A sewer network monitoring sensor based on a dual mode and a monitoring method thereof

Abstract

The application discloses a kind of based on dual-mode sewer network monitoring sensor and monitoring method thereof, belong to sensor technical field, including: ultrasonic unit, radar unit and processing unit, the ultrasonic unit and radar unit are electrically connected with processing unit;Processing unit is used to: acquisition ultrasonic unit's sound wave original signal and radar unit's radar original signal, according to the preset quick scene estimation rule determines the scene type of pipe network currently, according to scene type creates adaptive filtering strategy and dynamic fusion strategy, according to adaptive filtering strategy to sound wave original signal and radar original signal carry out filtering processing, obtain ultrasonic monitoring value and radar monitoring value, the ultrasonic monitoring value and radar monitoring value all include liquid level and flow rate;Based on dynamic fusion strategy to ultrasonic monitoring value and radar monitoring value carry out fusion, obtain sewer flow rate and liquid level.The application has the advantages that anti-interference ability is strong and fusion precision is high.

Description

Technical Field

[0001] This invention belongs to the field of sensor technology, specifically relating to a dual-mode drainage network monitoring sensor and its monitoring method. Background Technology

[0002] Municipal drainage networks are the core infrastructure of urban water circulation systems. Real-time and accurate monitoring of parameters such as flow and level is of great significance for flood control scheduling, network operation and maintenance, and water pollution prevention and control. With the advancement of smart city construction, traditional monitoring methods such as manual inspection and fixed-point sampling are no longer sufficient to meet the needs of refined management due to their poor real-time performance, limited coverage, and susceptibility to environmental interference. There is an urgent need for efficient and stable automated monitoring technologies.

[0003] Currently, the mainstream technologies for monitoring flow in drainage pipe networks often employ single-mode sensors, such as radar velocimeters and ultrasonic velocimeters. Radar velocimeters operate based on the principle of electromagnetic wave reflection, offering the advantage of non-contact measurement and strong resistance to interference from silt and floating debris. However, under low flow velocity and low liquid level conditions, the weak reflected signal and significant beam diffusion effect can easily lead to flow velocity measurement point shifts, making accuracy difficult to guarantee. Ultrasonic velocimeters, relying on the time-of-flight method (converting flow velocity by measuring the round-trip time difference of ultrasonic waves), offer high measurement accuracy and fast response, making them suitable for low to medium flow velocity scenarios. However, they are sensitive to water bubbles and pipe scaling. Bubbles can cause signal amplitude distortion, while scaling can cause ultrasonic wave attenuation and propagation path shifts, leading to measurement errors. Furthermore, electromagnetic interference can easily affect signal stability.

[0004] To compensate for the shortcomings of single-mode technology, radar-ultrasonic dual-mode sensors are gradually emerging. Their core idea is to combine the complementary advantages of the two modes to improve adaptability to complex operating conditions. However, existing dual-mode technology still faces many technical bottlenecks:

[0005] First, the fusion mechanism is simple: it mostly adopts fixed weight weighted fusion, lacks collaborative correction logic based on signal characteristics and operating conditions, and cannot dynamically adapt to different operating states such as full pipe / non-full pipe, low flow rate / rain and flood impact, thus limiting the fusion accuracy.

[0006] Secondly, the anti-interference algorithms lack specificity: the filtering strategies are mostly "one-size-fits-all" designs, without customizing solutions based on the differentiated interference characteristics of radar and ultrasonic signals (such as radar sediment interference and ultrasonic bubble interference), and cannot adapt to the interference compensation needs of long-term aging chemical conditions such as pipe scaling. Summary of the Invention

[0007] The purpose of this invention is to provide a dual-mode drainage network monitoring sensor and its monitoring method to solve the problems of low fusion accuracy and insufficient anti-interference ability in the existing technology.

