment fault is output. After the quality conditions are met, the start-up intention and the operating current are compared. If the electrical response is met, the flow rate or dissolved oxygen is then compared to see if it changes in the expected direction. When inconsistency in evidence first occurs, the current type and start time are written into the candidate exception; when the duration of the candidate exception is greater than or equal to the corresponding confirmation time, it is converted into a confirmed exception, and returns to normal when the condition disappears before the confirmation time is reached; after the triggering condition of the confirmed exception disappears, it enters the recovery candidate, and returns to normal when the recovery condition is continuously met and the recovery confirmation time is reached.

[0010] Preferably, the running status and evidence digest are encapsulated into a status event, persisted locally, and then uploaded to the central platform, including: The status conclusion is based on the status type, retaining the candidate start time, confirmation time, and recovery time; the triggering evidence and quality conclusion describe the decision path, and the rule version is used to fix the configuration on which the current decision is based; The state conclusion and the evidence summary before and after the trigger are encapsulated as a state event. The state event supports idempotent reception with a unique event identifier and supports gap detection and timing recovery with a local sequence number. Status events are first written to the persistent storage of the edge node, and after successful persistence, the local sequence number is added to the queue to be sent.

[0011] Preferably, after receiving the data via deduplication by the central platform, an acknowledgment is returned, and retransmission is implemented after a link interruption, including: After receiving a status event, the central platform verifies its integrity, removes duplicates based on the event's unique identifier, and detects gaps based on the local sequence number. Once the platform successfully persists the event, it returns the maximum consecutive received sequence number from the original connection, and edge nodes clean up confirmed records that are no greater than that sequence number. When the sequence number gap still exists, the edge node retransmits from the earliest missing sequence number. The repeated arrival of events only updates the reception confirmation and does not add new status events. During uplink interruption, data acquisition, framing, judgment, and local persistence continue, and data is retransmitted after the link is restored.

[0012] Preferably, the configuration packet is sent through the downlink channel, the version switch is completed at the state frame boundary, and the rollback is performed when the rollback conditions are met, including: Configuration packages are identified by their configuration version, their applicable device type is used to limit their use, and their physical scope, characteristic parameters, and judgment thresholds are retained. When the integrity verification passes and the new version is higher than the current version, the configuration package enters the candidate configuration area first. After the current state frame is completed, the new version is used simultaneously by time alignment, feature processing and state determination; when the number of processing restarts reaches the rollback threshold or the single processing delay is greater than the delay threshold and continuously reaches the delay confirmation time during the observation period, the previous version is reset as the effective version. The generated state frames, state conclusions, and state events retain the rule version at the time, and rollback does not rewrite historical results.

[0013] The present invention also provides a wastewater treatment facility operation status monitoring system, comprising: The data acquisition module is used to collect control intent, electrical response, and process response data from field data sources and convert them into unified point records; The framing module is used to align uniform point records according to a time window and mark their quality, generating a status frame with quality markings; The determination module is used to construct evidence feature groups from the state frame, perform quality-priority tests, and determine and transfer the state based on evidence consistency comparison after the proportion of effective samples reaches a preset threshold. The event module is used to encapsulate the running status and evidence digest into status events, and upload them to the central platform after local persistence. The aggregation module is used to receive duplicate data from the central platform and return confirmation, and to enable breakpoint retransmission after the link is interrupted. The configuration module is used to send configuration packages through the downlink channel, complete version switching at the state frame boundary, and perform rollback when the rollback conditions are met.

