A monitoring method and system for small flow water replenishment accumulation

Abstract

The application belongs to the technical field of power plant heat supply water supplement, and particularly discloses a small-flow water supplement accumulation monitoring method and system, which comprises the following steps: acquiring liquid level data of a heat supply water tank and pressure pulsation signals in a water supplement pipeline in parallel, and constructing a dual-mode sensing system; generating a liquid level reference trend line through dynamic analysis of the liquid level data to calculate a main path water volume; processing the pressure pulsation signals by using a dynamic inversion model to inversely calculate an auxiliary path equivalent water supplement amount; introducing a liquid surface calmness index as a dynamic decision basis, focusing on liquid level changes when the liquid surface is stable, and relying on pressure inversion when the liquid surface fluctuates, and performing dynamic weight collaborative correction on the main path and the auxiliary path data to finally accumulate and generate an accumulated water supplement amount. The application realizes long-term and reliable accumulation monitoring of a small water supplement amount under complex working conditions by intelligent fusion of multiple source data, and can adapt to water tanks of different shapes.

Description

Technical Field

[0001] This invention belongs to the field of power plant heating water supply technology, and relates to a monitoring method and system for small-flow water supply accumulation. Background Technology

[0002] Hot water supply systems are a common component of building facilities, providing hot water for domestic or industrial use. To maintain a stable water level, automatic water replenishment devices are typically installed. During operation, accurately monitoring and accumulating the system's water replenishment, especially small and slow replenishments caused by minor pipe leaks, valve leaks, or evaporation, is crucial for system energy consumption analysis, fault prediction, and water conservation management. Precise measurement of flow rate or capacity is a key aspect of achieving effective management and control.

[0003] In existing technologies, the cumulative monitoring of water replenishment typically employs a single technical method. One common approach is to directly install a small flow meter on the water replenishment pipe for measurement; however, conventional flow meters are insensitive or even unresponsive to extremely low flow velocities, such as dripping. Another method involves installing a level gauge in the hot water supply tank and indirectly estimating water volume changes based on changes in the liquid level. While this method can detect changes in water level, it has several shortcomings in practical applications.

[0004] The aforementioned existing technologies all have significant drawbacks. Flow meter-based methods suffer from measurement blind spots in low-flow-rate areas, leading to substantial cumulative errors. As for level gauge-based methods, firstly, the cross-section of hot water tanks is often irregular, making the conversion between level and volume complex and inaccurate; secondly, the water level in the tank fluctuates drastically due to factors such as concentrated water usage, pump start-up and shutdown, or temperature changes. This dynamic interference severely affects the accuracy of level measurement, causing drastic jumps in the water replenishment data calculated based on instantaneous levels, resulting in low reliability over the long term. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art and to achieve the above objectives, the present invention proposes the following technical solution: a monitoring method for small flow water replenishment accumulation, comprising: step 1, acquiring the liquid level height data of the hot water supply tank and the pressure pulsation signal in the water replenishment pipe, and generating dual-modal sensing data.

[0006] Step 2: Perform dynamic analysis on the liquid level height data in the dual-modal sensing data to generate a liquid level baseline trend line and a liquid surface calmness index.

[0007] Step 3: Obtain the volume-height lookup table that maps the liquid level to the volume of the hot water tank, and use the liquid level baseline trend line to look up the table and calculate the estimated volume of the main outlet water.

[0008] Step 4: Obtain a dynamic inversion model for converting pressure pulsation signals into equivalent water flow contributions, and process the pressure pulsation signals based on the model to generate the equivalent water replenishment volume for the auxiliary road.

[0009] Step 5: Using the intelligent decision module, the corrected water volume is obtained by dynamically weighting the estimated water volume of the main outlet, the equivalent water replenishment volume of the auxiliary outlet, and the calmness index of the liquid surface.

[0010] Step 6: Accumulate the corrected outflow volume to generate the cumulative water replenishment volume.

[0011] A second aspect of the present invention provides a monitoring system for small-flow water replenishment accumulation, comprising: a dual-modal sensing module, which acquires the liquid level height data of the hot water supply tank and the pressure pulsation signal in the water replenishment pipe, and generates dual-modal sensing data.

[0012] The dynamic analysis module performs dynamic analysis on the liquid level height data in the dual-modal sensing data to generate a liquid level baseline trend line and a liquid surface calmness index.

[0013] The volume calculation module retrieves a volume-height lookup table that stores the mapping relationship between the liquid level and volume of the hot water tank, and calculates the estimated volume of the main outlet water based on the liquid level baseline trend line.

[0014] The inversion processing module acquires a dynamic inversion model for inverting pressure pulsation signals into equivalent water flow contributions, and processes the pressure pulsation signals based on the model to generate the equivalent water replenishment volume for the auxiliary road.

[0015] The intelligent decision module performs dynamic weighted collaborative correction based on the estimated main outlet water volume, the equivalent replenishment water volume of the auxiliary road, and the calmness index of the liquid surface, to obtain the corrected outlet water volume.

[0016] The cumulative output module accumulates the corrected output water volume to generate the cumulative water replenishment volume.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention constructs a dual-modal sensing system by acquiring liquid level height data and pressure pulsation signal in the water replenishment pipeline in parallel, and introduces the liquid surface calmness index as a dynamic decision basis, thereby realizing dynamic collaborative correction of multi-source data under different working conditions, and thus improving the overall accuracy of small flow water replenishment cumulative monitoring.

[0018] (2) By distinguishing between long-term trends and short-term fluctuations and dynamically switching the trust weight of the data source according to the stability of the liquid surface, this invention helps to avoid measurement errors caused by factors such as user water usage and water body sloshing, ensuring the reliability and data consistency of long-term operation in complex and dynamic real environments, thereby enabling the system to have anti-interference capabilities.

[0019] (3) The present invention has wide engineering applicability and low implementation cost. It does not require large-scale modification of the hardware of the existing hot water supply system. Furthermore, it can adapt to water tanks with various non-uniform cross sections through the volume-height lookup table, making the monitoring scheme easy to integrate and apply to various residential, commercial and industrial hot water supply systems. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of the implementation steps of the method of the present invention.

[0022] Figure 2 This is a schematic diagram of the system module connections of the present invention. Detailed Implementation

[0023] 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.