[0008] To achieve the above objectives, the present invention adopts the following technical solution:

[0009] In a first aspect, the present invention provides a dual-mode drainage network monitoring sensor, the sensor comprising: an ultrasonic unit, a radar unit, and a processing unit, wherein the ultrasonic unit and the radar unit are both electrically connected to the processing unit;

[0010] The processing unit is used to: acquire the raw acoustic signal from the ultrasonic unit and the raw radar signal from the radar unit; determine the current scene type of the pipeline network according to the preset rapid scene prediction rules; create an adaptive filtering strategy according to the scene type; filter the raw acoustic signal and the raw radar signal according to the adaptive filtering strategy to obtain ultrasonic monitoring values ​​and radar monitoring values, both of which include liquid level and flow velocity; create a dynamic fusion strategy according to the scene type; and fuse the ultrasonic monitoring values ​​and the radar monitoring values ​​based on the dynamic fusion strategy to obtain the flow velocity and liquid level of the drainage pipe.

[0011] Preferably, the sensor further includes a communication unit, which is electrically connected to the processing unit;

[0012] The processing unit is equipped with a lightweight prediction model, which is used to predict the flow rate and liquid level in future periods using the flow rate and liquid level observation sequence as input, and obtain the predicted values ​​of flow rate and liquid level. The flow rate and liquid level observation sequence is created by fusing the flow rate and liquid level of the drainage pipe using a dynamic fusion strategy.

[0013] The communication unit is used to upload the flow velocity and liquid level of the drainage pipe and the predicted values ​​of flow velocity and liquid level to the monitoring terminal, which includes at least a cloud service terminal and a user terminal.

[0014] Secondly, the present invention provides a monitoring method for the aforementioned dual-modal drainage network monitoring sensor, the method comprising:

[0015] Acquire the raw acoustic signals from the ultrasonic unit and the raw radar signals from the radar unit;

[0016] The current scenario type of the pipeline network is determined based on the preset rapid scenario prediction rules;

[0017] An adaptive filtering strategy is created based on the scenario type. The original acoustic signal and radar signal are filtered according to the adaptive filtering strategy to obtain ultrasonic monitoring values ​​and radar monitoring values. Both ultrasonic monitoring values ​​and radar monitoring values ​​include liquid level and flow rate.

[0018] A dynamic fusion strategy is created based on the scenario type. The ultrasonic monitoring values ​​and radar monitoring values ​​are fused based on the dynamic fusion strategy to obtain the flow velocity and liquid level of the drainage pipe.

[0019] Preferably, the scenario types include: standard scenarios and non-standard scenarios. The non-standard scenarios are further divided into low flow rate and low liquid level scenarios, scaling scenarios, and rainstorm impact scenarios. The rapid scenario prediction rule is as follows:

[0020] The raw acoustic and radar signals are preprocessed. Based on the preprocessed raw acoustic and radar signals, the estimated flow rate and estimated liquid level are determined. Based on the estimated flow rate and liquid level, the flow rate change rate and liquid level rise rate are determined.

[0021] Determine whether the flow velocity mutation rate reaches the preset mutation rate and whether the liquid level rise rate reaches the preset rise rate. If so, the scenario type is a rainstorm impact scenario under non-standard scenarios.

[0022] If not, extract the attenuation coefficient of the original acoustic signal and the peak offset of the original radar signal, and determine whether the attenuation coefficient exceeds the first preset coefficient and whether the peak offset exceeds the preset frequency. If so, the scenario type is a scaling scenario under non-standard scenarios.

[0023] If not, determine whether the estimated flow rate is less than the preset flow rate and whether the estimated liquid level is less than the preset liquid level. If yes, the scenario type is a low flow rate and low liquid level scenario under non-standard scenarios. If no, the scenario type is a standard scenario.

[0024] Preferably, the adaptive filtering strategy is:

[0025] In a standard scenario, basic filtering rules are established, and the original acoustic and radar signals are filtered using these rules to obtain the ultrasonic and radar monitoring values ​​in the standard scenario.

[0026] In non-standard scenarios, the basic filtering rules are fine-tuned, and the original acoustic and radar signals are filtered using the fine-tuned basic filtering rules to obtain the ultrasonic and radar monitoring values ​​under non-standard scenarios.