[0014] Preferably, the framing module defines the start and end boundaries of the window using the local clock and assigns the records to the corresponding windows according to the acquisition time. The alignment tolerance is used to limit the maximum acquisition time difference of records within the same time slice. The status frame retains the window boundaries and alignment values ​​and carries the number of samples to be sampled, the number of valid samples, and the maximum time deviation. Records with inconsistent assignments are kept in an isolation queue. When the configuration versions are inconsistent, the version that has been effective at the beginning of the status frame shall prevail. The evidence feature group constructed by the judgment module is intermediate data that binds control intent, electrical response, process response and their respective quality conclusions. The judgment first compares the effective sample ratio of the necessary points. If any necessary point is lower than the first validity threshold, the output data is unreliable. After the quality conditions are met, the starting intent and operating current, electrical response and process response are compared in sequence. When the evidence inconsistency occurs for the first time, it is written as a candidate anomaly. When the duration of the candidate anomaly reaches the confirmation time, it is converted into a confirmed anomaly.

[0015] Preferably, the event module encapsulates the state conclusion and the evidence summary before and after the trigger into a state event. The state event supports idempotent reception with a unique event identifier and supports gap detection and timing recovery with a local sequence number. The state event is first written to persistent storage, and after successful persistence, the local sequence number is put into the queue to be sent. The aggregation module deduplicates events by their unique identifier and detects gaps by local sequence number. After successful persistence, it returns the maximum consecutive received sequence number. Edge nodes clean up acknowledgments no greater than this sequence number. If a sequence number gap exists, retransmission starts from the earliest missing sequence number. The configuration packages issued by the configuration module are identified by their configuration version and the applicable device type is used to limit the target users. When the new version is higher than the current version, it enters the candidate configuration area first. After the current status frame is completed, the new version is loaded by the framing module and the judgment module at the same time. If the rollback conditions are met during the observation period, the previous version is reset as the effective version.

[0016] Beneficial effects 1. This invention combines acquisition time, reception time, and location quality in the generation of status frames and performs a quality prerequisite check before determining the operating status. This ensures that data with incomparable time or insufficient effective samples remain unreliable, while ensuring that control, electrical, and process evidence in the same determination has consistent time boundaries. This reduces the possibility of data loss or clock drift being misjudged as equipment failure. The quality marker of the status frame is also retained with the status event, so that the operating status received by the central platform carries the data availability basis at the time the status was formed, thus improving the stability of determination under fluctuating field sampling conditions.

[0017] 2. Based on the sequential comparison of control intent, electrical response, and process response, this invention combines conditional transitions between candidate anomalies, confirmed anomalies, and recovery candidates to form an operating state that includes the failed evidence level and the corresponding rule version. It also suppresses repeated alarms caused by instantaneous fluctuations near the threshold, thereby alleviating the problem of difficulty in locating abnormal links when only observing a single electrical quantity or a single process quantity. The failed evidence level and the state transition time are both retained in the state event, so that the alarm output simultaneously indicates the link where the anomaly is located and the duration, thereby shortening the path from the platform event back to the on-site inspection object.

[0018] 3. This invention employs a mechanism that first persists event packets locally before uploading them, and then the platform deduplicates them using unique identifiers and returns the largest consecutive sequence number. Combined with configuration version switching at the state frame boundary and failure rollback, this mechanism can maintain edge judgment and event timing during uplink interruptions. After connection recovery, it enables breakpoint retransmission and avoids duplicate generation of platform events, thereby reducing the impact of network jitter and configuration switching on monitoring continuity. The receiving confirmation and local cleanup use the same consecutive sequence number as the boundary, and configuration rollback also retains the version used by historical state events. This ensures that the running state can be restored according to the original judgment order after link recovery and version recovery, thus enhancing the continuity of event evidence. Attached Figure Description

[0019] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a system module diagram of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments. It should be understood that the specific embodiments described herein are merely for explaining the invention and are not intended to limit the scope of protection of the invention.