[0024] Example 1

[0025] Please see Figure 1 As shown, the present invention proposes a monitoring method for small flow water replenishment accumulation, including: Step 1, acquiring the liquid level height data of the hot water supply tank and the pressure pulsation signal in the water replenishment pipeline, and generating dual-modal sensing data.

[0026] The specific implementation process involves installing a non-contact level sensor, such as an ultrasonic or radar level gauge, on the top of the hot water tank to measure the liquid level in real time. Simultaneously, a high-frequency response pressure sensor is installed on the water supply pipe connected to the tank to detect subtle pressure fluctuations caused by minute water flow. These two sensors transmit the collected liquid level data and pressure pulsation signals to a dual-modal sensing module via wired or wireless means, respectively.

[0027] The dual-modal sensing module is a hardware unit integrating data acquisition and preprocessing functions. Its function is to align the timestamps and unify the formats of two heterogeneous data streams from different physical sensors. In this method, the module packages the liquid level reading at the same moment with the pressure fluctuation waveform data within the corresponding time period into a data pair, thereby generating a dual-modal sensing data with a synchronized timestamp.

[0028] The liquid level height data is a series of discrete values, representing the height of the water surface in the tank relative to a certain reference point; the pressure pulsation signal is a segment of high-frequency sampled continuous waveform data, reflecting the dynamic changes in pressure inside the pipeline.

[0029] For example, at time T, the level sensor measures a liquid level of 2.503 meters. Simultaneously, the pressure sensor collects a pressure signal with an average pressure of 0.2 MPa, but containing several weak high-frequency fluctuations, within the past second. The dual-modal sensing module packages the 2.503-meter value and this pressure waveform data, and marks it with a timestamp of time T, forming a set of dual-modal sensing data.

[0030] Step 2: Perform dynamic analysis on the liquid level height data in the dual-modal sensing data to generate a liquid level baseline trend line and a liquid surface calmness index.

[0031] In a preferred embodiment, the step of dynamically analyzing the liquid level height data in the dual-modal sensing data to generate a liquid level reference trend line and a liquid surface calmness index includes: smoothing the liquid level height data, extracting its long-term trend, and generating a liquid level reference trend line.

[0032] Calculate the standard deviation of the liquid level height data fluctuation within a preset time window to generate a liquid surface calmness index.

[0033] The liquid surface calmness index is used to quantify the real-time stability of liquid level height data.

[0034] The specific implementation process involves sending the dual-modal sensing data to a dynamic analysis module, which contains two parallel processing logics:

[0035] The first step is to apply a low-pass filtering algorithm, such as moving average or Kalman filtering, to smooth the continuously received raw liquid level data. This aims to filter out short-term, high-frequency fluctuations caused by factors such as slight water surface swaying and sensor noise, thereby extracting a long-term trend that reflects the actual water volume changes in a smooth manner. This processed data sequence forms the liquid level baseline trend line.

[0036] For example, the system employs an exponentially weighted moving average algorithm. The core of this algorithm lies in assigning higher weights to recent data, thereby effectively filtering out instantaneous disturbances while sensitively tracking the true trend of liquid level changes. The algorithm steps are as follows: in each sampling period t, the following formula is executed to update the liquid level baseline trend line value: In the formula, This represents the currently calculated baseline value of the liquid level trend line. It is the raw liquid level height data at the current moment obtained from the liquid level sensor. It is the historical liquid level baseline trend line value calculated in the previous sampling period. When the system starts, it can be initialized to the original liquid level height value obtained by the liquid level sensor in the first sampling period. It is a preset smoothing coefficient, and its engineering significance lies in adjusting the system's response speed to new data. The smaller the value, the stronger the smoothing effect, and the better the generated liquid level baseline trend line reflects long-term changes, making it suitable for scenarios with gentle water usage patterns. A larger value indicates a more sensitive response. This parameter is typically preset based on the volume of the hot water tank and typical water usage patterns during on-site commissioning, with common engineering values ​​ranging from 0.01 to 0.1. Then, all liquid level baseline trend line values ​​over several past sampling periods are calculated to obtain a smooth curve showing the change of the liquid level baseline trend line values ​​over time; this is the liquid level baseline trend line.

[0037] The second step is to calculate the surface calmness index. Raw liquid level data is collected within a preset sliding time window, such as a preset 10-second time window, and the standard deviation of all data points within this time window is calculated. This standard deviation is defined as the surface calmness index, and its value directly quantifies the real-time stability of the liquid surface: the smaller the index value, the calmer the liquid surface and the more reliable the liquid level reading; the larger the index value, the more turbulent the liquid surface fluctuations.

[0038] The algorithm steps are as follows: The system first calculates the standard deviation of the liquid level height data within the preset time window. The system then executes a formula to calculate the liquid surface calmness index: In the formula, The liquid surface calmness index is limited to a range of 0 to 1. The closer the value is to 1, the calmer the liquid surface and the higher the reliability of the data. The standard deviation of the liquid level height data, calculated using a standard statistical formula within a sliding time window containing N latest sampling points, directly reflects the severity of liquid level fluctuations. This is a preset dimensionless scaling factor used to map the range of standard deviation values ​​that may occur on-site to an exponential interval of 0 to 1. This factor is obtained by recording the standard deviations of the water tank under static and typical maximum fluctuation conditions during the system commissioning phase, and then calibrating accordingly.

[0039] Through this step, the system successfully decomposes the raw, noisy liquid level height data into a smooth liquid level baseline trend line and a liquid surface calmness index that measures its stability, providing high-quality input for subsequent co-calibration.

[0040] For example, the system first receives liquid level data from the past 10 seconds, which fluctuates slightly between 2.500 meters and 2.506 meters. Using a moving average, the system outputs a baseline liquid level trend line with a current value of 2.503 meters. Simultaneously, the standard deviation of the raw data over these 10 seconds is calculated, yielding a surface calmness index of 0.002.

[0041] Step 3: Obtain the volume-height lookup table that maps the liquid level to the volume of the hot water tank, and use the liquid level baseline trend line to look up the table and calculate the estimated volume of the main outlet water.

[0042] In a preferred embodiment, the step of obtaining a volume-height lookup table that stores the mapping relationship between the liquid level and volume of the hot water tank, and then looking up the table based on the liquid level reference trend line to calculate the estimated volume of the main outlet water includes: using the liquid level reference trend line, determining the corresponding volume value based on the volume-height lookup table, and obtaining the current volume value and historical volume value.

