Liquid level dynamic balance regulating control method for sewage lifting pump station

Abstract

This invention discloses a dynamic balance regulation and control method for liquid level in sewage lifting pump stations, relating to the field of sewage pump station control technology. The method includes configuring different principle liquid level detection devices and high-high and low-low liquid level switches in the wet well; collecting pump frequency, current, start-up / shutdown, and fault information and preprocessing it to generate validity labels and signal quality tags; calculating liquid level reliability based on pump reachability, cross-consistency, and quality tags; outputting fused liquid level, fused reliability, and distortion type labels; establishing a well chamber volume-liquid level relationship; forming physical consistency deviation and cumulative deviation based on the net inflow / outflow water difference; reconstructing the soft-sensor liquid level when necessary and using liquid level switches for clamp correction; switching graded control and limit interlocking connections based on fused reliability and consistency status, and recording and tracing. This method reduces false start-up / shutdown and false alarms, lowers the risk of overflow and dry running, suppresses water hammer and frequent start-up / shutdown, and improves control continuity and maintenance efficiency.

Description

Technical Field

[0001] This invention relates to the field of sewage pumping station control technology, specifically to a method for dynamic balance regulation and control of liquid level in sewage lift pumping stations. Background Technology

[0002] Wastewater lifting pump stations are mostly located in low-lying areas such as municipal drainage networks, industrial parks, and residential communities to lift gravity-concentrated wastewater to subsequent pipe networks or treatment facilities. The wet wells of the pump stations serve as collection and storage facilities. Inflow water varies daily, due to rainfall, and discharge patterns. They are often operated unattended or minimally staffed, relying on automatic start / stop based on liquid level signals, frequency conversion regulation, and multiple pump activation / deactivation to balance the risks of overflow at high liquid levels and dry running at low liquid levels. Typical control methods include threshold-based start / stop and pump rotation based on liquid level switches, constant liquid level frequency conversion regulation based on wet well liquid level feedback, and pump activation / deactivation based on liquid level change rate or operating frequency thresholds (e.g., Chinese patent publication numbers CN101922443A and CN101761490A). The "Design Code for Automation Systems of Drainage Pumping Stations" requires that water level detection devices should be redundant and that an alarm be triggered when the difference between two sets of readings reaches a certain proportion; when controlling the switch quantity according to water level, a liquid level switch should be used; the "Technical Code for Operation, Maintenance and Safety of Urban Drainage Pipelines and Pumping Stations" (2016 edition) requires that liquid level instrument sensors be cleaned and calibrated regularly.

[0003] Other literature suggests using multiple pressure sensors and taking the median to determine the liquid level (Chinese Patent Publication No. CN109709993A), or using pump operating information to detect components and output fault or limit liquid level alarms (Chinese Patent Publication No. CN103075331A); there are also designs with dual controllers and dual redundant liquid level sensors to achieve unattended drainage (Publication No. CN214225776A). In actual wet well media and environments, floating foam, grease and silt, condensation, reflections from well components, and hydraulic changes can cause liquid level detection to jump, get stuck, or drift slowly. When the control method uses liquid level as the main control variable for start-stop or frequency conversion regulation, liquid level errors will gradually accumulate and cause threshold misjudgment or deviation in adjustment direction, resulting in decoupling of pump status from actual water level, leading to frequent start-stop, insufficient pumping volume, excessive pumping, and even overflow discharge, environmental risks, or abnormal wear and protection behavior caused by low liquid level circulation.

[0004] Therefore, the problem with existing technology is:

[0005] When the wet well of the sewage lifting pump station is unattended or poorly staffed, and the liquid level detection device is affected by media contamination, foam condensation, structural reflection, and hydraulic fluctuations, resulting in jumps, jamming, or slow drift, the existing start-stop frequency conversion control, redundant alarms, and self-check alarms that use liquid level as the main control variable lack constraints. This leads to distorted liquid level participating in start-stop or adjustment judgments, and the control commands are inconsistent with the actual liquid level state, causing safety and operational deviations. Summary of the Invention

[0006] (a) Technical problems to be solved

[0007] To address the shortcomings of existing technologies, this invention provides a dynamic balance regulation and control method for liquid level in sewage lifting pump stations. Based on pump accessibility, cross-consistency, and quality markers, the liquid level reliability is calculated, and a fused liquid level, fused reliability, and distortion type label are output. A well chamber volume-liquid level relationship is established, and physical consistency deviation and cumulative deviation are generated according to the net inflow-outflow water difference. If necessary, the soft-sensor liquid level is reconstructed and corrected by a liquid level switch. Based on the fused reliability and consistency status, graded control and limit interlocking are switched and recorded for traceability. This method reduces false start / stop and false alarms, lowers the risk of overflow and dry running, suppresses water hammer and frequent start / stop, and solves the technical problems described in the background art.

[0008] (II) Technical Solution

[0009] To achieve the above objectives, the present invention provides the following technical solution:

[0010] A method for dynamic balance regulation and control of liquid level in sewage lifting pump stations includes: arranging two types of liquid level detection devices and high-high and low-low liquid level switches; collecting and preprocessing liquid level signals and pump operation information to obtain multi-channel liquid level data, validity labels, and signal quality labels; calculating the reliability score of each liquid level based on the pump action reachability criterion and multi-channel liquid level cross-consistency criterion formed by the signal quality labels and pump operation information; and weighting and fusing the liquid level reliability scores to obtain the fused liquid level, fused reliability, and distortion type label.

[0011] Establish the well chamber volume-liquid level relationship and convert the fused liquid level into volume change. Compare it with the net inflow / outflow difference estimated from the pump set operation information to obtain the physical consistency deviation. When the fused reliability is low or the physical consistency deviation exceeds the limit, reconstruct the soft measurement liquid level and correct it with a liquid level switch. Generate pump set control commands based on the fused liquid level, fused reliability, physical consistency deviation and liquid level switch. Switch the graded dynamic balance control and limit interlocking pipe according to the fused reliability and physical consistency deviation. Output maintenance suggestions and traceability records.

[0012] Furthermore, non-contact level gauges, submersible static pressure level gauges, high-high level switches, and low-low level switches are installed in the wet well to collect variable frequency operation commands and feedback, operation and fault status, and motor current. Range verification, glitch removal, rate of change constraints, and sampling period alignment are performed on the level signal to generate validity labels and signal quality marks.

[0013] Furthermore, update cycle and communication discrimination are performed on the liquid level signal, and out-of-bounds, long-term no-update, communication abnormality and self-diagnosis abnormality are written into the validity label; lost wave, over-range and echo quality difference are parsed into structured fields and written into the signal quality label, and the pump start and stop times are recorded to form a pump action event sequence.

[0014] Furthermore, when the validity is marked as available and the signal quality label meets the set conditions, a response time window after pump start-up and a reverse response time window after pump shutdown are set. Based on the pump action event sequence and pump group operation information, the reachability of pump action is judged for the liquid level trend. The soft threshold difference and hard threshold difference of multiple liquid level differences are combined to form a cross-consistency criterion, and the liquid level confidence score is calculated.

[0015] Furthermore, the multi-level liquid levels are weighted and robustly fused based on the liquid level credibility score to obtain the fused liquid level, and the fused credibility is generated simultaneously. Based on the pump action reachability judgment, cross consistency criterion and liquid level change mode output distortion type label, the fused credibility is used as the judgment quantity for control level switching.

[0016] Furthermore, a well chamber volume-level lookup table is established based on wet well geometry data and stored with monotonically increasing height and volume nodes. During operation, within the sampling interval where the fusion reliability meets the set conditions, the well chamber volume-level lookup table is segmented and corrected based on the fused liquid level and pump operation information to form an equivalent cross-sectional area curve.

[0017] Furthermore, the fusion liquid level is converted into volume change by looking up the well chamber volume-liquid level table, and compared with the drainage volume determined by the outflow rate measurement and the inflow volume determined by the pump shutdown interval to form the net inflow-outflow difference, thus obtaining the physical consistency deviation; the physical consistency deviation is accumulated according to the sampling period to obtain the cumulative deviation, and the consistency status is output.

[0018] Furthermore, the drainage volume in the net inflow-outflow difference is estimated based on the frequency feedback of the variable frequency operation, the motor current and the outlet water pressure, combined with the pump type characteristics, and calibrated with a slowly updated correction factor; the physical consistency deviation is accumulated according to the sampling period to obtain the cumulative deviation, and the systematic distortion judgment is output based on the cumulative deviation.

[0019] Furthermore, in cases of low fusion reliability and excessive physical consistency deviation, the soft measurement level is obtained by integral deduction of the well chamber volume based on the net inflow-outflow water difference and back-calculation through well chamber volume-liquid level lookup table; when the high-high liquid level switch is triggered, the soft measurement level is corrected to be no lower than the corresponding height, and when the low-low liquid level switch is triggered, the soft measurement level is corrected to be no higher than the corresponding height.