[0021] Example 1 Please see Figure 1 As shown, this embodiment provides a method for monitoring the operational status of a wastewater treatment facility. Edge nodes are set up on-site at the wastewater treatment facility. An edge node refers to a computing device or software operating environment deployed close to the on-site data source to perform data standardization, time alignment, status determination, and network outage caching. The on-site data source consists of a controller, motor protector or frequency converter, and process sensors. Specifically, the method includes the following steps: S1. Collect on-site data and generate a status frame with quality markers; The acquisition adapter reads raw messages from the field network or serial link, acquiring control intent, electrical response, and process response data. Control intent refers to the controller's start / stop status or frequency requirements for the equipment; electrical response refers to the motor action represented by operating feedback and current; process response refers to changes in water level, flow rate, or dissolved oxygen after equipment operation. The acquisition adapter converts Boolean states, integer register values, or floating-point sensor values ​​into unified point records. A unified point record is a single acquisition record after normalization of numerical types and units of measurement, identified by facility and equipment identifiers, and includes the point type, acquisition time, and reception time. The recorded values ​​and units are used for subsequent feature processing, the raw quality code is used to retain the validity of the data source, and the incrementing sequence number is used to detect missing or duplicate records of the same point.

[0022] After unified location records are entered into the time window, edge nodes define the start and end boundaries of the window using their local clocks and assign records to the corresponding windows according to their acquisition time. The sampling period is the interval between adjacent planned acquisition times for the same location, which can be configured between 0.2 seconds and 60 seconds. Preferably, the sampling period is 1 second. Alignment tolerance is used to limit the maximum acquisition time difference of records within the same time slice, and can be selected between 20 milliseconds and 1000 milliseconds. Records with a time difference less than or equal to the selected alignment tolerance are assigned to the same slice; those with a time difference greater than the selected alignment tolerance are retained separately and marked as unaligned. Preferably, the alignment tolerance is 100 milliseconds.

[0023] Records within a time window are generated into status frames after quality inspection. A status frame is an aligned data object corresponding to the same facility, the same equipment, the same time window, and the same configuration version. The status frame retains the window boundaries and alignment values, and carries the number of samples to be taken, the number of valid samples, and the maximum time deviation; the defect code indicates the reason for missing, out-of-bounds, duplicate, or time misalignment, while the quality mark is used to determine whether subsequent evidence can participate in the equipment status determination.

[0024] When the window closes, edge nodes also verify the facility identifier, equipment identifier, and configuration version of the unified location record. Records with inconsistent attribution are kept in an isolation queue with an attached contextual defect conclusion and do not participate in the current state frame. When the configuration version is inconsistent, the version that was already in effect at the beginning of the state frame is used, and the newer version is kept in the candidate configuration area until the next time window boundary before participating in framing. The current state frame therefore only receives records with consistent attribution and version, and the resulting quality tags and evidence feature groups use the same context. When configuration packets or other device records arrive during window processing, they do not change the judgment criteria of the state frame. After the attribution and version verification is completed, the current time window is encapsulated into a state frame for continued use in evidence construction and operational status determination.

[0025] Quality inspection compares the deviation between the receiving time and the acquisition time, the occurrence of consecutive missing values, and the range of physical quantities. The clock anomaly threshold can be set between 0.5 seconds and 10 seconds; samples with deviations exceeding this threshold retain their original values ​​and are marked with a clock anomaly flag. Preferably, the clock anomaly threshold is set to 2 seconds. The water level physical range can be selected with upper and lower boundaries within the equipment's measurement range. Readings falling within the closed interval are retained as valid boundary data; readings below the lower boundary or above the upper boundary are marked as exceeding the sensor's boundary. Preferably, the water level physical range is 0 meters to 6 meters; readings exactly equal to the lower or upper boundary are still considered valid boundary data. Original values ​​marked as exceeding the boundary are only used for sensor anomaly events and are not included in equipment fault determination.