[0043] Calculate the difference between the current volume value and the historical volume value to generate the volume change.

[0044] Based on the volume change, the estimated volume of water discharged from the main pipeline is determined.

[0045] The objective of this invention is to, for hot water tanks with non-uniform cross-sectional shapes, such as cylindrical, cuboid, or irregular shapes, use table lookup and interpolation to calculate the liquid level baseline trend line output by the dynamic analysis module into a physically meaningful estimated value of the main outlet water volume.

[0046] The specific implementation process begins with relying on a pre-defined hot water tank volume-height lookup table. This table is generated during the system installation and commissioning phase through integration or field calibration of the specific hot water tank's geometric model. It stores a series of precise mapping relationships between liquid level heights and corresponding water volumes in the form of high-density discrete points, such as height increments in millimeters or centimeters. This table solidifies the non-uniform shape characteristics of the water tank and forms the basis for subsequent accurate calculations.

[0047] In each calculation cycle, obtain the current liquid level baseline trend line value. Compared with the historical liquid level benchmark trend line value of the previous period Since these two values ​​are typically between the preset discrete height values ​​included in the volume-height lookup table, the system uses the periodic liquid level trend line value as an index. Based on the volume-height lookup table, it determines the lower limit height point that is less than and closest to the corresponding periodic liquid level trend line value, and the upper limit height point that is greater than and closest to the corresponding periodic liquid level trend line value. It calculates the ratio of the difference between the trend line value and the lower limit height point to the height difference between the upper and lower limit height points. This ratio is multiplied by the corresponding volume difference and then superimposed onto the volume value corresponding to the lower limit height point, thereby obtaining the value corresponding to the volume-height lookup table. and Precisely corresponding current volume Volume of the previous period .

[0048] After obtaining these two precise volume values, the volume change within that period can be calculated using the following formula: In the formula, This is the calculated estimated volume of water flowing out of the main pipeline, representing the amount of water reduction reflected only by changes in liquid level within a calculation period. The volume of the water tank in the previous period was obtained by looking up a table and interpolating. This is the tank volume for the current cycle, obtained using the same method. When A positive value indicates that water has been released from the system.

[0049] This step transforms complex geometric conversion problems into efficient table lookups and simple algebraic operations, ensuring the real-time performance and accuracy of the calculations.

[0050] For example, continuing from the previous example, the current liquid level baseline trend value is 2.503 meters. The volume calculation module queries the volume-height lookup table and obtains the corresponding current volume of 3.154 cubic meters. In the previous calculation cycle, the recorded historical liquid level baseline trend value was 2.510 meters, corresponding to a historical volume of 3.168 cubic meters. The module calculates the volume change as follows: Therefore, the estimated volume of water discharged from the main pipeline for this cycle is 14 liters.

[0051] Step 4: Obtain a dynamic inversion model for converting pressure pulsation signals into equivalent water flow contributions, and process the pressure pulsation signals based on the model to generate the equivalent water replenishment volume for the auxiliary road.

[0052] In a preferred embodiment, the step of acquiring a dynamic inversion model for inverting pressure pulsation signals into equivalent flow contributions, and processing the pressure pulsation signals based on the model to generate equivalent supplementary water volume for the auxiliary channel, includes: identifying fluctuation patterns characterizing minute flow rates from the pressure pulsation signals and generating pressure feature vectors.

[0053] The pressure feature vector is input into the dynamic inversion model to calculate the instantaneous equivalent flow rate.

[0054] The instantaneous equivalent flow rate is integrated over time to generate the equivalent water replenishment volume for the auxiliary road.

[0055] The specific implementation process is as follows: During the system deployment phase, a series of known small flow rates, such as a few milliliters to tens of milliliters per minute, are generated in the pipeline, and signals from pressure sensors are collected simultaneously. A mapping relationship between the pressure signal fluctuation pattern and the actual flow rate is established, and this relationship is solidified into a preset dynamic inversion model. This model is a nonlinear mapping function built based on machine learning algorithms, typically a multivariate regression model or a lightweight neural network trained based on field experimental data.

[0056] During system operation, after receiving high-frequency pressure data from the water supply pipeline, the inversion processing module first performs signal feature extraction. The system performs time-frequency domain transformation on the original pressure pulsation signal, such as using short-time Fourier transform or wavelet transform, to identify fluctuation patterns characterizing minute flow rates. A fluctuation pattern is a signal fingerprint experimentally calibrated in fluid mechanics, specifically referring to the weak oscillating waveform with a expected frequency range and amplitude attenuation characteristics caused by microscopic water hammer effects or fluid turbulence on the pressure sensor when water flows through pipe valves or orifices at a minute velocity, such as dripping or linear flow.

[0057] Once this pattern is identified, the system extracts its key parameters, such as dominant frequency, amplitude, and information entropy, and encapsulates these parameters to generate a pressure feature vector. The pressure feature vector is a standardized multidimensional numerical array that transforms complex analog waveform signals into a set of digital features that can be directly processed by a computer model.

[0058] The system then uses this pressure feature vector as input to a pre-deployed dynamic inversion model. This model functions similarly to a virtual flow meter, calculating the current flow velocity within the pipe based on the input feature vector—that is, determining the instantaneous equivalent flow rate. The instantaneous equivalent flow rate is an estimated physical quantity representing the theoretical water velocity flowing through the pipe within the current millisecond-level time slice, typically measured in milliliters per second.

[0059] Finally, within each calculation cycle, the system performs time integration on the continuously output instantaneous equivalent flow rate values. Time integration is a mathematical accumulation operation that involves multiplying the instantaneous flow rate at discrete time points by the sampling time interval and continuously summing these tiny volume increments. Through this process, the system generates the equivalent makeup water volume for the auxiliary path. The equivalent makeup water volume for the auxiliary path refers to the total volume of tiny makeup water that accumulates within a complete statistical cycle, calculated through pressure signal inversion in the blind zone that cannot be detected by the main flow meter or level gauge.

[0060] The algorithm steps are as follows: In each analysis cycle, the calculation system extracts the energy spectral density within a predetermined frequency band from the pressure pulsation signal. and dominant frequency The system uses these features as input to the model. Then, it calls the dynamic inversion model to calculate the equivalent water replenishment volume of the auxiliary road using a formula: In the formula, The equivalent water replenishment volume for the generated auxiliary road represents the water replenishment volume retrieved from the pressure signal within one calculation cycle. It is the functional expression of the dynamic inversion model, which receives the energy spectral density. and dominant frequency As input, output an instantaneous equivalent flow value. This refers to the length of the calculation period. For example, a simplified linear regression model can be expressed as: ,in , , The coefficients are determined during model calibration.