[0020] Furthermore, a four-level control hierarchy is set up: normal dynamic balance, downgraded conservative, distortion takeover, and extreme interlock takeover. The control hierarchy is switched based on the fusion credibility and physical consistency deviation.

[0021] Hysteresis, minimum running time, and minimum downtime are set at each control level. Maintenance suggestions are output and traceability records are recorded when distortion type labels continue to appear and when cumulative deviation exceeds the limit. The cumulative deviation is the deviation obtained by accumulating the physical consistency deviation according to the sampling period.

[0022] (III) Beneficial Effects

[0023] This invention provides a method for dynamic balance regulation and control of liquid level in sewage lifting pump stations, which has the following beneficial effects:

[0024] Information is collected by liquid level detection devices based on different principles, along with high-high and low-low liquid level switches. The same liquid level signals and pump operation information are preprocessed, sampling period aligned, validity labeled, and signal quality marked. This facilitates subsequent calculations to achieve consistent time and the same measurement range, effectively avoiding false triggering due to glitches, out-of-bounds errors, or communication anomalies. Pump reachability criteria, signal quality markings, and cross-consistency criteria are linked to generate liquid level reliability, obtaining a fused liquid level and fused reliability. Distortion type labels are output, allowing for differentiated handling based on states such as jumps, jams, drifts, and loss, avoiding single alarms and single-channel switching.

[0025] A well chamber volume-level relationship is established, and consistency verification of fused level changes and net inflow / outflow difference is implemented within the well chamber volume domain. Physical consistency deviation and cumulative deviation are used to identify systematic distortions caused by long-term fouling, common-mode drift, and slow changes in pump output, enhancing the ability to identify slow-changing anomalies. When fusion reliability decreases or physical consistency deviation exceeds limits, the soft-sensor level is reconstructed. Upper and lower boundary clamping corrections are performed by high-high and low-low level switches, ensuring that the level reference is continuous and subject to hard boundary constraints during sensor distortion, reducing the risk of control mismatch under distortion conditions.

[0026] The system triggers hierarchical dynamic balance control and limit interlocking takeover based on reliability and consistency status. It utilizes confirmation time, minimum operating time, minimum downtime, and rate of change to limit the closed-loop action chain, minimizing start-up, shutdown, and hydraulic shock. Distortion type labels, cumulative deviations, control levels, and control commands are compiled into traceability records, and maintenance recommendations are output. Actions such as cleaning, calibration, checking for obstructions, and ventilation are linked to corresponding abnormal evidence, reducing troubleshooting and enabling timely restoration of operation. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of the structure of the dynamic balance adjustment and control method for the sewage lift pump station of the present invention. Detailed Implementation

[0028] 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.

[0029] Please see Figure 1 This invention provides a method for dynamic balance regulation and control of liquid level in a sewage lifting pump station, comprising:

[0030] The medium inside a wet well contains solid impurities, grease, and fibers, and the liquid surface is prone to foaming and fluctuations. In addition, the high humidity inside the well makes it easy for condensation to occur. The liquid level detection device and communication link may experience short-term jumps, intermittent loss, or slow drift during long-term operation.

[0031] Step 1: At the wet well site, convert the multi-source liquid level information and pump operation information into a usable data sequence under the same time reference, and attach traceable validity labels and signal quality labels to each data point, so that Step 2 can directly use the aligned and labeled data to calculate the liquid level reliability.

[0032] The wet wells of sewage lifting pump stations are enclosed spaces characterized by high humidity, easy condensation, and impurities in the medium. The liquid surface is often accompanied by foam, floating matter, and backflow. In this environment, level detection devices may experience short-term jumps, intermittent loss of data, or slow drift. Pump start-up and shutdown, check valve closure, and pipeline back pressure can also cause non-monotonic fluctuations in the liquid level over short periods. If the data acquisition side directly uses unprocessed level measurements as control input, subsequent analysis may easily misinterpret abnormal measurements as actual level changes, triggering unnecessary start-ups, shutdowns, or erroneous frequency converter adjustments. Therefore, it is crucial to ensure data availability, traceability, and time alignment, and to generate validity markers and signal quality indicators.

[0033] Starting with the on-site structure, the installation heights of the first liquid level detection device, the second liquid level detection device, and the limit liquid level switch are uniformly mapped to the zero-point reference of the wet well. The distance from the measuring point to the zero-point reference is fixed as a configuration value of the acquisition system, so that the acquisition output is directly a liquid level measurement value under the same dimension, thereby reducing redundant conversions and interpretation discrepancies on the control side.

[0034] Among them, non-contact level gauges are installed on the well cover or well top support, and are obstructed by the installation angle and internal well components; submersible hydrostatic level gauges are suspended inside the well, and are determined by the fixed point height and conductor tension. Since neither type of gauge has a zero-point reference on the data acquisition side, the first level measurement value in subsequent steps... Second liquid level measurement value If there is a discrepancy, a fixed bias will be entered, which will cause the cross-consistency judgment to be triggered prematurely, and the stability of the flag will decrease.

[0035] First, the bottom of the well is determined as the zero-point reference, and the installation reference surface at the top of the well or the well cover is used as the measurement reference surface. Then, the output of the first liquid level detection device is converted to zero, and the static pressure output of the second liquid level detection device is also converted to zero. The trigger heights of the high-high liquid level switch and the low-low liquid level switch are written into the event record of the acquisition system as limit anchoring events, so that the liquid level interval boundary corresponding to the time of the limit event can be directly traced back.

[0036] First, the data acquisition side converts the raw distance measurement of the first liquid level detection device into the first liquid level measurement value. During the conversion, the zero-position reference is used as the reference, and the height difference between the installation reference surface and the zero-position reference is written into the acquisition system as a fixed configuration value, so that... The value directly represents the height of the liquid level relative to the zero reference. Secondly, the acquisition side converts the original static pressure of the second liquid level detection device into the measured value of the second liquid level. During the conversion, the vertical distance from the center of the sensor diaphragm to the zero reference is used as a fixed configuration value, and the triggering states of the high-high liquid level switch and the low-low liquid level switch are recorded as limit anchoring events. The trigger timestamp and the corresponding installation height configuration value are retained in the event record.

[0037] Under the same implementation approach, the first liquid level detection device can be a radar liquid level gauge or an ultrasonic liquid level gauge, and the second liquid level detection device can be a submersible hydrostatic liquid level gauge or a differential pressure liquid level gauge; the high-high liquid level switch and the low-low liquid level switch can be a float liquid level switch or an electrode liquid level switch, all of which can be used as alternative implementation methods.

[0038] When in use, the first liquid level measurement value Compared with the second liquid level measurement value On the data acquisition side, under the same zero-point reference, fixed installation offsets do not participate in subsequent cross-consistency judgments. Limit anchoring events bind limit heights to timestamps, facilitating the search for validity marker boundaries. Upon encountering an abnormal liquid level, the investigation can be directly traced back to the installation configuration and switch trigger records, shortening the troubleshooting process.

[0039] When the pump set switches between power frequency and variable frequency, there is a short-term asynchronous phenomenon in the command feedback. The motor current or power may be affected by changes in the power grid and mechanical load, resulting in transient spikes. If the acquisition side only marks the instantaneous value and not the state transition point, the subsequent steps of judging the liquid level response after the pump starts may mistakenly regard the start-up stage, in which no effective water is produced, as the pump has been effectively running, thus deviating from the reachability criterion.

[0040] The data acquisition side selects the minimum set of fields for pump unit operation information, which should include at least operating status, fault status, frequency or speed command feedback, motor current or power, number of start-stop cycles, and cumulative runtime. Then, the points of change in operating status, fault status, and frequency change from zero to non-zero are uniformly recorded as pump action events. Each pump action event is then compared with those in the same sampling period. , By binding to the same timestamp, step two can reference the pump action event time window without having to reverse-engineer the pump action time.

[0041] First, the acquisition side records the pump unit's operating and fault states using edge triggering. When a pump changes from stopped to running, running to stopped, normal to fault, or fault recovery, a pump action event is generated, and the event type and timestamp are retained, forming a pump action event sequence. Second, the acquisition side samples and writes the variable frequency drive (VFD) operating frequency or speed command and feedback within the same period, simultaneously writing the command value and feedback value within the same sampling period, along with the motor current or power for that period. This ensures that each sampling period provides a basis for determining whether the pump is effectively outputting power, serving as a time window for subsequent steps.

[0042] Pump unit operating information is converted into a traceable sequence of pump action events, starting with the first liquid level measurement. Compared with the second liquid level measurement value There are clear start and end boundaries on a time axis; frequency conversion commands and feedback are recorded in parallel in the same sampling period, so that subsequent steps can distinguish between the start-up stage and the effective output stage.

[0043] By aligning asynchronously sampled multi-source signals to the sampling period, short-term glitches are suppressed during the alignment process, ensuring that subsequent programs can read the signals in each sampling period. and Alignment value.