[0026] S2. Perform quality prerequisite checks on the state frames and construct evidence feature groups; When the quality conclusion of a status frame meets the processing conditions, the edge node generates an evidence feature set from it. The evidence feature set is intermediate data that binds the device's control intent, electrical response, and process response to their respective quality conclusions. The control intent retains the command status and duration; the electrical response retains the operating feedback duration and median current value; and the process response retains the flow rate change or dissolved oxygen change during the response period. The feature window length is used to limit the historical duration of a single status determination and can be selected between 30 seconds and 1800 seconds. Preferably, the feature window length is 300 seconds. The effective sample ratio is the percentage of samples that pass the quality inspection within the feature window out of the total samples to be sampled, retaining only its name, percentage unit, and determination direction not lower than the first validity threshold. Preferably, the first validity threshold is 80%.

[0027] Only after all facility identifiers, equipment identifiers, time windows, and configuration versions in the evidence feature group are consistent can the operational status be determined. The determination first compares the effective sample ratio of necessary points; if any necessary point falls below the first validity threshold, the output data is deemed unreliable and a corresponding defect code is given, but no equipment fault is output. After the quality conditions are met, the start-up intention and operating current are compared; if the electrical response is met, the flow rate or dissolved oxygen is then compared to see if it changes in the expected direction. This sequence allows evidence levels that fail to pass to directly proceed to the status conclusion, facilitating the differentiation between control, electrical, and process aspects.

[0028] S3. State determination and migration based on evidence consistency comparison; The operating current threshold for pump equipment is the minimum current required to determine if the motor has generated a valid electrical response. It can be set between 0.1 A and 500 A according to the equipment's rated operating conditions. Preferably, the operating current threshold is 3.0 A; a current exactly equal to this threshold is considered to have generated an electrical response. If there is an intention to start but the current is below this threshold and continues until the candidate confirmation time, the status type is "start-up failure"; if there is no intention to start but the current is greater than or equal to this threshold and continues until the candidate confirmation time, the status type is "uninstructed operation". The flow response threshold is the minimum increment of the outlet flow rate after pump startup relative to the pre-start reference, and can be 0.1 m³ / s. 3 / h to 100m 3 The threshold should be selected between / h. Preferably, the flow response threshold is 0.5m. 3 When the increment is exactly equal to the threshold, the process response boundary is considered reached. If the current is greater than or equal to the operating current threshold, but the flow rate increment is lower than the flow rate response threshold and continues to reach the process response period, the status type is insufficient process response.

[0029] The dissolved oxygen response threshold for the blower is the minimum increase in dissolved oxygen relative to the pre-operation baseline after the equipment starts operating, and can be set between 0.05 mg / L and 2.0 mg / L. If the blower has generated an electrical response, but the dissolved oxygen increase is below this threshold and continues to reach the process response period, the status type is insufficient aeration response. The hydraulic load determination simultaneously checks the water level, influent flow rate, and effluent flow rate. A hydraulic load imbalance candidate is generated only when the combined conditions of the water level being higher than the first water level threshold, the influent flow rate being higher than the first influent threshold, and the effluent flow rate being lower than the first effluent threshold continue to reach the hydraulic confirmation time. The first water level threshold is preferably 1.2 meters; the first influent threshold is 80% of the design treatment flow rate; the first effluent threshold is 50% of the design treatment flow rate; and the hydraulic confirmation time is preferably 60 seconds.

[0030] When inconsistency first occurs, the edge node writes the current type and start time into a candidate anomaly. If the duration of a candidate anomaly is greater than or equal to the corresponding confirmation time, it becomes a confirmed anomaly. If the conditions disappear before the confirmation time is reached, it returns to normal. Preferably, the start-up failure confirmation time is 15 seconds, the process response period is 30 seconds, and the recovery confirmation time is 60 seconds. After the triggering conditions of a confirmed anomaly disappear, it enters the recovery candidate. If the recovery conditions are continuously met until the recovery confirmation time is reached, it returns to normal. If it crosses the boundary again before the recovery confirmation time is reached, it remains a confirmed anomaly. The migration logic between the candidate anomalies, confirmed anomalies, and recovery candidates distinguishes between instantaneous boundary crossings, continuous anomalies, and stable recovery. When the quality of a necessary point is not met, it enters a data unreliable state. After the quality recovers and continues until the quality recovery time is reached, it re-enters normal evaluation, thus preventing unreliable data from directly changing the equipment fault state. The preferred quality recovery time is 10 seconds.