[0061] Through this step, the system can keenly capture the "fingerprint" of minute water replenishment events from a seemingly stable pressure background and quantify them into an equivalent water replenishment volume of the auxiliary road with actual physical meaning, thereby making up for the blind spot of liquid level measurement under low flow conditions.

[0062] For example, following the pressure pulsation signal acquired in the previous step, the system detected a weak but persistent oscillation signal with a frequency concentrated between 45Hz and 50Hz within a 10-second window. The system identified this as a typical "minor internal valve leakage" pattern and generated a pressure feature vector with the following content: [center frequency: 48Hz, amplitude: 0.02bar, information entropy: 0.85]. The inversion processing module input this vector into the dynamic inversion model, and the model outputs an instantaneous equivalent flow rate of 0.05 liters / minute. Subsequently, the integration unit calculates the flow rate within these 10 seconds, i.e., (0.05 liters / minute ÷ 60 seconds / minute) × 10 seconds, ultimately generating an equivalent auxiliary channel water replenishment volume of 0.0083 liters. This value is then sent to the accumulation module, thus accurately recording this minute water replenishment event that the level gauge could not detect.

[0063] Step 5: Using the intelligent decision module, the corrected water volume is obtained by dynamically weighting the estimated water volume of the main outlet, the equivalent water replenishment volume of the auxiliary outlet, and the calmness index of the liquid surface.

[0064] In a preferred embodiment, the step of using an intelligent decision module to perform dynamic weighting and collaborative correction based on the estimated main outlet volume, the equivalent replenishment volume of the auxiliary road, and the calmness index of the liquid surface to obtain the corrected outlet volume includes: dynamically allocating the main road weight and the auxiliary road weight based on the estimated main outlet volume and the equivalent replenishment volume of the auxiliary road according to the calmness index of the liquid surface.

[0065] When the liquid level calmness index indicates that the liquid level is stable, increase the weight of the main road and decrease the weight of the auxiliary road.

[0066] When the liquid level calmness index indicates liquid level fluctuations, reduce the weight of the main road and increase the weight of the auxiliary road.

[0067] The estimated effluent volume of the main road and the equivalent replenishment volume of the auxiliary road are weighted and fused based on the weight of the main road and the weight of the auxiliary road to generate the corrected effluent volume.

[0068] The specific implementation process is as follows: In order to achieve intelligent fusion of multi-source data, the intelligent decision module receives the estimated volume of the main outlet water. Equivalent water replenishment volume of auxiliary roads And the key liquid surface calmness index Then, a dynamic weighting collaborative correction will be performed. Dynamic weighting collaborative correction is an adaptive data fusion strategy. Its core is not to rely on any single data source, but to dynamically adjust the contribution ratio of different data sources in the final result calculation based on the credibility of real-time conditions.

[0069] This method first dynamically allocates the weights of the main pipeline and the auxiliary pipeline based on the estimated effluent volume of the main pipeline and the equivalent makeup water volume of the auxiliary pipeline, according to the level of the surface calmness index. The main pipeline weight and the auxiliary pipeline weight are two dimensionless coefficients between 0 and 1, whose sum is always equal to 1, representing the importance of the estimated main pipeline value and the estimated auxiliary pipeline value in the final fusion calculation, respectively. The logic of weight allocation is as follows: the main pipeline weight is positively correlated with the surface calmness index, while the auxiliary pipeline weight is negatively correlated with it. The higher the calmness index, the closer the main pipeline weight is to its set upper limit, and vice versa.

[0070] The specific calculation of weights can be achieved through the following weight mapping function: and In the formula, These are the weights assigned to the estimated volume of water discharged from the main pipeline. It is the weight assigned to the equivalent water replenishment of the auxiliary road. Both are dimensionless numbers between 0 and 1, and their sum is 1. It is a weighting mapping function. The simplest implementation is a linear function, such as directly mapping the value of the surface calmness index to the weight value assigned to the estimated volume of water flowing out of the main channel. In this situation, the higher the calmness index, the greater the weight of the main road.

[0071] After obtaining the dynamic weights, the system performs data fusion using the formula. Calculate the corrected effluent volume, where, This is the corrected output water volume.

[0072] Through this step, the system adaptively adjusts its trust level for different data sources based on the reliability of the current system state. This ensures that the macroscopic accuracy of liquid level measurement is fully utilized when the liquid level is stable, while relying on the robustness of pressure inversion to microscopic disturbances when the liquid level fluctuates. As a result, the system obtains continuous, stable and accurate measurement results throughout the entire dynamic range.

[0073] For example, continuing from the above example, the estimated main outlet water volume is 14 liters, the equivalent auxiliary water supply is 0.1 liters, and the surface calmness index is 0.002. Assuming the weight adjustment threshold is 0.05, since 0.002 is far below this threshold, the intelligent decision module determines that the surface is stable and selects a high main outlet weight combination, i.e., main outlet weight 0.95 and auxiliary outlet weight 0.05. The corrected outlet water volume is calculated as (14 × 0.95) + (0.1 × 0.05) = 13.3 + 0.005 = 13.305 liters.

[0074] In a further preferred embodiment, the step of dynamically allocating the weights of the main road and the auxiliary road based on the estimated volume of water discharged from the main road and the equivalent water replenishment volume of the auxiliary road according to the surface calmness index includes: obtaining a weight adjustment threshold, comparing the surface calmness index with the threshold, and generating a judgment result of surface stability or surface fluctuation.

[0075] Based on the judgment, the weights of the main road and the auxiliary road will be switched to one of the predetermined weight combinations.

[0076] The predetermined weight combination includes a high main road weight combination and a high auxiliary road weight combination.

[0077] The specific implementation process is as follows: The system first acquires a preset weight adjustment threshold, which is a specific quantified liquid surface calmness index value, used in engineering as a clear dividing line for judging the transition of liquid surface state from stable to fluctuating. During the system initialization phase, this weight adjustment threshold is calibrated according to the on-site hydraulic characteristics, typically set between 0.7 and 0.9. Simultaneously, the system presets a first preset value and a second preset value, both of which are dimensionless high-confidence weights, typically ranging from 0.8 to 0.95.