[0044] Non-contact level gauges and submersible hydrostatic level gauges have different output cycles, and communication link jitter causes the sampling time to be inconsistent. If the difference is calculated directly based on the most recent sampling value, the level change caused by the time misalignment will be interpreted as a divergence of the sensor, which will cause the signal quality mark to be reduced prematurely and affect the reliability of subsequent steps.

[0045] With sampling period A unified time scale is generated to determine the sampling time for each sampling cycle. An alignment value is generated at the sampling time of each liquid level signal. After the alignment value is generated, glitch removal is performed. Glitch removal adopts a continuous discrimination: if the jump occurs only in a very short time and meets the liquid level change envelope allowed by the pump action event, it is judged as a glitch and suppressed; if the jump still exists in multiple sampling cycles, it is retained and handed over to steps two and three for discrimination.

[0046] First, piecewise linear interpolation is used to map the original sampling sequence to a unified sampling time to obtain the aligned first liquid level measurement value. Align with the second liquid level measurement value Interpolation form:

[0047]

[0048] Where: sampling time The current sampling point on the continuous time axis; the original time. Less than And with The most recent raw sampling time; raw time Greater than and Align with the most recent original sampling time; align liquid level values : Bottom-top fluid level height; original fluid level value : The liquid level height output by the corresponding sensor;

[0049] right and A short-duration, persistent threshold is used to remove glitch data. The threshold is determined by the start-up and steady-state phases provided by the pump action event. The start-up phase allows for short-term fluctuations but requires the data to return to the envelope within two consecutive sampling periods. In the steady-state phase, adjacent differences of aligned values ​​should not repeatedly cross the validity boundary. After glitch removal... and These are used as preprocessed outputs for and , respectively.

[0050] In the sampling period Generate uniform sampling time Then, for each original sampling sequence, the interpolation endpoints are taken according to the following rules:

[0051] Left endpoint time Take the satisfaction The most recent original sampling time; the right endpoint time Take the satisfaction The most recent original sampling time; if in There are no endpoints on either side, or If the maximum allowable alignment gap duration is exceeded, the corresponding liquid level measurement value for that sampling cycle is deemed unusable, and the corresponding validity label is set to unusable. The positive real number is set by the control cabinet parameters; in engineering practice, it is determined by the communication cycle and the allowable signal interruption duration on site.

[0052] When using this method, the sampling time should be aligned with the interpolation, and the first liquid level measurement value should be used. Second liquid level measurement value Simultaneously readable, cross-consistency judgment is unaffected by sampling misalignment. Short-term spikes are suppressed through a continuous threshold, reducing abnormal spikes caused by communication jitter or transient echo mis-locking.

[0053] A single out-of-range judgment cannot cover common distortions in wet wells, such as sludge-covered hydrostatic diaphragms causing slow response, or non-contact level gauges drifting due to changes in the echo path caused by foam. These distortions may not cross the range in the short term, but rather continuously deviate from the true level. If the acquisition side does not provide traceable quality information, subsequent steps will find it difficult to distinguish the difference between the actual level change and the measurement quality degradation.

[0054] After deburring and Based on this, physical boundary verification and maximum rate of change verification are performed first to generate binary validity labels. and Subsequently, the four types of information—boundary violation, rate of change violation, diagnostic anomaly, and failure to update—were converted into signal quality labels. and The validity flag indicates whether the data from this channel has entered the calculation chain within the current sampling period. The signal quality flag indicates the strength of the data quality under the premise of usability and is bound to the liquid level measurement value for output, ensuring that step two can read the triplet data within the same sampling period.

[0055] right and Generate validity labeling tags respectively and The system is unavailable when the liquid level is below the physical lower limit but above the physical upper limit; it is also unavailable when the liquid level difference between adjacent sampling periods is greater than the envelope of the maximum rate of change. This unavailability is recorded along with the corresponding pump action event time window. A signal quality marker is generated based on the cumulative degree of the violation term, and the signal quality marker uses exponential decay.

[0056]

[0057] Where: signal quality label : Values ​​are taken from closed intervals This value indicates the signal strength within the sampling period; a smaller value indicates weaker quality. (Penalty coefficient) Positive real numbers, used to adjust for physical boundary violation terms. Attenuation intensity; penalty coefficient Positive real numbers, used to adjust for terms that violate the rate of change. Attenuation intensity; penalty coefficient Positive real numbers, used to adjust for diagnosing abnormalities or not updating violation terms. Attenuation intensity; boundary violation term : A non-negative real number used to characterize the severity of boundary breaches; the larger the breach, the larger the value. (Example: Physical lower limit liquid level) Physical upper limit liquid level Normalized out-of-bounds range;

[0058] rate of change violation : A non-negative real number used to characterize the degree to which the liquid level difference exceeds the envelope; the greater the exceedance, the larger the value. The maximum liquid level change rate limit and sampling period are used as the basis for the value. Formulate the allowable variation for a single period and normalize the excess portion; diagnose violations. : A non-negative real number used to characterize the cumulative degree of signal loss, over-range, or communication anomaly. The more times it occurs, the larger the value. The discrete cumulative quantity is directly generated from the sensor diagnostic bit and communication status. For example, signal loss, over-range, and communication anomaly each correspond to a discrete event count, which is accumulated according to the sampling period.

[0059] After the signal quality marker is generated, it is compared with... , as well as , The outputs are concatenated within the same sampling period, and the meaning of the fields remains unchanged across all sampling periods. In use, the validity flag distinguishes physical out-of-bounds and unreachable rates of change as unusable data, preventing subsequent steps from performing calculations on unusable sampling periods; the signal quality flag converts diagnostic anomalies and minor violations into continuous attenuation amounts, enabling subsequent steps to differentiate quality levels among usable data and form stable, weighted inputs.

[0060] Step 2: Convert the relationship between multi-source liquid level and pump action into reliability and fused liquid level. Step 3: Used to form a physical consistency state based on the well chamber volume-liquid level relationship. Step 4: Used to execute graded control and link limit interlocks based on the fused reliability and consistency state.

[0061] The liquid level detection in the wet well of a sewage lifting pump station is not only affected by the accuracy of the sensor, but also by the state of the medium on site and the pump operation process: First, foam and floating objects can cause non-contact level gauges to experience echo lock-in changes in a short period of time, which is reflected in the first liquid level measurement value. The readings may be subject to sudden spikes or short-term distortions; secondly, the sensor diaphragm of an immersion hydrostatic level gauge may be covered with sludge or grease, resulting in a second level measurement value. The response is sluggish or drifts slowly; furthermore, pump start-up, shutdown, check valve rebound, and backflow surges can cause non-monotonic changes in the wet well fluid level in localized periods. If the controller only judges based on instantaneous differences or single-point thresholds, it is easy to misjudge the fluid level disturbance caused by pump action as sensor distortion, or to misjudge sensor distortion as a true fluid level change. To ensure a stable input basis for the physical consistency verification in step three and the hierarchical control in step four, step two needs to use the first validity label generated in step one. Second validity label With the first signal quality mark Second signal quality mark Building upon this foundation, pump action reachability and cross-consistency are introduced, explicitly transforming availability, reliability, and interpretability into reproducible scores and labels, thereby avoiding the establishment of control links on a single, untraceable measurement.

[0062] Furthermore, usability and trustworthiness are expressed in layers: usability is determined by... , Given, credibility is determined by , This is given in conjunction with pump accessibility. The hierarchy revolves around eliminating unusable features and demoting usable but low-quality features, avoiding the misinterpretation of... , Binary removal and , The strength of continuous quality is mixed into the same judgment, thereby stabilizing the input space before subsequent reachability criteria are introduced. The key is to mark the first validity indicator. Second validity label As a hard gate, mark the quality of the first signal. Second signal quality mark As a soft gating mechanism, the effective boundaries of both are fixed within the same sampling period, so that the reliability calculation for each sampling period can be reproduced.

[0063] A confusing situation often occurs in wet well sites: sensor readings are not out of bounds, but diagnostic information indicates signal loss or communication abnormalities; in this case, the liquid level measurement value may remain unchanged for several sampling periods. If not checked first... , Hard gating is implemented, and the unchanged old value is mistaken for stable liquid level, which in turn creates false consistency in the pump operation reachability judgment.

[0064] Another scenario is that the reading is available but the quality is degraded, for example... Gradual decay occurs; if it is not used as a soft-gated input before accessibility judgment, the accessibility penalty will be forced to bear the entire burden of explaining quality degradation, causing the penalty threshold to be adjusted too sensitively, thus triggering deweighting even during normal liquid surface disturbances. Therefore, hierarchical gating shaping is performed first, so that subsequent penalty quantities only explain mechanism consistency and do not mix in usability loss.

[0065] Within each sampling period, the controller first reads the first validity flag. With the second validity label mark When a certain signal's validity flag is marked as unavailable, that signal will not participate in cross-consistency comparison and fusion level generation within the current sampling period, and the reason for its unavailability will be recorded in step four. Subsequently, the controller reads the first signal quality flag. With the second signal quality mark This is used as a quality prior input into the subsequent credibility score, and a consistent update rhythm is set for the cases of continuous no update and intermittent recovery: when there is no continuous update, the quality prior remains in a decaying state, and after the update is resumed, the quality prior recovers in a smooth manner, avoiding large fluctuations in credibility between two adjacent sampling periods, thus providing a basis for the subsequent fusion level stability.