[0027] Preferably, the basic filtering rule for the original radar signal is as follows: the original radar signal is acquired at a preset sampling frequency, beam scattering intensity and spectrum data are extracted, and the proportion of spectral clutter is statistically analyzed based on the spectrum data; it is determined whether the beam scattering intensity is greater than the preset intensity and whether the proportion of spectral clutter is greater than the preset proportion. If so, the original radar signal is filtered using a 128th-order FIR low-pass filter; if not, the original radar signal is filtered using a 32nd-order FIR low-pass filter. The filtered original radar signal is then subjected to self-healing verification. After the self-healing verification is passed, the radar monitoring value under the standard scenario is obtained.

[0028] Preferably, fine-tuning the basic filtering rules of the original radar signal includes:

[0029] In low flow rate and low liquid level scenarios, the filtering parameters of the FIR low-pass filter in the standard scenario are adjusted, and the radar monitoring value obtained in the adjusted standard scenario is used as the radar monitoring value in the low flow rate and low liquid level scenario. The filtering parameters of the FIR low-pass filter include at least the order and the cutoff frequency.

[0030] In the scaling scenario, the frequency offset is input into the pre-built frequency-thickness fitting equation to obtain the scaling thickness. The radar monitoring value is corrected according to the scaling thickness and the pre-built correction model. The self-healing verification is performed again on the corrected radar monitoring value. After the self-healing verification passes, the radar monitoring value in the scaling scenario is obtained.

[0031] In the rainstorm impact scenario, the filtering parameters of the FIR low-pass filter and the self-healing verification parameters in the standard scenario are adjusted, and the radar monitoring value obtained in the adjusted standard scenario is used as the radar monitoring value in the rainstorm impact scenario.

[0032] Preferably, the basic filtering rule for the original acoustic signal is:

[0033] The raw ultrasonic signal is acquired at a preset sampling frequency, and the time-domain amplitude is extracted.

[0034] The signal attenuation coefficient is determined based on the time-domain amplitude, and the interference type is determined based on the signal attenuation coefficient. The interference type is either bubble interference or flow field anomaly. When the interference type is bubble interference, a second-order Volterra filter is used to filter the original acoustic signal. When the interference type is flow field anomaly, a rejection-interpolation rule is used to process the original acoustic signal to obtain the processed original acoustic signal. The stability of the processed original acoustic signal is adjusted based on the AGC closed-loop adjustment rule to obtain the adjusted original acoustic signal. The ultrasonic monitoring value under the standard scenario is determined based on the adjusted original acoustic signal.

[0035] Preferably, the fine-tuning of the basic filtering rules for the original sound wave signal is as follows:

[0036] In low flow rate and low liquid level scenarios, the filtering parameters of the Volterra filter and the adjustment parameters of the AGC closed-loop adjustment rule in the standard scenario are adjusted, and the ultrasonic monitoring value obtained in the adjusted standard scenario is used as the ultrasonic monitoring value in the low flow rate and low liquid level scenario. The filtering parameters of the Volterra filter include at least the order.

[0037] In the scaling scenario, the threshold for judging the interference type in the standard scenario is reconstructed and the adjustment parameters of the AGC closed-loop regulation rule are adjusted to obtain the ultrasonic monitoring value in the scaling scenario.

[0038] In the rainstorm impact scenario, the Volterra filter in the standard scenario is replaced with the Butterworth filter, and the adjustment parameters of the AGC closed-loop regulation rule are adjusted to obtain the ultrasonic monitoring values ​​under the rainstorm impact scenario.

[0039] Preferably, the dynamic fusion strategy is as follows:

[0040] In the standard scenario, the ultrasonic monitoring values ​​and radar monitoring values ​​are fused according to the extended Kalman filter algorithm to obtain the flow velocity and liquid level of the drainage pipe in the standard scenario.