[0078] In each calculation cycle, the intelligent decision module will calculate the liquid surface calmness index in real time. With weight adjustment threshold The comparison and the judgment logic are as follows:

[0079] if Greater than The system determines that the liquid level is stable. At this point, the main route data based on the liquid level is highly reliable, and the system switches the main route weight and auxiliary route weight to a combination with a high main route weight, executing the formula: and In the formula, the weight of the estimated volume of water discharged from the main pipeline is... It was directly set to the first preset value. The weight of the equivalent water replenishment volume of the auxiliary road It is then set to 1 minus accordingly. .

[0080] Conversely, if Not greater than The system's judgment result is liquid level fluctuation. At this point, the reliability of the liquid level data decreases, and the system switches the weights to the high auxiliary road weight combination, executing the formula: and In the formula, the weight of the equivalent water replenishment volume of the auxiliary road is... It was directly set to the second preset value. This reflects that the system at this moment places more trust in the inversion results from the pressure pulsation signal, while the weight of the estimated main outlet water volume is lower. Then adjust accordingly to 1 minus .

[0081] The predetermined weight combination consists of two sets of fixed weight configuration parameters representing extreme trust levels, which are also preset during system initialization.

[0082] In a high main road weighting combination, the main road weight is set to an extremely high value close to 1, such as 0.95, while the secondary road weight is correspondingly set to an extremely low value, i.e. This combination demonstrates absolute confidence in the estimated volume of water flowing out of the main pipeline when the liquid level is stable.

[0083] In a high-weighted auxiliary road combination, the auxiliary road weight is set to an extremely high value close to 1, such as 0.9, while the main road weight is correspondingly set to an extremely low value, i.e. This combination reflects the decision-making logic of prioritizing the inversion results of pressure pulsation signals when the liquid level fluctuates.

[0084] By introducing this explicit threshold judgment and discretized weight combination, this method simplifies the complex dynamic weight adjustment process into a logical switch. The system no longer performs continuous proportional calculations, but instead makes decisive and predictable trust switches between stable and fluctuating operating conditions, thereby improving the response speed and engineering robustness of the entire monitoring system while ensuring the core logic.

[0085] For example, suppose the weight adjustment threshold is set to 0.8, the high main road weight combination is {main road: 0.95, secondary road: 0.05}, and the high secondary road weight combination is {main road: 0.1, secondary road: 0.9}.

[0086] Scenario 1: At a certain moment, the system calculates the liquid surface calmness index to be 0.92. Since 0.92 > 0.8, the system determines that the liquid surface is stable and immediately adopts a high main path weight combination, that is, the main path weight is set to 0.95 and the auxiliary path weight is set to 0.05 for subsequent weighted fusion calculation.

[0087] Scenario 2: At another moment, due to concentrated water usage by users, the liquid surface sloshes, and the calmness index drops sharply to 0.65. Since 0.65 ≤ 0.8, the system determines that the liquid surface is fluctuating and immediately switches to the high auxiliary road weight combination, setting the main road weight to 0.1 and the auxiliary road weight to 0.9. Thus, during this period, the system mainly relies on the equivalent water replenishment volume of the auxiliary road to estimate the total volume change.

[0088] Step 6: Accumulate the corrected outflow volume to generate the cumulative water replenishment volume.

[0089] In a preferred embodiment, the specific implementation process is as follows: after the intelligent decision module outputs the corrected effluent volume for each calculation cycle, the system enters the final accumulation and output stage. This step is executed by the cumulative output module, whose engineering purpose is to continuously sum the incremental data at discrete time points, thereby forming a cumulative total with macroscopic statistical significance, and presenting it in a way that is perceptible to the user.

[0090] At the heart of this process is a continuously running accumulator. In programming, the accumulator is typically represented as a static or global variable, used to continuously store and update the total accumulated water replenishment value throughout the system's uninterrupted operation. At the end of each calculation cycle, the system executes the following formula: The latest corrected effluent volume is added to the accumulator, where, The cumulative water replenishment volume after the current update is the sum of all corrected outflow volumes since the system started or was last reset, and the unit is volume, such as cubic meters. This represents the cumulative water replenishment amount in the previous calculation period. When the system is first started, the initial value of this value is set to zero. This represents the corrected outflow volume of the just-completed calculation cycle, output by the intelligent decision module.

[0091] By repeatedly performing this accumulation operation in each sampling cycle, the system aggregates a series of discrete, intelligently corrected volume increments into a continuously increasing cumulative water replenishment curve. This final cumulative water replenishment value provides users with information about the total amount of water replenished to the hot water supply system over a period of time due to normal consumption, evaporation, or minor leaks. It serves as a direct basis for energy consumption analysis, cost accounting, and fault diagnosis.

[0092] For example, suppose that at the beginning of a calculation cycle, the system's internally stored cumulative water replenishment volume is 12.583 cubic meters. During this cycle, after all the complex calculations in the preceding steps, the intelligent decision module finally outputs a corrected output water volume of 0.015 cubic meters. At the end of this calculation cycle, the cumulative output module performs an accumulation operation: the new cumulative water replenishment volume = 12.583 + 0.015 = 12.598 cubic meters. Subsequently, the system updates this new value of 12.598 to the local display screen and reports it to the remote monitoring center through the communication interface, thus completing a complete closed-loop process of "measurement-correction-accumulation-output".

[0093] In a further preferred embodiment, the method further includes: encapsulating the cumulative water replenishment volume to generate a standardized data message.

[0094] Standardized data messages are sent to an external monitoring system via a communication interface.

[0095] The cumulative water replenishment volume is recorded and displayed on the external monitoring system.

[0096] The specific implementation process is as follows: after the cumulative output module completes the accumulation operation of the corrected effluent volume, the system will execute the final data presentation and upload task. The engineering purpose of the cumulative output module is to transform the internal calculation results into interactive information that is useful to the external world.

[0097] At the end of each calculation cycle, the process first adds the accumulated corrected effluent volume generated in that cycle to the total variable to update the cumulative replenishment volume. Then, the following output and display actions are performed:

[0098] For digital display, the system sends the updated cumulative water replenishment in a preset format, such as retaining two decimal places and including the unit liters or cubic meters, to the local human-machine interface, such as an LCD screen, via the driver interface. The refresh rate is usually set between 0.5 Hz and 2 Hz to ensure visual stability and data immediacy.