[0066] First, the controller during the sampling period Under the time reference, the first validity marker will be marked. With the second validity label mark As a hard gating condition, it is written into the credibility calculation process: when If unavailable, the reliability score of the first liquid level will be determined within this sampling period. The base value is set to zero and the cross-connection consensus participation right of that path is frozen; when If unavailable, freeze the second liquid level reliability score in the same manner. The base value.

[0067] Under hard gating, the quality of the first signal is marked. Second signal quality mark As a soft-gated prior, and with smooth updates: when diagnostic information experiences continuous signal loss or communication problems, maintain... or The decay state and duration interval are recorded; when the diagnostic information returns to normal and or When it becomes available again, or The ascending step size approaches the same as the current diagnostic state, and the soft-gated prior does not mutate during the recovery phase.

[0068] When in use, hard gating from , Within the boundary of the unavailable sampling period, eliminate pseudo-stability after the old value stagnates; soft gating from , The prior art explicitly causes a decrease in quality, and the subsequent reachability penalty only explains the inconsistency in the mechanism without any loss of usability; the smooth update suppresses the abrupt change in the confidence input within adjacent sampling periods, providing a stable premise for the stability of the subsequent fused liquid level.

[0069] Furthermore, the pump operation information is transformed into a time window constraint for the expected liquid level response. The deviation of the actual liquid level change from this constraint is then transformed into a non-negative penalty, which is then aggregated in a smooth manner to form a unified penalty for the reliability score. The key is that the penalty is not given as a binary conclusion based on whether the constraint is met, but rather as a continuous expression of the degree of deviation. This creates an interpretable transition range between the natural fluctuations of the wet well liquid level and the actual distortion of the sensor, avoiding frequent switching of simple threshold logic near the boundary.

[0070] The pump operation reachability criterion is necessary because sensor distortion is often highly correlated with the pump operation period: after the pump starts, if the pump has established an effective water outlet channel, the wet well level should show an explainable downward trend within a limited time; if the level does not drop or rises instead, there is one or a combination of level measurement distortion, pump failure to effectively discharge water, and abnormal backflow.

[0071] On the other hand, after the pump stops, if there is no external venting path, the wet well level should not continue to drop in a short period of time; if it does drop continuously, it may indicate one or a combination of level measurement distortion, check valve leakage, or pipeline backflow. and The difference cannot distinguish the above mechanism. Therefore, the pump action event is used as the time window anchor point, the trend that should occur is transformed into a penalty quantity, and then combined with the cross-consistency penalty quantity into the score.

[0072] Within each sampling period, the most recent start-up event and the most recent shutdown event are read from the pump action event sequence in step one, and the sampling period is used as the basis for further analysis. The observation window for particle size distribution during startup and shutdown is set. Subsequently, the first liquid level measurement value is recorded. Compared with the second liquid level measurement value Three types of violations are calculated: the first type reflects insufficient cumulative decrease within the start-up observation window; the second type reflects abnormal cumulative decrease within the shutdown observation window; and the third type reflects the degree of cross-consistency deviation between the two liquid levels within the same sampling period. All three types of violations are non-negative and can be physically interpreted as the degree of deviation. Finally, the three types of violations are aggregated into an reachability consistency potential using a smoothing maximum operator, and based on this potential, [the following is used to...]. and Continuous weight reduction is performed. The smoothing maximum operator is used because wet well fluid levels may exhibit multiple slight deviations simultaneously at certain times. Directly taking the maximum value would be sensitive to a single spike, while direct summation would amplify multiple slight deviations into severe deviations. The smoothing maximum operator provides a controllable transition between the two, making the score insensitive to spikes but sensitive to continuous deviations.

[0073] First, three types of violation variables are constructed, and an reachability consistency potential is formed. For any liquid level measurement, a starting violation variable is generated. Shutdown violation amount Contradictions :

[0074] Start-up violation is used to characterize the degree of insufficient cumulative drop in liquid level within the start-up observation window; shutdown violation is used to characterize the degree of abnormal cumulative drop in liquid level within the shutdown observation window; and divergence violation is used to characterize... and When both are available, the deviation of the two liquid level differences exceeds the soft threshold, and a strong penalty interval is entered when the difference crosses the hard threshold. Then, a smoothing maximum operator is used to form an reachability consistency potential:

[0075]

[0076] Where: reachability consistency potential : A non-negative real number used to aggregate multiple types of violations into a single penalty potential; smoothing coefficient : Positive real number, used to adjust how close the smoothing maximum operator is to the maximum value. The larger it is, the closer it is to taking the largest of the three. The smaller the value, the closer it is to the smoothed average of the three; initiation violation amount : A non-negative real number used to characterize the degree of deviation when the cumulative drop in liquid level is insufficient within the initial observation window; if the cumulative drop in liquid level is insufficient within the initial observation window, a deviation occurs; if the liquid level rises, the deviation increases. The cumulative drop is represented by... It means that, among them This serves as the starting point for initiating the observation window. The expected minimum descent envelope is obtained by multiplying the lower bound of the achievable descent rate by the running time. The lower bound of the achievable descent rate can be obtained from the commissioning parameter table by indexing the frequency conversion feedback frequency.

[0077] Stoppage violation : A non-negative real number used to characterize the degree of deviation of the cumulative abnormal drop within the shutdown observation window; if the cumulative drop in liquid level exceeds the allowable envelope within the shutdown observation window, a violation occurs. The allowable envelope can be obtained by multiplying the upper bound of the allowable drop rate during shutdown by the duration of shutdown.

[0078] Disagreement violation quantity : A non-negative real number used to characterize the degree of deviation between the two liquid levels in cross-consistency and can map the soft threshold and hard threshold range; normalize the excess amount after the difference between the two liquid levels exceeds the soft threshold, and allow the difference to continue to increase linearly after exceeding the hard threshold so that the penalty increases as the difference increases.

[0079] Serial number symbol : is a positive integer in the set of integers Used to distinguish the first liquid level measurement value Compared with the second liquid level measurement value The penalty potential calculation channel; sampling time Values ​​are taken according to the sampling period. Discretized time series, used to indicate the calculation time corresponding to the current sampling period;

[0080] Secondly, the reachability consistency potential is introduced into the reliability score and coupled with hard gating and soft gating. For any level measurement, the level reliability score is generated using the following formula.

[0081]

[0082] Liquid level reliability score : Values ​​are taken from closed intervals Used to characterize at the sampling time The reliability of the liquid level measurement is indicated by a higher value; validity is also indicated by a label. : Values Used for hard gating. for The road is unavailable at that time. A value of 1 allows entry into the scoring process; signal quality flag. Closed interval It is used for soft-gated priors to characterize the impact of discrepancies between diagnostic information and preprocessing on quality.

[0083] Penalty coefficient : is a positive real number used to adjust the strength of the reachability consistency potential's effect on score decay. The larger the value, the more pronounced the weight reduction caused by slight deviations; reachability consistency potential. : Non-negative real numbers, used to aggregate multiple types of violations and include them in the score decay term; ordinal number symbol : Values Distinguish between the first and second scoring channels; sampling time The value range is based on the sampling period. Discretized time series;

[0084] When in use, the pump action reachability is transformed into a continuous penalty potential and enters the scoring decay term, so that the mechanism deviations such as the pump has started but the liquid level does not drop, or the pump has stopped but the liquid level continues to drop can be expressed at the scoring level; the cross consistency deviation enters the same penalty potential with the divergence violation amount, avoiding making the difference judgment an isolated threshold, so that the subsequent fusion liquid level generation has a consistent weight basis.

[0085] Furthermore, to , Multi-source input compression into fusion liquid level and will , The weighting significance is explicitly implemented in the fusion process, ensuring that the fusion result is consistent with the scoring logic.

[0086] The fusion liquid level is not simply taken as the average or from a single path, but rather within the same sampling period... , As a weight, it forms a continuous output and provides explicit division-to-zero protection and output constraints for boundary cases where neither side is trustworthy, providing a calculable input for introducing the well chamber volume-level relationship in step three.

[0087] If the fusion process is separated from the scoring process, contradictions will arise on-site, such as the scoring process deeming the data unreliable while the fusion process still heavily utilizes that data, making the consistency deviation check in step three unexplainable. On the other hand, when or If the fusion level is still forcibly output in a fixed form when the channel is unavailable within a certain sampling period, it is easy to incorporate the old values ​​of the unavailable channel into the fusion result, causing fusion level lag. Therefore, hard gating, soft gating, and penalties are necessary. , The results are directly substituted into the fusion formula, and a zero-division protection coefficient is introduced to ensure that the fusion liquid level still has a clear numerical generation rule in extreme cases.