[0031] S4. Encapsulate the judgment result into a status event and complete local persistence; State transitions generate state conclusions, which are centered on the state type and retain candidate start time, confirmation time, and recovery time. Triggering evidence and quality conclusions describe the decision path, and rule versions are used to fix the configuration on which the current decision is based. State conclusions and evidence summaries before and after triggering are encapsulated as state events. State events are data objects used to transmit a change in operational state and its necessary evidence to the central platform. They support idempotent reception with unique event identifiers and gap detection and timing recovery with local sequence numbers. Evidence summaries retain passed and failed evidence and their quality markers, while integrity check values ​​are used by the platform to check the transmission results before storage.

[0032] Status events are first written to the persistent storage of the edge node, and after successful persistence, the local sequence number is added to the queue to be sent.

[0033] S5. Upload status events to the central platform, receive them after deduplication, and return confirmation. Then, resume transmission at the breakpoint if the link is interrupted. After receiving a status event, the central platform verifies its integrity, deduplicates it based on the event's unique identifier, and detects gaps based on the local sequence number. Once the platform successfully persists the data, it returns the largest consecutive received sequence number from the original connection. Edge nodes only clean up acknowledged records with a sequence number no greater than that number. When a sequence number gap still exists, the edge nodes retransmit from the earliest missing sequence number, ensuring that duplicate events only update the received acknowledgments without creating new status events.

[0034] During uplink interruption, data acquisition, status frame generation, evidence feature group construction, operational status determination, and local persistence continue. The offline judgment threshold can be selected between two and five consecutive transmission failures; preferably, three consecutive transmission failures are used. A single timeout can be set between 2 and 30 seconds; preferably, a single timeout is 5 seconds.

[0035] The retry interval can be increased progressively, with an upper limit configured between 60 and 600 seconds. Preferably, the upper limit of the retry interval is 600 seconds. When the number of pending events reaches the cache limit, the cache limit is preferably 2000 entries. Unrecovered anomalies and evidence of the first trigger are still retained. Heartbeat summaries are merged according to time intervals, so that the causal information of the failure event is not affected when storage degrades.

[0036] S6. Send configuration packages through the downlink channel, complete version switching at the status frame boundary, and perform rollback when conditions are met, while maintaining the independence of the boundary between monitoring and control. Configuration feedback uses a downlink connection independent of uplink events. The configuration package is versioned judgment data generated by the central platform, identifying old and new versions by configuration version, limiting the applicable devices by applicable device type, and retaining the physical range, characteristic parameters, and judgment thresholds. When the integrity check passes and the new version is higher than the current version, the configuration package first enters the candidate configuration area; after the current status frame is completed, the new version is then used simultaneously for time alignment, feature processing, and status judgment. The observation period can be set between 300 seconds and 3600 seconds; preferably, it is 600 seconds.

[0037] If the number of process restarts reaches the rollback threshold during the observation period, or if a single processing delay exceeds the delay threshold and continuously reaches the delay confirmation time, the previous version is reset as the effective version. Preferably, the rollback threshold is 3 process restarts, the delay threshold is 2 seconds, and the delay confirmation time is 30 seconds. Generated status frames, status conclusions, and status events retain the rule version at the time of generation; rollback does not rewrite historical results.

[0038] The monitoring link takes control intent as the observed input, and provides the platform with operational status and status events for status updates and alarm output. The judgment results of edge nodes are not written to the controller start / stop channel along the monitoring link, thus maintaining a clear boundary between status monitoring and field control execution.