[0099] For data transmission to external monitoring systems, the system encapsulates cumulative water replenishment, current liquid level, calmness index, and other relevant status information into standardized data packets. A standardized data packet is a data structure that follows a public communication protocol, mapping internal variables to agreed-upon addresses or topics, ensuring mutual understanding and communication between devices from different manufacturers. The system then periodically sends these data packets to external monitoring systems, such as Building Automation Systems (BAS) or Supervisory Control Systems (SCADA), via a physical communication interface, such as an Ethernet port. The data transmission cycle can be configured according to the needs of the higher-level system, such as the BAS or SCADA, with typical engineering settings ranging from 5 seconds to 300 seconds to achieve seamless data integration with the entire heating monitoring network.

[0100] After receiving the data, the external monitoring system will parse it, record it to the historical database, and display the cumulative water replenishment in the form of digital readings, historical trend curves or dashboards on the large screen in the central monitoring room or on the software interface of the engineer's station, thereby realizing remote centralized monitoring and management.

[0101] For example, the system calculates the current cumulative water replenishment to be 12.583 cubic meters. It encapsulates this floating-point value into a Modbus TCP packet and writes it to a holding register at address 40001. The host computer monitoring system reads the value of this register according to a preset 10-second polling cycle and displays it in the "Cumulative Water Replenishment to Tank" section of the heating system monitoring overview screen. At the same time, it stores this value along with a timestamp in the database for subsequent energy consumption analysis and report generation.

[0102] In a further preferred embodiment, the method further includes: identifying a high-confidence calibration window in which the liquid surface calmness index remains above a preset calibration threshold.

[0103] Within this window, the baseline water replenishment is calculated based on the volume-height lookup table and the liquid level baseline trend line.

[0104] The baseline water replenishment volume is compared with the equivalent water replenishment volume of the auxiliary road generated in the same period, and the model parameters of the dynamic inversion model are updated based on the difference.

[0105] The specific implementation process is as follows: To ensure the long-term accuracy of the system, this invention integrates a self-diagnosis and online calibration mechanism. This mechanism is automatically triggered by a background self-calibration module at a preset long period, such as once a week or once a month. Its engineering purpose is to proactively compensate for measurement drift caused by sensor aging, pipe scaling, or changes in the physical properties of water.

[0106] Upon triggering, the system first scans its internally stored historical data to identify high-confidence calibration windows. A high-confidence calibration window is a historical time interval that meets the following stringent conditions: a sufficiently long duration, such as more than one hour, during which the surface calmness index remains consistently above a preset calibration threshold, such as 0.98, and the estimated main outlet water volume remains consistently zero or close to zero. These conditions collectively ensure that the tank is in a stable, low-volume replenishment state within this time window.

[0107] After finding the high-confidence window, the system will perform calibration on the dynamic inversion model and the volume-height lookup table respectively:

[0108] (i) For the calibration of the dynamic inversion model, the system uses the small, stable rise of the liquid level reference trend line within the window to obtain a reliable water replenishment reference. Because the liquid level is stable and there is no outflow during this period, the volume increase corresponding to the change in the liquid level reference trend line can be regarded as the true cumulative water replenishment during this period. The system first obtains the liquid level height at the beginning and end of the window, and performs interpolation calculations based on the volume-height lookup table to calculate the exact starting and ending volumes corresponding to these two height values. The difference between the two volumes is the reference water replenishment. This is the actual value. Simultaneously, the system accumulates the sum of the equivalent water replenishment volume for the auxiliary road calculated by the dynamic inversion model within this window. This is the model's predicted value.

[0109] Subsequently, through the formula Update the parameters of the dynamic inversion model, where, After this calibration update, it is a new set of internal parameters for the dynamic inversion model. Its feature is a vector, which means that the model contains multiple parameters that need to be adjusted, such as the weight coefficients of each feature in linear regression, or the weights and biases of a neuron in a certain layer of a neural network. This is the parameter vector that the model was using before this calibration. The learning rate is a dimensionless constant used to control the magnitude of each adjustment to avoid over-calibration. In engineering practice, the learning rate is usually started from orders of magnitude such as 0.1, 0.01, 0.001, etc., and an optimal value is selected by observing the model's performance on the test dataset to ensure convergence without causing oscillations. It is the total error within the calibration window, that is, the overall deviation between the model's predicted value and the actual value. It is the average vector of the pressure features input to the model within this window, such as amplitude and dominant frequency.

[0110] (ii) For the calibration of the volume-height lookup table, the system searches for another characteristic event: a complete large-flow water replenishment process. This event is marked as follows: the system detects that the rate of liquid level rise is consistently higher than the rise rate threshold within a high-confidence window, and the total rise exceeds the minimum amplitude threshold. The rise rate threshold is a preset critical value representing the rate of liquid level change, for example, 0.01 m / min. The minimum amplitude threshold is a preset critical value representing the total change in liquid level, for example, 0.5 m.

[0111] To obtain the standard volume for calibration, the system uses the start and end timestamps of the marked high-flow-rate water replenishment event as the query basis. After the event ends, a data request is initiated to an external system holding the standard volume data through a preset communication interface and protocol. This request aims to query the cumulative flow value recorded by the external coarse flow meter within the aforementioned timestamp interval. The value returned by the external system is then used as the total standard volume for this calibration. Meanwhile, the system uses the same initial and final liquid level baseline values, consults the internal volume-height lookup table to obtain the corresponding volume values, and then calculates the corresponding volume increment. .

[0112] like and If the relative deviation exceeds the calibration trigger threshold, the system calculates a correction scaling factor. ',like The calibration trigger threshold is a preset percentage value, such as 5%, which defines how much deviation is unacceptable and requires calibration. This correction scaling factor is then used to adjust the mapping relationship of the relevant segments in the volume-height lookup table, thereby realigning the geometric model with physical reality.

[0113] Through this dual calibration mechanism, the system achieves self-evolution and accuracy maintenance.