[0088] Within each sampling period, the first liquid level confidence score is read first. Reliability score of the second liquid level And confirm that the rating has been included. , The hard-gating effect. Subsequently, the controller... , Generate fused liquid level for weighting And will integrate credibility The output is co-located to the extent that the fusion result can be referenced in subsequent steps.

[0089] The generation of credibility fusion does not pursue complex models, but emphasizes interpretability: when and All are relatively high and and If the discrepancy is small, the fusion credibility is high; if only one score is high, the fusion credibility decreases but is not zero; if both scores are low, the fusion credibility enters a low position and triggers a strong prompt for the distortion type label, enabling step four to enter a more conservative control level.

[0090] First, generate fusion liquid levels by weighting based on credibility. And set the division by zero protection factor.

[0091]

[0092] Where: fusion liquid level : Values ​​are the liquid level height range from the bottom of the well to the top of the well; First liquid level reliability score : Value Used for the first liquid level measurement value Weighting in fusion; second liquid level credibility score : Used for the second liquid level measurement value Weighting in fusion; first liquid level measurement value : The range of liquid level height from the bottom of the well to the top of the well; the second liquid level measurement value Values ​​are taken from the liquid level height range from the bottom of the well to the top of the well; zero is divided by the protection coefficient. : A positive real number much less than 1, used to avoid the denominator being zero, which would make the fusion liquid level uncalcible; sampling time : based on sampling period Discretized time series;

[0093] Fusion credibility The outputs are parallelized within the same sampling period, and the parallel outputs are constrained by both the total score and the discrepancy violation. and The total volume is large, and the volume of discrepancies and contradictions is also large. Take the higher value; only one path's score is greater than the other path's score. Taking the median value indicates that the fusion is closer to a single path; both paths have low scores, large discrepancies and violations, and are in a strong penalty range (at a low level). Entering a lower level indicates that step four involves entering a more conservative level.

[0094] The controller can use a lookup table method to map the score sum-disagreement violation interval to... In the lookup table method, the entries are entered on-site by the debugging personnel in a fixed order to form a version record, thus avoiding rule drift.

[0095]

[0096] Fusion credibility Closed interval Its function is to characterize the fusion liquid level. This can be used to assess the reliability of subsequent steps three and four; First liquid level reliability score : The second liquid level credibility score serves as a fundamental component of the overall credibility assessment. : It serves as a fundamental component of fusion credibility; fusion divergence penalty coefficient : is a positive real number, which adjusts the attenuation of the divergence violation quantity on the fusion reliability; divergence violation quantity : Non-negative real number, used to characterize the degree of divergence between the two liquid levels and included in the fusion reliability penalty term; sampling time : based on sampling period Discretized time series;

[0097] When using, the fusion liquid level With rating , Coupling within the same cycle ensures consistency between the fusion result and the scoring logic, preventing contradictory interpretations between the scoring and fusion when referenced in step three. Fusion credibility. It is generated using an interpretable lookup table method, enabling subsequent hierarchical control to have reproducible trigger values ​​without relying on external models.

[0098] The challenge of on-site maintenance lies in the fact that even if the system detects inconsistencies between the two liquid levels, maintenance personnel still need to know which cause to check first. If only the output difference exceeds the limit, maintenance personnel may need to repeatedly open the well cover to check, and they cannot determine whether to clean the probe first, check the hydrostatic diaphragm first, or check the communication wiring first. On the other hand, the selection of the control level also depends on the distortion mechanism: jump-type distortion is more suitable for short-term freezing and weighted suppression, drift-type distortion is more suitable for continuous weight reduction and prompting for cleaning or calibration, and stuck-type distortion is more suitable for prompting for obstruction and fixed-point checks. Therefore, using distortion type labels as a bridge between control and maintenance allows the graded control in step four to be based not only on numerical magnitude but also on the mechanism category.

[0099] Therefore, within each sampling period, the first signal quality marker is compared first. With the second signal quality mark The change pattern: If there is a short-term sharp decline followed by a quick recovery, and at the same time the initiation of the violation and divergence violation values ​​increases briefly, it tends to be judged as a jump type; if It is in a state of decay for a long time and the shutdown violation occurs repeatedly. If the difference exhibits hysteresis across multiple sampling periods, it is likely to be classified as drifting or stuck. or If the device repeatedly switches from available to unavailable, it is identified as a lost device and a prompt will be made to check the communication link and power supply.

[0100] Subsequently, the judgment result is written into the distortion type label, and the corresponding maintenance statement is output within the same sampling period. The maintenance statement is described in the form of specific action-object-location. For example, the maintenance statement for a non-contact level gauge refers to cleaning the probe surface and checking for installation obstructions, while the maintenance statement for an immersion hydrostatic level gauge refers to cleaning the diaphragm, checking the ventilation path, and verifying the fixed point height. Finally, the controller associates the distortion type label with the current... , Bind and write trace records so that step four can reference tags and display the clear reason on the alarm interface when switching control levels.

[0101] Among them, the distortion type label adopts a priority-based writing mechanism: when or When the condition is unavailable and persists, prioritize writing to the lost type; when , Available but or Write a jump type when there is a short-term, rapid decay accompanied by a sudden increase in the start-up violation within the pump action event time window; when or When the level remains low for an extended period and is accompanied by accumulated violations, it is classified as drift type; when the level difference remains close to zero for an extended period and is significantly inconsistent with the pump's operational reachability, it is classified as stuck type.

[0102] This priority mechanism ensures the uniqueness of labels within the same sampling period, preventing multiple contradictory labels from being generated simultaneously within a single sampling period. Uniqueness: Only one distortion type label is output in any given sampling period; Priority: The missing type is determined first. or If the system is unavailable or communication is abnormal, then determine the jump type (single-cycle difference exceeds the jump threshold and returns within a limited period), then determine the stuck type (pump effective output is established but liquid level difference is close to zero for a long time), then determine the drift type (discrepancy violation and cumulative deviation). (Indicating continuous growth in the same direction); retention and release: Once a tag is written, it must be retained for at least a minimum retention period. Retrieval requires the corresponding trigger condition to disappear and the fusion credibility to be met. Conditions for continuous recovery.

[0103] Secondly, a mapping table is fixed between distortion type labels and maintenance statements, and this table is saved in read-only format in the controller, supporting version updates. Each column of the mapping table contains four statements: the object to be inspected, the location to be inspected, the visible phenomenon, and the suggested action. Maintenance personnel can perform on-site inspections without interpreting the scoring formula. With labels and maintenance statements fixed in a corresponding mapping table, maintenance actions have shifted from guessing the cause based on numerical values ​​to looking up the object based on the label, consistent with the control level selection interface in step four. Uniqueness and priority mechanisms ensure that labels do not conflict within the same sampling period, and traceability records can be used for event playback and responsibility interface display.

[0104] Step 3: The actual evolution of the wet well level is limited by the inflow, outflow, and well chamber cross-section; the output of the level detection device can only reflect the measurement results in the height domain. Step 2 outputs the fused level. It can suppress single-path jumps, but if common-mode drift, cross-sectional abrupt change conversion error, or pump group drainage capacity changes occur, the height domain remains stable but the mechanism is not closed.

[0105] Therefore, the fusion liquid level Through well chamber volume mapping The data is converted to a volumetric domain, and the pump unit operating information is solidified into a drainage volume change rate. With the rate of change of inflow water volume Then calculate the physical consistency deviation. Consistency status; when the consistency status indicator height domain input no longer has interpretability, then extrapolate the soft-sensor liquid level. The boundary is anchored using a level switch, so that step four obtains traceable input data.

[0106] Specifically, the wet well geometry and pump output are uniformly transformed into a volume domain expression, so that subsequent deviation calculations no longer rely on the intuitive height domain threshold. A lookup table structure of height nodes and volume nodes explicitly writes cross-sectional changes into the controller, ensuring that volume conversion remains traceable at well chamber steps and pipe intrusion heights. Simultaneously, soft-sensing liquid level measurement... The height needs to be calculated from the volume domain, therefore the volume inverse calculation operator must be made public. The specific implementation process.

[0107] First, the controller generates height and volume node tables: Prioritizing the entry of key heights and cross-sectional dimensions of the well chamber based on as-built data; when data is missing, maintenance personnel measure the bottom of the well, the bottom of the inlet pipe, the wall penetration height of the outlet pipe, and the reference surface of the well top at the inspection window, and then enter the cross-sectional dimensions of each height interval into the control cabinet. The controller then accumulates these dimensions to obtain the volume nodes. To ensure repeatable node table entry, the control cabinet provides sequential prompts on the entry interface in ascending order of height, and immediately displays the calculated cross-sectional area of ​​the interval after each entry. Maintenance personnel can then immediately verify whether there are any errors such as treating millimeters as meters or inner diameters as outer diameters.