[0039] Example 2 Please see Figure 2 As shown, this embodiment provides a wastewater treatment facility operation status monitoring system, including the following modules: The acquisition module is used to collect control intent, electrical response, and process response data from field data sources and output a unified point record. The acquisition adapter in the acquisition module is responsible for protocol and numerical type conversion, and the output of the acquisition module is connected to the input of the framing module through the internal message channel of the edge node.

[0040] The framing module aligns and marks the quality of uniform location records according to a time window, generating a status frame with quality markings. The framing module only accepts records that are recognizable by facility identifier, equipment identifier, and effective configuration version; records that fail to be recognized are placed in an isolation queue. After the status frame is encapsulated, it is delivered to the decision module via an internal message channel; the original readings do not bypass the framing module's path.

[0041] The decision module constructs evidence feature groups from state frames and determines the operational status based on quality prerequisite checks, evidence consistency comparisons, and state transitions. The input interface of the decision module checks whether the state frame and the current state machine record belong to the same device, and whether the state frame, feature parameters, and decision threshold use the same configuration version. If any context mismatch occurs, the previously confirmed device state is retained, and a context defect conclusion is output to prevent the merging of evidence across devices or versions.

[0042] The event module encapsulates the running status and evidence digest into status events, persists them locally, and then uploads them via the uplink event interface. The local storage index of the event module is sorted by local sequence number, and the pending queue only stores the sequence numbers that have been successfully persisted. After receiving the confirmation, the event module cleans up records that have been covered by the largest consecutive sequence number, and when a sequence number gap is found, it puts the earliest missing sequence number and the records following it back into the pending queue.

[0043] The aggregation module performs integrity checks, deduplication, and sequence number gap detection on status events, updates facility operating status, and returns a receipt confirmation. Deployed on the central platform, its data entry operation uses the event's unique identifier as the idempotent key; when the same event arrives again, only the received status is merged. Receipt confirmations are generated only after the status event has been persisted, ensuring that the event module's cleanup criteria align with the platform's actual storage progress.

[0044] The configuration module generates and distributes configuration packages, verifies candidate configurations, switches configuration versions at status frame boundaries, and restores the previous version when rollback conditions are met during the observation period. This module outputs configuration packages to the central platform and has candidate and active configuration areas at the edge nodes. The switching event is triggered by the current status frame completion marker output by the framing module, and the new version is then loaded jointly by the framing and decision modules. The switching result and failure reason are sent back to the central platform via the downlink configuration connection's return channel, enabling the central platform to determine the actual configuration version used by the edge nodes.

[0045] During system operation, unified location recording can only be input along the connection from the acquisition module to the framing module, and the judgment module can only use status frames that have completed quality marking. The status conclusion is then persisted by the event module and, after successful persistence, is handed over to the aggregation module via the uplink event interface. The aggregation module's receipt confirmation is then returned to the event module along the original connection, and the version switching result of the configuration module is then sent back to the central platform via the return channel. When there is a context mismatch, unverified quality, unpersisted event, or the configuration has not yet reached the status frame boundary, the corresponding subsequent path remains closed.

[0046] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for monitoring the operational status of a wastewater treatment facility, characterized in that, include: Control intent, electrical response, and process response data are collected from field data sources and converted into unified point records. The unified point records are then aligned by time windows and marked with quality to generate status frames with quality markings. Evidence feature groups are constructed from state frames, quality-priority checks are performed, and state determination and migration are carried out based on evidence consistency comparison after the proportion of effective samples reaches a preset threshold. The running state and evidence summary are encapsulated into state events, and after local persistence, they are uploaded to the central platform. After being deduplicated and received by the central platform, a confirmation is returned, and breakpoint retransmission is achieved after the link is interrupted; configuration packets are sent through the downlink channel, version switching is completed at the boundary of the status frame, and rollback is performed when the rollback conditions are met.