[0114] For example, the system identified a high-confidence calibration window between 3:00 AM and 4:00 AM. During this period, the liquid level baseline trend line steadily rose from 2.501 meters to 2.503 meters, and the baseline water replenishment volume calculated by looking up the table was 5.0 liters. Meanwhile, the cumulative total of the equivalent water replenishment volume for the auxiliary road calculated by the dynamic inversion model was 4.8 liters. The system detected a deviation of 0.2 liters and immediately activated the calibration algorithm, fine-tuning the weighting coefficients corresponding to the frequency components related to the deviation in the inversion model, thus slightly improving the model's response sensitivity to this type of minor water replenishment pattern.

[0115] In a further preferred embodiment, the method further includes: obtaining an alarm threshold and a duration threshold.

[0116] Determine if the liquid surface calmness index remains below the alarm threshold and exceeds the duration threshold, and generate an alarm trigger signal.

[0117] Based on the alarm trigger signal, a system alarm indicating abnormal fluctuations in the liquid level is output.

[0118] Specifically, to enhance the system's fault diagnosis capabilities, this invention also includes an alarm triggering mechanism based on persistent abnormalities in the liquid surface state. The core of this mechanism lies in the continuous monitoring of the liquid surface calmness index over time to distinguish between normal instantaneous fluctuations and long-term anomalies that may indicate physical failure.

[0119] The implementation process is as follows: the system first acquires two preset parameters: a preset alarm threshold and a duration threshold. The preset alarm threshold is a low liquid surface calmness index value, such as 0.3, which represents a significantly unstable liquid surface. The duration threshold defines the maximum duration for which the system can tolerate this unstable state, such as 3 to 5 minutes; exceeding this duration indicates that the anomaly is no longer accidental.

[0120] During system operation, a dedicated alarm monitoring logic unit runs in the background. This unit obtains the latest liquid surface calmness index in each calculation cycle. and compare it with the preset alarm threshold. Comparison:

[0121] if Below Instead of immediately triggering an alarm, the system starts or increments an internal timer.

[0122] If in a subsequent calculation cycle, Restore to The timer will then be reset to zero, indicating that the anomaly is temporary and no alarm is required.

[0123] If the timer's accumulated time reaches the preset duration threshold That is, in continuous If the liquid surface calmness index remains below the preset alarm threshold for an extended period of time, the system determines that a continuous abnormal event has occurred and generates an alarm trigger signal.

[0124] Based on the alarm trigger signal, the system will immediately output a system alarm that clearly indicates abnormal fluctuations in the liquid level. The alarm can be triggered in two ways, including but not limited to: the system activating a local audible and visual alarm to issue a warning signal, while simultaneously displaying a clear fault code on the human-machine interface, such as "E01 - Continuous Liquid Level Fluctuation". Alternatively, the system may send a message indicating the alarm or a high-priority SNMP trap message to an external monitoring system via its communication interface, containing the alarm type and timestamp of occurrence.

[0125] This continuous monitoring logic helps filter out instantaneous fluctuations caused by normal water usage peaks or brief hydraulic shocks, and only issues alarms for long-term, persistent anomalies that may indicate deeper problems such as stuck water supply valves, pipe resonance, or sensor malfunctions. This improves the accuracy and effectiveness of alarms and reduces misjudgments by maintenance personnel.

[0126] For example, the alarm threshold is set to 0.3, and the duration threshold is set to 5 minutes. At 2:00 PM, due to a jammed water supply valve, water flowed into the tank in pulsating patterns, causing the water level calmness index to plummet to 0.25 and remain below 0.3. The anomaly timer started. For the next 5 minutes, the index fluctuated between 0.2 and 0.28. When the timer reached 5 minutes, the system officially generated an alarm trigger signal, immediately displaying a message on the local screen: "Warning: Abnormal fluctuations in the water level have lasted for 5 minutes. Please check the status of the water supply valve," and sending an emergency alarm containing this information to the central monitoring room.

[0127] Example 2

[0128] Please see Figure 2 As shown, based on Embodiment 1, the second aspect of the present invention provides a monitoring system for small-flow water replenishment accumulation, comprising: a dual-modal sensing module, a dynamic analysis module, a volume calculation module, an inversion processing module, an intelligent decision module, and an accumulation output module.

[0129] The dual-modal sensing module acquires the liquid level data of the hot water tank and the pressure pulsation signal in the water supply pipe, and generates dual-modal sensing data.

[0130] The dynamic analysis module performs dynamic analysis on the liquid level height data in the dual-modal sensing data to generate a liquid level baseline trend line and a liquid surface calmness index.

[0131] The volume calculation module obtains a volume-height lookup table that stores the mapping relationship between the liquid level and volume of the hot water tank, and calculates the estimated value of the main outlet water volume based on the liquid level baseline trend line.

[0132] The inversion processing module acquires a dynamic inversion model for inverting pressure pulsation signals into equivalent water flow contributions, and processes the pressure pulsation signals based on the model to generate equivalent supplementary water volume for the auxiliary road.

[0133] The intelligent decision module performs dynamic weighted collaborative correction based on the estimated main outlet water volume, the equivalent replenishment water volume of the auxiliary road, and the calmness index of the liquid surface to obtain the corrected outlet water volume.

[0134] The cumulative output module accumulates the corrected water volume to generate the cumulative water replenishment volume.

[0135] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. A monitoring method for the cumulative effect of small-flow water replenishment, characterized in that, include: Step 1: Obtain the liquid level data of the hot water tank and the pressure pulsation signal in the water supply pipe to generate dual-modal sensing data; Step 2: Perform dynamic analysis on the liquid level height data in the dual-modal sensing data to generate a liquid level baseline trend line and a liquid surface calmness index; Step 3: Obtain the volume-height lookup table that maps the liquid level to the volume of the hot water tank, and use the liquid level baseline trend line to look up the table and calculate the estimated volume of the main outlet water. Step 4: Obtain the dynamic inversion model for converting pressure pulsation signals into equivalent water flow contributions, and process the pressure pulsation signals based on the model to generate the equivalent water replenishment volume for the auxiliary road. The process of acquiring a dynamic inversion model for converting pressure pulsation signals into equivalent flow contributions, and processing the pressure pulsation signals based on this model to generate equivalent supplementary water supply for the auxiliary channel, includes: The fluctuation patterns representing minute flow rates are identified from pressure pulsation signals, and pressure feature vectors are generated. The pressure feature vector is input into the dynamic inversion model to calculate the instantaneous equivalent flow rate. The instantaneous equivalent flow rate is integrated over time to generate the equivalent water replenishment volume for the auxiliary road. Step 5: Using the intelligent decision module, the dynamic weights of the estimated main outlet water volume, the equivalent water replenishment volume of the auxiliary road, and the calmness index of the liquid surface are dynamically weighted and coordinated to obtain the corrected outlet water volume. The intelligent decision module performs dynamic weighted collaborative correction based on the estimated main outlet water volume, the equivalent auxiliary water supply, and the surface calmness index to obtain the corrected outlet water volume, including: Based on the calmness index of the liquid surface, the weights of the main road and the auxiliary road are dynamically allocated according to the estimated volume of water discharged from the main road and the equivalent water replenishment volume of the auxiliary road. When the liquid level calmness index indicates that the liquid level is stable, increase the weight of the main road and decrease the weight of the auxiliary road. When the liquid level calmness index indicates liquid level fluctuations, reduce the weight of the main road and increase the weight of the auxiliary road. The estimated effluent volume of the main road and the equivalent water replenishment of the auxiliary road are weighted and fused based on the weight of the main road and the weight of the auxiliary road to generate the corrected effluent volume. Step 6: Accumulate the corrected outflow volume to generate the cumulative water replenishment volume.