[0108] Secondly, well chamber volume mapping Piecewise linear interpolation is used to solidify deterministic rules:

[0109]

[0110] Where: well chamber volume mapping : A mapping from non-negative real numbers to non-negative real numbers; liquid level height : This represents the liquid level height range from the bottom to the top of the well; left end height node : A node in the height node table that satisfies Used for piecewise interpolation of the left endpoint; right height node. : The value is a node in the height node table and satisfies Node volume value : A non-negative real number representing the volume of the well chamber corresponding to the left-hand height node; node volume value : is a non-negative real number used for the well chamber volume corresponding to the right-hand height node;

[0111] Among them, the volume inverse calculation operator It uses reverse search of the volume node table + piecewise linear interpolation: first define the volume interval and then calculate the height. The interval positioning is a binary search to avoid the delay of sequential traversal. The piecewise linear interpolation is a linear function, which makes it easier for maintenance personnel to check the calculation process. The input process strictly checks that the height node and the volume node are strictly incremental.

[0112] When in use, well chamber volume mapping Transitional changes in cross-sections are formalized into node table information, ensuring that volume conversion and consistency verification maintain the same baseline at points of cross-sectional change. Monotonicity checks proactively expose input errors, preventing subsequent deviations. Unexplainable perturbations are introduced by incorrect mapping.

[0113] Furthermore, integrate pump unit operation information with reliability. Coupling, forming the rate of change of drainage volume With the rate of change of inflow water volume An feasible estimation process ensures that consistency deviations reflect both measurement issues and the mechanistic differences between pump output and inflow variations. Discharge volume change rate. Using a tiered approach that prioritizes instrumentation and uses table lookup as a fallback, the rate of change in influent volume is... Only when the pump stops and the fusion credibility is Continuous and reliable in-window updates.

[0114] Drainage volume change rate The estimation is graded and solidified according to the available signals on site: if the effluent flow rate signal is available, it is directly converted into the rate of change of drainage volume. If the outlet flow rate signal is unavailable but the outlet pressure signal is available, the rate of change of drainage volume is obtained by indexing the pump performance curve table using frequency conversion feedback and outlet pressure. The index uses bilinear interpolation to ensure continuous output as frequency and pressure change between table nodes; if the outlet pressure signal is unavailable, the drainage volume change rate is obtained by indexing the frequency-power-drainage volume change rate table using frequency conversion feedback and motor power. The index uses piecewise linear interpolation to keep the output continuous when the power node changes.

[0115] To absorb long-term deviations caused by pump wear, impeller fouling, and valve opening variations, the controller allows slowly updated correction factors to be superimposed on the table output. These correction factors are only updated when the consistency is normal and the fusion confidence level is [value missing]. Continuous and reliable window execution, and write the correction factor update time to the retrospective record.

[0116] Secondly, the rate of change of water volume Generated only within the pump shutdown observation segment: The controller filters the fusion confidence level within this observation segment. A continuous and reliable sampling period will correspond to the fusion liquid level. Through well chamber volume mapping Convert to a volume sequence, then divide the difference between adjacent volumes by the sampling period. Candidate values ​​were obtained, and occasional liquid level fluctuations were suppressed using the median rule; the resulting influent volume change rate... Write in increments with a fixed step size to avoid sudden changes within a single sampling period.

[0117] During use, the rate of change of drainage volume The hierarchical path covers different instrument configurations, ensuring that the consistency calculation framework does not change the input interface when signals are missing; influent volume change rate Pump shutdown and fusion reliability Dual constraints reduce the contamination of the influent backflow by the distortion window and stabilize soft measurement input.

[0118] Furthermore, the fusion liquid level is controlled within the volumetric domain. Changes and rate of change of drainage volume , rate of change of water volume The mechanism limitations were compared to form a bias. With consistency, and in the cumulative deviation Output soft-sensor liquid level when triggered .

[0119] The height change is mapped via the well chamber volume. Converted to volume change, and then compared with the expected net volume change rate (inflow volume change rate). Subtract the rate of change of drainage volume By comparing quantities of the same dimension, false deviations can be avoided at abrupt changes in the height domain. To ensure that the state output closely matches the start-up and shutdown transients of the wet well, the state persistence is bound to the pump's operating time window, preventing unnecessary upgrades from being triggered during the start-up transition.

[0120] First, the controller defines physical consistency deviations in the volume domain. :

[0121]

[0122] Where: Physical consistency deviation : A real number representing the difference between the observed rate of change and the expected net rate of change in the volumetric domain; well chamber volume mapping : A mapping from non-negative real numbers to non-negative real numbers; fused liquid level : The liquid level height range from the bottom to the top of the well, used as the liquid level input for deviation calculation; sampling period. : Positive real number, used for discrete difference time scale and consistent with step one;

[0123] Rate of change of inflow volume : A non-negative real number used to characterize the inflow intensity calculated and gradually written back during the pump shutdown window; the rate of change of discharge volume. : A non-negative real number used to characterize the drainage capacity obtained from instruments or tables; sampling time The value range is based on the sampling period. Discretized time series, used to indicate the current calculation time.

[0124] Controller based on The amplitude range and consistency status of the persistence counter output: The amplitude range is jointly determined by the estimated source level, the pump operation stage, and the effective volume range of the wet well. The controller maps the estimated source level to the allowable bandwidth using a lookup table, and then briefly widens or tightens the bandwidth based on the pump operation stage. The persistence counter uses a sampling period... To achieve step-by-step accumulation, the counter's increment / decrement rules are tied to the pump action event time window to avoid misjudging the transient volume rebound during startup as a long-term deviation. When the pump is in the startup transition phase, only the decay count is updated without escalating the status. Once the pump enters the stable phase, the system outputs "Normal Consistency," "Consistency Warning," and "Severe Consistency" according to the continuous rules, and records the drainage volume change rate in the status explanation field. The estimation source level allows for a traceable explanation of the same bias under different estimation uncertainties.

[0125] During use, deviation Calculated in the volume domain and mapped from the well chamber volume. By absorbing cross-sectional changes, consistency judgments remain interpretable at steps and intrusion heights. The consistency state is tied to the pump's operating time window, reducing interference from start-stop transients on the state output and making the state more closely aligned with the wet well mechanism.

[0126] Furthermore, in the deviation Construct cumulative deviation based on To capture slow-changing deviations, and in cases of severe or cumulative deviations in consistency. Triggering Soft-Measuring Liquid Level The deduction is based on the net volume change rate (inflow volume change rate). Subtract the rate of change of drainage volume As a driving force, avoid using distorted fusion liquid levels. The short-term fluctuations were then corrected by using a high-high liquid level switch and a low-low liquid level switch to anchor the simulation at the boundary.

[0127] Cumulative deviation An evolutionary constraint is adopted that allows for accumulation of the same sign beyond the allowable band and a decrease in the allowable band during regression. The incremental and decremental updates are bound to the pump action time window: only decrement is allowed in the start-up transition phase, while both increment and decrement are allowed in the stable phase. This makes the system sensitive to continuous deviations but insensitive to transient disturbances.

[0128] Define the allowable deviation threshold and the cumulative threshold as control cabinet parameters, and update the cumulative deviation as follows: :

[0129] when When the deviation exceeds the allowable threshold, the cumulative deviation increases; when Upon returning to the allowable band, the cumulative deviation decreases by a fixed attenuation amount, but not less than 0. A smoothing threshold function can be used to avoid boundary jitter caused by hard line transitions, for example:

[0130]

[0131] Cumulative deviation : A non-negative real number, used to characterize the degree of continuous accumulation of deviation exceeding the limit; sampling period : A positive real number, used as the time scale for cumulative updates; physical consistency deviation : is a real number, used to accumulate input; tolerance threshold for deviation. : is a non-negative real number, used to define the boundary of the tolerance band for deviation. Exceeding this value will trigger accumulation; smoothing coefficient : Positive real number, used to adjust the steepness of the transition in the smoothing threshold function. The larger the value, the closer it is to the hard threshold;

[0132] Attenuation : A non-negative real number, used to apply a downward trend to the cumulative deviation when the deviation is brought back or slightly exceeds the limit, to prevent... Permanent accumulation; sampling time : based on sampling period Discretized time series, used as the current update time.

[0133] The consistent state can be output according to the following rules: If Exceeding the severe deviation threshold or If the cumulative threshold is exceeded, the output consistency is severely compromised; otherwise, if... Exceeding the warning deviation threshold or If the cumulative warning threshold is exceeded, a consistency warning will be output; otherwise, a consistency normal warning will be output.

[0134] When soft sensing is enabled, the controller simulates the soft sensing liquid level. :

[0135]

[0136] Where: soft liquid level measurement : The range of liquid level height from the bottom to the top of the well; volumetric inverse operator The mapping from non-negative real numbers to liquid level height ranges is achieved using reverse search and piecewise interpolation.

[0137] For any volume value First locate the interval so that Then calculate in reverse:

[0138]

[0139] volume : Values ​​range from non-negative real numbers; function is to inversely calculate inputs; volume inverse calculation operator The range of values ​​is a mapping from non-negative real numbers to the liquid level height interval. Its function is to convert the volume back to the liquid level height. Other symbols are the same as above. definition.