[0041] In non-standard scenarios, the extended Kalman filter algorithm is optimized, and the optimized extended Kalman filter algorithm is used to fuse ultrasonic monitoring values ​​and radar monitoring values ​​to obtain the flow velocity and liquid level of the drainage pipe in non-standard scenarios.

[0042] The beneficial effects of this invention are:

[0043] 1. After determining the current scene type of the pipeline network through the rapid scene prediction rule, the present invention creates an adaptive filtering strategy based on the scene type. The adaptive filtering strategy can use different filtering methods for the original acoustic signal and radar signal under different scenes to achieve differentiated filtering processing and improve anti-interference capability.

[0044] 2. The present invention also creates a dynamic fusion strategy based on scene type. The dynamic fusion strategy is used to fuse ultrasonic monitoring values ​​and radar monitoring values, which can adapt to different fusion scenarios and improve fusion accuracy. Attached Figure Description

[0045] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings:

[0046] Figure 1 This is a block diagram of a dual-modal drainage network monitoring sensor provided in one embodiment of the present invention;

[0047] Figure 2 This is a flowchart of a monitoring method provided in one embodiment of the present invention. Detailed Implementation

[0048] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in conjunction with the accompanying drawings and descriptions of the embodiments or the prior art. Obviously, the following description of the structure of the accompanying drawings is 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. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0049] Example 1

[0050] Figure 1 This is a block diagram of a dual-modal drainage network monitoring sensor provided in one embodiment of the present invention. Figure 1 As shown, this embodiment provides a dual-mode drainage network monitoring sensor, which includes an ultrasonic unit, a radar unit, and a processing unit. The ultrasonic unit and the radar unit are both electrically connected to the processing unit.

[0051] The ultrasonic unit in this embodiment includes a transmitting circuit, a receiving circuit, and an AGC (Automatic Gain Control) control circuit. The transmitting circuit is connected to the processing unit, and the receiving circuit is connected to the processing unit through the AGC control circuit. The processing unit in this embodiment can be an STM32H7 series microprocessor.

[0052] The transmitting circuit consists of a high-frequency oscillator, a power amplifier, and a driving tube. The oscillator generates a high-frequency excitation signal of 40kHz to 200kHz (preferably 100kHz, to balance penetration and anti-interference). After being amplified by the power amplifier (gain 20 to 40dB), the ultrasonic transducer is controlled by the driving tube to emit sound waves.

[0053] The receiving circuit includes a preamplifier, a bandpass filter, and a detector circuit. The weak echo signal (mV level) received by the transducer is amplified to V level by the preamplifier (built with a low-noise operational amplifier, noise figure ≤2dB), then filtered out by the bandpass filter (center frequency is the same as the transmission frequency, bandwidth ±5kHz) to remove electromagnetic interference and water clutter. Finally, the signal amplitude and timing characteristics are extracted by the detector circuit to provide the basic signal for time difference calculation.

[0054] The AGC control circuit is used to dynamically adjust the amplifier's amplification factor (adjustment range 0~30dB, step size 0.2~0.5dB) to form a "detection-calculation-adjustment-feedback" closed loop, compensating for signal attenuation caused by scaling and bubbles.

[0055] The radar unit in this embodiment includes a radio frequency transmitting circuit and a radio frequency receiving circuit, both of which are connected to the processing unit.

[0056] The radio frequency (RF) transmitting circuit includes a frequency synthesizer, a power amplifier, and a transmitting antenna. The frequency synthesizer (such as AD9910) generates a 24GHz / 77GHz RF signal (24GHz is preferred for drainage scenarios, balancing measurement distance and equipment cost). The signal frequency is linearly modulated over time, with a modulation bandwidth of 1~2GHz. The power amplifier amplifies the signal power to 10~20dBm, and the signal is directionally transmitted via a microstrip antenna with a gain ≥15dBi to suppress beam spread effects (suitable for low liquid level scenarios).

[0057] The radio frequency (RF) receiving circuit consists of a receiving antenna, a mixer, and an intermediate frequency (IF) amplifier. The receiving antenna captures the echo signal reflected from the water body, which, along with the local oscillator signal from the transmitting circuit, is fed into the mixer to output an IF signal (10~100MHz). The IF amplifier (low-noise design, gain 40~60dB) amplifies and conditions the IF signal, filters out RF interference, and provides a clear signal source for spectrum analysis.