2. The monitoring method for small-flow water replenishment accumulation according to claim 1, characterized in that, The dynamic analysis of liquid level height data in the dual-modal sensing data to generate a liquid level baseline trend line and a liquid surface calmness index includes: Smooth the liquid level data, extract its long-term trend, and generate a liquid level baseline trend line; Calculate the standard deviation of the liquid level height data fluctuation within a preset time window to generate a liquid surface calmness index.

3. The monitoring method for small-flow water replenishment accumulation according to claim 1, characterized in that, The process of obtaining a volume-height lookup table that maps the liquid level to the volume of the hot water tank, and then using this table to calculate the estimated volume of the main outlet water based on the liquid level baseline trend line, includes: Using the liquid level baseline trend line, the corresponding volume value is determined based on the volume-height lookup table to obtain the current volume value and historical volume values; Calculate the difference between the current volume value and the historical volume value to generate the volume change; Based on the volume change, the estimated volume of water discharged from the main pipeline is determined.

4. The monitoring method for small-flow water replenishment accumulation according to claim 1, characterized in that, The dynamic allocation of main road weights and auxiliary road weights based on the estimated main road outlet volume and the equivalent auxiliary road makeup volume according to the liquid surface calmness index includes: Obtain the weight adjustment threshold, compare the liquid surface calmness index with the threshold, and generate a judgment result of liquid surface stability or liquid surface fluctuation; Based on the judgment, the weights of the main road and the auxiliary road will be switched to one of the predetermined weight combinations; The predetermined weight combination includes a high main road weight combination and a high auxiliary road weight combination.

5. The monitoring method for small-flow replenishment accumulation according to claim 1, characterized in that, The method further includes: The cumulative water replenishment volume is encapsulated to generate a standardized data message; Standardized data messages are sent to an external monitoring system via a communication interface; The cumulative water replenishment volume is recorded and displayed on the external monitoring system.

6. The monitoring method for small-flow replenishment accumulation according to claim 1, characterized in that, The method further includes: Identify high-confidence calibration windows where the liquid surface calmness index remains consistently higher than a preset calibration threshold. Within this window, the baseline water replenishment is calculated based on the volume-height lookup table and the liquid level baseline trend line; The baseline water replenishment volume is compared with the equivalent water replenishment volume of the auxiliary road generated in the same period, and the model parameters of the dynamic inversion model are updated based on the difference.

7. The monitoring method for small-flow water replenishment accumulation according to claim 1, characterized in that, The method further includes: Get the alarm threshold and duration threshold; Determine if the liquid surface calmness index remains below the alarm threshold and exceeds the duration threshold, and generate an alarm trigger signal accordingly; Based on the alarm trigger signal, a system alarm indicating abnormal fluctuations in the liquid level is output.

8. A monitoring system for cumulative small-flow water replenishment, characterized in that, include: The dual-modal sensing module acquires the liquid level data of the hot water tank and the pressure pulsation signal in the water supply pipe, and generates dual-modal sensing data. The dynamic analysis module performs dynamic analysis on the liquid level height data in the dual-modal sensing data to generate a liquid level baseline trend line and a liquid surface calmness index. The volume calculation module obtains a volume-height lookup table that stores the mapping relationship between the liquid level and volume of the hot water tank, and calculates the estimated volume of the main outlet water based on the liquid level baseline trend line. The inversion processing module acquires a dynamic inversion model for inverting pressure pulsation signals into equivalent water flow contributions, and processes the pressure pulsation signals based on the model to generate the equivalent water replenishment volume for the auxiliary channel. A dynamic inversion model is obtained to convert pressure pulsation signals into equivalent flow contributions. Based on this model, the pressure pulsation signals are processed to generate the equivalent make-up water volume for the auxiliary channel, including: The fluctuation patterns representing minute flow rates are identified from pressure pulsation signals, and pressure feature vectors are generated. The pressure feature vector is input into the dynamic inversion model to calculate the instantaneous equivalent flow rate. The instantaneous equivalent flow rate is integrated over time to generate the equivalent water replenishment volume for the auxiliary road. The intelligent decision module performs dynamic weighted collaborative correction based on the estimated main outlet water volume, the equivalent replenishment water volume of the auxiliary road, and the calmness index of the liquid surface, to obtain the corrected outlet water volume. The intelligent decision module performs dynamic weighted collaborative correction based on the estimated main outlet water volume, the equivalent replenishment volume of the auxiliary outlet, and the surface calmness index to obtain the corrected outlet water volume, including: Based on the calmness index of the liquid surface, the weights of the main road and the auxiliary road are dynamically allocated according to the estimated volume of water discharged from the main road and the equivalent water replenishment volume of the auxiliary road. When the liquid level calmness index indicates that the liquid level is stable, increase the weight of the main road and decrease the weight of the auxiliary road. When the liquid level calmness index indicates liquid level fluctuations, reduce the weight of the main road and increase the weight of the auxiliary road. The estimated effluent volume of the main road and the equivalent water replenishment of the auxiliary road are weighted and fused based on the weight of the main road and the weight of the auxiliary road to generate the corrected effluent volume. The cumulative output module accumulates the corrected output water volume to generate the cumulative water replenishment volume.

Images