[0140] Well chamber volume mapping : A mapping from nonnegative real numbers to nonnegative real numbers; sampling period : Positive real number, convert the net rate of change into a discrete volume increment and maintain consistency with step one; inflow volume change rate : A non-negative real number used to deduce input terms, representing the intensity of inflow; rate of change of drainage volume. : is a non-negative real number used to deduce the input term and characterize the drainage capacity; sampling time : based on sampling period Discretized time series, used to indicate the current simulation moment;

[0141] The soft-measurement liquid level obtained from the simulation Entering extreme clamping correction: When the consistency status returns from severe to warning or normal, the controller will soft-measure the liquid level. With fusion liquid level A consistency alignment is performed, and the difference between the two is used as the initial bias for subsequent deductions, thereby avoiding input jumps caused by sudden switching back after a long period of control. If the high-high liquid level switch is triggered, the clamping condition is set to be no lower than the corresponding height; if the low-low liquid level switch is triggered, the clamping condition is set to be no higher than the corresponding height. The clamping event is written into the traceability record for reference in step four.

[0142] During use, cumulative deviation Enhanced identification of persistent slight deviations makes common-mode drift and long-term contamination more likely to trigger connection conditions; soft-sensor liquid level measurement. From the rate of change of drainage volume With the rate of change of inflow water volume Driven and via volume inverse operator The reverse calculation ensures that the generated path is consistent with the water inlet and outlet mechanism of the wet well, thus preventing the continued transmission of distorted measurements.

[0143] Step 4: The risk of wet well environmental control is not solely due to level error, but rather the simultaneous decline in measurement reliability and the degree of closure of the inflow / outflow mechanism: [This is related to] the fusion of liquid level... Physical consistency deviations may occur due to foam, dirt, or communication jitter, resulting in jumps or slow drifts. and cumulative deviation If the pump set's drainage, inflow changes, and well chamber volume evolution are not closed-loop and a single regulation is still used, it will cause frequent pump set start-ups and shutdowns, insufficient or excessive pumping, and in extreme cases, trigger overflow and dry running protection.

[0144] Therefore, based on the credibility of integration The consistency status has two thresholds; above the threshold, the fusion level is used. Achieving a balanced regulation combining static and dynamic forces, using a soft-touch liquid level sensor below the threshold. The trend and level switch boundaries achieve conservative control, maintaining the interlock priority between the high-high level switch state and the low-low level switch state under any circumstances.

[0145] Specifically, the high-high liquid level switch state and the low-low liquid level switch state are placed before any adjustment action, and the hierarchical determination is changed from instantaneous threshold to continuous evidence + maintenance constraint.

[0146] First, establish the boundary semantics of the control actions: a high-high level switch indicates that drainage must be prioritized, and a low-low level switch indicates that the pump must be stopped first to prevent dry running; within these boundary semantics, the regulation strategy is allowed to take effect. This structure ensures that subsequent level switching does not change the determination criteria of the safety boundary, thereby avoiding the same level switch being assigned different meanings at different levels.

[0147] In each sampling period Initially, the controller reads the states of the high-high level switch and the low-low level switch, using these as the primary criteria for judgment within this sampling period: when the high-high level switch is triggered, the controller outputs a pump start or pump increase command and maintains this until the high-high level switch state is deactivated and the preset holding time is met; when the low-low level switch is triggered, the controller outputs a pump stop or pump start prohibition command and maintains this until the low-low level switch state is deactivated and the preset holding time is met. Subsequently, the controller reads the extreme clamping event output in step three: if the extreme clamping event occurs, the soft-measured liquid level... It is regarded as a reference quantity that has been anchored to the boundary, and the anchoring action shall not be interpreted as a measurement anomaly during the current sampling period, so as to avoid repeated takeover and repeated retreat near the boundary.

[0148] In an equivalent implementation, the high-high level switch and the low-low level switch can be either float level switches or electrode level switches; the interlock output can act on the inverter enable terminal or the power frequency contactor coil terminal; the holding logic is implemented using a control cabinet timer and synchronized with the sampling period. Alignment.

[0149] In use, the interlock priority takes precedence over any other level of decision execution, ensuring consistent semantics between overflow prevention and dry-run prevention boundaries under all operating conditions; extreme clamping events incorporate the boundary anchoring process into the control chain, enabling soft-sensor level measurement. The evidence remains traceable near the boundary; the duration of the maintenance suppresses repeated actions at the moment of interlock release, ensuring a continuous connection between interlock exit and subsequent adjustments.

[0150] Furthermore, enhance the credibility of the fusion. As a reliable threshold for measurement, the consistent state is used as a mechanistic closure threshold, and a minimum retention time and unidirectional migration rule are introduced to ensure that level switching is based on continuous evidence. This structure utilizes the synergy of the outputs of steps two and three: when both point to risk, it quickly enters a conservative level; when there is a short-term divergence, it first maintains the level and waits for evidence to accumulate, reducing erroneous switching.

[0151] In each sampling period Internal read fusion reliability Consistency status, physical consistency deviation and cumulative deviation When the consistency state is normal and the fusion credibility is... When the preset trust conditions are continuously met, the system enters the dynamic balance level; when the consistency status is in the warning or fusion trust level... When the level falls within the preset range, it enters a degraded adjustment level, expands the target level band, and extends the pump confirmation time to reduce the frequency of actions; when the consistency status is severe or cumulative deviation... Continuous growth and fusion credibility If the preset reliability conditions are not met continuously, the system will enter the conservative control level and the soft-sensor liquid level will be switched off. As a trend reference, the fusion liquid level Reduced to an alarm display reference value. A minimum hold time is applied for each level change; within this hold time, migration to a safer level is allowed, but regression to a more aggressive level is not permitted. After the hold time ends, if the consistency status and fusion confidence level... If the rollback conditions are still met, then a step-by-step rollback will be allowed.

[0152] When using it, integrate credibility. The formation of a double threshold with the consistency state ensures that the hierarchical decision is simultaneously constrained by measurement reliability and mechanism closure; the minimum hold time and unidirectional migration suppress hierarchical oscillations, so that the control action sequence has a continuous interpretation chain during playback.

[0153] Degradation adjustment retains fusion liquid level The adjustment input is reduced, but the intensity of the action is decreased. When returning to the dynamic balance level, the frequency setpoint of the inverter operation is gradually returned according to the preset change rate limit to avoid the frequency jump at the moment of level return, so as to make the transition between dynamic balance and control smooth.

[0154] Furthermore, within the dynamic equilibrium level, the fusion liquid level The main control output is the variable frequency drive frequency setpoint. When the variable frequency drive capacity is insufficient, the pump is increased after a confirmation period; when the liquid level drops and the operating intensity decreases, the pump is decreased after a confirmation period. Within the degraded regulation level, the liquid level is still adjusted according to the setpoint. While maintaining the primary control variable, the action density is reduced through a wider target band, longer confirmation time, and tighter integral limit. The action chain is linked to the pump start-stop transition: within the preset transition section after pump start-up, the controller only allows the frequency setpoint to increase monotonically and prohibits triggering the pump reduction confirmation timer; within the preset transition section after pump stop-up, it only allows the frequency setpoint to decrease monotonically and prohibits triggering the pump increase confirmation timer. This isolates short-term fluctuations caused by check valve rebound and backflow surge within the transition section, preventing start-stop disturbances from being mistaken for regulation deviations.

[0155] When in dynamic equilibrium, the controller determines the liquid level based on the fusion liquid level. The deviation from the preset liquid level target band generates the variable frequency operating frequency setpoint. The proportional-integral-derivative control uses forward Euler integral and two-point difference differentiation, and sets integral and derivative limits. When the frequency setpoint reaches the upper limit and the liquid level is integrated... When the stable section remains above the upper boundary of the target band, the controller triggers a pump start confirmation timer. After the confirmation timer expires, the standby pump is put into operation. Based on the number of start-stop cycles and the cumulative running time collected in step one, the controller selects a rotation order from the available pump groups to give the long-running pump groups a rest window. If the operation feedback is not established within the preset confirmation time after the pump start command is issued, the pump is marked as a fault isolation object and the next candidate pump is assigned to take over the pump start task.

[0156] When the fusion liquid level When the frequency drops and remains below the target band midline during the stable phase, and the frequency setpoint is near the lower limit, the controller triggers a pump reduction confirmation timer. After the confirmation timer expires, one mains frequency pump is deactivated or a standby pump is shut down. The deactivation priority is given to pump groups with a high number of start-stop cycles and a long cumulative operating time, preventing wear differences from accumulating on a single pump over a long period. If the shutdown feedback is not confirmed within the preset confirmation time after the deactivation command is issued, the pump remains running and is included in the fault isolation alert to prevent the risk of dry running at low liquid levels due to shutdown failure. Each pump has a minimum operating time and a minimum shutdown time set, and a rate of change limit is set for the frequency setpoint to ensure that the frequency varies across multiple sampling periods. The internal changes are gradual to avoid repeated check valve activation. The pump addition confirmation timer only activates when the consistency status is normal or warning and the fusion confidence level is reached. Starts when the pump is in the available range; the pump reduction confirmation timer starts only when the pump set has met the minimum running time and the low-low level switch has not been triggered; any confirmation timer is reset when the level changes or the interlock is triggered.