2. The method for monitoring the operational status of a wastewater treatment facility according to claim 1, characterized in that, Control intent, electrical response, and process response data are collected from field data sources and converted into unified point records, including: Collect control intent, electrical response, and process response data. Control intent refers to the start / stop status or frequency requirements issued by the controller to the equipment. Electrical response refers to the motor action represented by the operating feedback and current. Process response refers to the changes in water level, flow rate, or dissolved oxygen after the equipment is activated. Boolean states, integer register values, or floating-point sensor values ​​are converted into unified point records. The unified point records are assigned by facility identifiers and equipment identifiers, and carry the point type, acquisition time, and reception time. Records from the same facility, the same equipment, the same time window, and the same configuration version are generated into status frames after quality inspection.

3. The method for monitoring the operational status of a wastewater treatment facility according to claim 1, characterized in that, Align the unified location records by time window and mark their quality to generate a status frame with quality markings, including: The start and end boundaries of the window are defined by the local clock, and the records are assigned to the corresponding windows according to the acquisition time. The alignment tolerance is used to limit the maximum acquisition time difference of records within the same time slice. When the time difference is less than or equal to the alignment tolerance, the records are assigned to the same slice. When the time difference is greater than the alignment tolerance, the records are retained and marked as unaligned. The status frame retains the window boundaries and alignment values, and carries the number of samples to be sampled, the number of valid samples, and the maximum time deviation. The defect code indicates the reason for missing, out-of-bounds, duplicate, or time misalignment. Records with inconsistent attribution are kept in the isolation queue. When the configuration versions are inconsistent, the version that has been in effect at the start of the status frame shall prevail, and the newer version shall be retained until the next time window boundary.

4. The method for monitoring the operational status of a wastewater treatment facility according to claim 1, characterized in that, Evidence feature groups are constructed from state frames, and quality-priority checks are performed. After the proportion of valid samples reaches a preset threshold, state determination and transition are performed based on evidence consistency comparison, including: The evidence feature group is intermediate data that binds the control intent, electrical response, and process response of the equipment with their respective quality conclusions. Once the facility identifier, equipment identifier, time window, and configuration version are all consistent, it enters the operation status determination. First, the effective sample ratio of the necessary points is compared. If any necessary point is lower than the first validity threshold, the data is output as unreliable and the corresponding defect code is output, but no equipment fault is output. After the quality conditions are met, the start-up intention and the operating current are compared. If the electrical response is met, the flow rate or dissolved oxygen is then compared to see if it changes in the expected direction. When inconsistency in evidence first occurs, the current type and start time are written into the candidate exception; when the duration of the candidate exception is greater than or equal to the corresponding confirmation time, it is converted into a confirmed exception, and returns to normal when the condition disappears before the confirmation time is reached; after the triggering condition of the confirmed exception disappears, it enters the recovery candidate, and returns to normal when the recovery condition is continuously met and the recovery confirmation time is reached.

5. The method for monitoring the operational status of a wastewater treatment facility according to claim 1, characterized in that, The runtime status and evidence digest are encapsulated into state events, persisted locally, and then uploaded to the central platform, including: The status conclusion is based on the status type, retaining the candidate start time, confirmation time, and recovery time; the triggering evidence and quality conclusion describe the decision path, and the rule version is used to fix the configuration on which the current decision is based; The state conclusion and the evidence summary before and after the trigger are encapsulated as a state event. The state event supports idempotent reception with a unique event identifier and supports gap detection and timing recovery with a local sequence number. Status events are first written to the persistent storage of the edge node, and after successful persistence, the local sequence number is added to the queue to be sent.

6. The method for monitoring the operational status of a wastewater treatment facility according to claim 1, characterized in that, After receiving the data through the central platform for deduplication, an acknowledgment is returned, and retransmission is implemented after a link interruption, including: After receiving a status event, the central platform verifies its integrity, removes duplicates based on the event's unique identifier, and detects gaps based on the local sequence number. Once the platform successfully persists the event, it returns the maximum consecutive received sequence number from the original connection, and edge nodes clean up confirmed records that are no greater than that sequence number. When the sequence number gap still exists, the edge node retransmits from the earliest missing sequence number. The repeated arrival of events only updates the reception confirmation and does not add new status events. During uplink interruption, data acquisition, framing, judgment, and local persistence continue, and data is retransmitted after the link is restored.