[0157] When in the degraded regulation level, the controller expands the target band and extends the pump increase and decrease confirmation time, while tightening the integral limit to reduce start-stop operations during measurement fluctuations or mechanism warnings; if the distortion type label points to the jump type, then the fusion liquid level... Apply a short-term freeze window, within which only actions moving in the safe direction are allowed to be output, and mark the start and end timestamps of the freeze window in the traceback record.

[0158] When using, the fusion liquid level The variable frequency drive and the confirmed timing pump activation / deactivation form a continuous action chain, making the liquid level return process reproducible and independent of instantaneous fluctuations. Minimum operating time, minimum downtime, and rate of change limits explicitly incorporate hydraulic and mechanical constraints into the action layer, reducing frequent start-stop cycles and frequency abrupt changes.

[0159] Soft level measurement within the conservative control level The trend is used as a start / stop preparation signal, with high-high and low-low level switch states as hard constraints. Maintenance recommendations and traceability record fields are standardized, ensuring that takeover operations go beyond conservative approaches and form executable recovery paths. Soft-sensor level measurement... It is only used as a trend reference and not directly for fine-tuning. Therefore, the actions are mainly based on boundary priority and duration constraints, and the distortion type label and consistency status are used to point to the inspection object.

[0160] Upon entering the conservative control level, the controller maintains the interlock priority of sub-step 401 and uses soft-sensing liquid level measurement. Prepare for pump start-up conditions in advance by monitoring the continuous upward trend of the liquid level, using soft-mounted liquid level measurement. Prepare for pump shutdown conditions in advance based on the continuous downward trend, so that start-up and shutdown are not dependent on distorted fusion liquid level. To avoid prolonged pump operation without interruption or inactivity, the controller is set with a maximum continuous operating time and a maximum continuous shutdown time. When the maximum continuous operating time is reached, the pump is forcibly shut down and waits for the soft-sensor liquid level. If the trend is reversed and then resumed, the pump will be forcibly started when the longest continuous downtime is reached, and drainage will continue until the high-high liquid level switch is triggered; the above durations are all written into the traceability record to explain the basis of the action.

[0161] Maintenance is recommended to be triggered by a combination of distortion type label + consistency status + validity flag: drift type with cumulative deviation. If the growth continues, prompt to clean the probe or diaphragm and verify the zero-point reference configuration; for lost probes, indicate the first validity marker. Or a second validity label When repeatedly unavailable, prompt to check communication terminals and power supply, and verify signal quality markings; stuck type with physical consistency deviation. If the pump continues to deviate from its stable position, a prompt will appear to check for installation obstructions and ensure the static water pipe is clear. The traceability record field is updated for each sampling period. Write, at least including fusion level , Integration credibility Soft level measurement Physical consistency deviation Cumulative deviation The system includes information on consistency status, distortion type label, control level, pump start / stop commands, and frequency setpoint source descriptions, along with the corresponding operational feedback status to distinguish between commands issued and equipment actions taken. During level switching, interlock triggering, and extreme pressure events, the traceability record additionally includes an event timestamp, a retention duration value, and a summary of the pump action event sequence at that time, enabling maintenance personnel to locate the trigger point along the event chain without having to reverse-engineer the process.

[0162] When using, the soft-mounted liquid level sensor... Together with the interlock boundary, it supports the trend start-up and shutdown of the control level, enabling control actions to be independent of distorted fusion liquid levels. Furthermore, it is still subject to boundary capping; maintenance recommendations are triggered jointly by distortion type labels, consistency status, and validity marking flags, ensuring that maintenance actions are consistent with the reasons for takeover and point to specific inspection objects.

[0163] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0164] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0165] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0166] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0167] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for dynamic level balancing regulation control of a sewage lifting pump station, characterized in that: include, Two types of liquid level detection devices, along with high-high and low-low liquid level switches, are deployed to collect liquid level signals and pump operation information. These signals are then preprocessed to obtain multi-channel liquid level data, validity labels, and signal quality labels. Based on the signal quality labels and pump operation information, a pump action reachability criterion and a multi-channel liquid level cross-consistency criterion are formed. The reliability score of each liquid level channel is calculated, and the liquid level reliability scores are weighted and fused to obtain the fused liquid level, fused reliability, and distortion type label. Establish the well chamber volume-liquid level relationship and convert the fused liquid level into volume change. Compare it with the net inflow / outflow difference estimated from the pump set operation information to obtain the physical consistency deviation. When the fused reliability is low or the physical consistency deviation exceeds the limit, reconstruct the soft measurement liquid level and correct it with a liquid level switch. Generate pump set control commands based on the fused liquid level, fused reliability, physical consistency deviation and liquid level switch. Switch the graded dynamic balance control and limit interlocking pipe according to the fused reliability and physical consistency deviation. Output maintenance suggestions and traceability records.

2. The liquid level dynamic balance adjustment and control method according to claim 1, characterized in that: Non-contact level gauges, submersible hydrostatic level gauges, high-high level switches, and low-low level switches are installed in the wet well to collect variable frequency operation commands and feedback, operation and fault status, and motor current. Range verification, glitch removal, rate of change constraints, and sampling period alignment are performed on the level signals to generate validity labels and signal quality marks.

3. The liquid level dynamic balance adjustment and control method according to claim 2, characterized in that: The update cycle and communication are judged for the liquid level signal. Out-of-bounds, long-term no-update, communication abnormality and self-diagnosis abnormality are written into the validity label. Lost wave, over-range and echo quality difference are parsed into structured fields and written into the signal quality label. The pump start and stop times are recorded to form a pump action event sequence.

4. The liquid level dynamic balance adjustment and control method according to claim 3, characterized in that: When the validity is marked as available and the signal quality flag meets the set conditions, a response time window after pump start-up and a reverse response time window after pump shutdown are set. Based on the pump action event sequence and pump group operation information, the reachability of pump action is judged for the liquid level trend. The soft threshold difference and hard threshold difference of multiple liquid level differences are combined to form a cross-consistency criterion, and the liquid level confidence score is calculated.

5. The liquid level dynamic balance adjustment and control method according to claim 4, characterized in that: The fused liquid level is obtained by weighted and robust fusion of multiple liquid levels based on the liquid level credibility score, and the fusion credibility is generated simultaneously. The fusion credibility is used as the criterion for switching control levels based on the pump action reachability judgment, cross consistency criterion and liquid level change mode output distortion type label.

6. The liquid level dynamic balance adjustment and control method according to claim 5, characterized in that: A well chamber volume-level lookup table is established based on wet well geometry data and stored with monotonically increasing height and volume nodes. During operation, within the sampling interval where the fusion reliability meets the set conditions, the well chamber volume-level lookup table is segmented and corrected based on the fused liquid level and pump operation information to form an equivalent cross-sectional area curve.

7. The liquid level dynamic balance adjustment and control method according to claim 6, characterized in that: The fusion liquid level is converted into volume change by looking up the well chamber volume-liquid level table, and compared with the drainage volume determined by the outflow rate measurement and the inflow volume determined by the pump shutdown interval to form the net inflow-outflow difference, thus obtaining the physical consistency deviation; the physical consistency deviation is accumulated according to the sampling period to obtain the cumulative deviation, and the consistency status is output.

8. The liquid level dynamic balance adjustment and control method according to claim 7, characterized in that: The drainage volume in the net inflow-outflow difference is estimated based on the frequency feedback of the variable frequency operation, the motor current and the outlet water pressure, and the pump type characteristics, and is calibrated with a slowly updated correction factor; the physical consistency deviation is accumulated according to the sampling period to obtain the cumulative deviation, and the systematic distortion judgment is output based on the cumulative deviation.

9. The liquid level dynamic balance adjustment and control method according to claim 8, characterized in that: In cases of low fusion reliability and excessive physical consistency deviation, the soft measurement level is obtained by integrating the well chamber volume based on the net inflow-outflow water difference and back-calculating the well chamber volume-liquid level using a table; when the high-high liquid level switch is triggered, the soft measurement level is corrected to be no lower than the corresponding height, and when the low-low liquid level switch is triggered, the soft measurement level is corrected to be no higher than the corresponding height.

10. The liquid level dynamic balance adjustment and control method according to claim 9, characterized in that: The system sets up four control levels: normal dynamic balance, downgraded conservative, distortion takeover, and extreme interlock takeover, and switches control levels based on fusion credibility and physical consistency deviation. Hysteresis, minimum running time, and minimum downtime are set at each control level. Maintenance suggestions are output and traceability records are recorded when distortion type labels continue to appear and when cumulative deviation exceeds the limit. The cumulative deviation is the deviation obtained by accumulating the physical consistency deviation according to the sampling period.

Images