7. The method for monitoring the operational status of a wastewater treatment facility according to claim 1, characterized in that, The configuration package is sent via the downlink channel, the version switch is completed at the state frame boundary, and the rollback is performed when the rollback conditions are met, including: Configuration packages are identified by their configuration version, their applicable device type is used to limit their use, and their physical scope, characteristic parameters, and judgment thresholds are retained. When the integrity verification passes and the new version is higher than the current version, the configuration package enters the candidate configuration area first. After the current state frame is completed, the new version is used simultaneously by time alignment, feature processing and state determination; when the number of processing restarts reaches the rollback threshold or the single processing delay is greater than the delay threshold and continuously reaches the delay confirmation time during the observation period, the previous version is reset as the effective version. The generated state frames, state conclusions, and state events retain the rule version at the time, and rollback does not rewrite historical results.

8. A wastewater treatment facility operation status monitoring system, characterized in that, include: The data acquisition module is used to collect control intent, electrical response, and process response data from field data sources and convert them into unified point records; The framing module is used to align uniform point records according to a time window and mark their quality, generating a status frame with quality markings; The determination module is used to construct evidence feature groups from the state frame, perform quality-priority tests, and determine and transfer the state based on evidence consistency comparison after the proportion of effective samples reaches a preset threshold. The event module is used to encapsulate the running status and evidence digest into status events, and upload them to the central platform after local persistence. The aggregation module is used to receive duplicate data from the central platform and return confirmation, and to enable breakpoint retransmission after the link is interrupted. The configuration module is used to send configuration packages through the downlink channel, complete version switching at the state frame boundary, and perform rollback when the rollback conditions are met.

9. A wastewater treatment facility operation status monitoring system according to claim 8, characterized in that, The framing module defines the start and end boundaries of the window using the local clock and assigns records to the corresponding window according to the acquisition time. The alignment tolerance is used to limit the maximum acquisition time difference of records within the same time slice. The status frame retains the window boundaries and alignment values ​​and carries the number of samples to be sampled, the number of valid samples, and the maximum time deviation. Records with inconsistent assignments are kept in the isolation queue. When the configuration versions are inconsistent, the version that has been effective at the beginning of the status frame shall prevail. The evidence feature group constructed by the judgment module is intermediate data that binds control intent, electrical response, process response and their respective quality conclusions; the judgment first compares the effective sample ratio of the necessary points, and the output data is unreliable when any necessary point is lower than the first validity threshold; Once the quality conditions are met, the startup intent and operating current, electrical response and process response are compared sequentially. When the evidence is inconsistent for the first time, it is written as a candidate anomaly. When the duration of the candidate anomaly reaches the confirmation time, it is converted into a confirmed anomaly.

10. A wastewater treatment facility operation status monitoring system according to claim 8, characterized in that, The event module encapsulates the state conclusion and the evidence summary before and after the trigger into a state event. The state event supports idempotent reception with a unique event identifier and supports gap detection and timing recovery with a local sequence number. The state event is first written to persistent storage, and after successful persistence, the local sequence number is put into the queue to be sent. The aggregation module deduplicates events by their unique identifier and detects gaps by local sequence number. After successful persistence, it returns the maximum consecutive received sequence number. Edge nodes clean up acknowledgments no greater than this sequence number. If a sequence number gap exists, retransmission starts from the earliest missing sequence number. The configuration packages issued by the configuration module are identified by their configuration version and their applicable device type. When the new version is higher than the current version, it enters the candidate configuration area first. After the current status frame is completed, the new version is loaded by the framing module and the judgment module at the same time. If the rollback conditions are met during the observation period, the previous version will be reset as the effective version.

Images