[0058] In this embodiment, the processing unit is used to: collect the raw acoustic signals from the ultrasonic unit and the raw radar signals from the radar unit; determine the current scenario type of the pipeline network according to a preset rapid scenario prediction rule; the scenario types in this embodiment include: standard scenarios and non-standard scenarios, and the non-standard scenarios are divided into low flow rate and low liquid level scenarios, scaling scenarios, and rainstorm impact scenarios; create an adaptive filtering strategy and a dynamic fusion strategy according to the scenario type; filter the raw acoustic signals and raw radar signals according to the adaptive filtering strategy to obtain ultrasonic monitoring values ​​and radar monitoring values, both of which include liquid level and flow rate; and fuse the ultrasonic monitoring values ​​and radar monitoring values ​​based on the dynamic fusion strategy to obtain the flow rate and liquid level of the drainage pipe.

[0059] In this embodiment, after determining the current scene type of the pipeline network through rapid scene prediction rules, an adaptive filtering strategy is created based on the scene type. The adaptive filtering strategy can use different filtering methods for the original acoustic and radar signals under different scenes to achieve differentiated filtering processing and improve anti-interference capability. In addition, a dynamic fusion strategy is also created based on the scene type. The dynamic fusion strategy is used to fuse the ultrasonic monitoring values ​​and radar monitoring values, which can adapt to different fusion scenarios and improve fusion accuracy.

[0060] As a further optimization of this embodiment, the sensor further includes: a communication unit, which is electrically connected to the processing unit;

[0061] The processing unit is equipped with a lightweight prediction model, which is used to predict the flow rate and liquid level in future periods using the flow rate and liquid level observation sequence as input, and obtain the predicted values ​​of flow rate and liquid level. The flow rate and liquid level observation sequence is created by fusing the flow rate and liquid level of the drainage pipe using a dynamic fusion strategy.

[0062] The communication unit is used to upload the flow velocity and liquid level of the drainage pipe and the predicted values ​​of flow velocity and liquid level to the monitoring terminal, which includes at least a cloud service terminal and a user terminal.

[0063] The lightweight prediction model in this embodiment can be a hybrid model of LSTM (Long Short-Term Memory) and CNN (Convolutional Neural Network). The lightweight prediction model is used to predict the flow rate and liquid level in the future. Based on the flow rate and liquid level in the future, the corresponding scene type is identified, and an early warning is sent to the monitoring terminal based on the flow rate and liquid level in the future and the corresponding scene type.

[0064] Example 2

[0065] Figure 2 This is a flowchart of a monitoring method provided in one embodiment of the present invention. Figure 2 As shown, this embodiment provides a monitoring method, which is applied to the dual-mode drainage network monitoring sensor in Embodiment 1. The method includes steps S10 to S40.

[0066] Step S10: Collect the original acoustic signal from the ultrasonic unit and the original radar signal from the radar unit. In this embodiment, the original acoustic signal from the ultrasonic unit is the received signal output by the ultrasonic unit under default parameters (including parameters such as the frequency of the high-frequency excitation signal generated by the oscillator, the amplification factor of the power amplifier, and the echo amplification factor, which have been listed in detail in Embodiment 1). Similarly, the original radar signal is the received signal output by the radar unit under default parameters (including parameters such as the frequency of the frequency synthesizer, the frequency modulation bandwidth, the signal power factor, and the antenna gain).

[0067] Step S20: Determine the current scenario type of the pipeline network according to the preset rapid scenario prediction rules.

[0068] In this embodiment, the scenario types include: standard scenarios and non-standard scenarios. The non-standard scenarios are further divided into three types: low flow rate and low liquid level scenarios, scaling scenarios, and rainstorm impact scenarios. The rapid scenario prediction rule in this embodiment is:

[0069] First, the raw acoustic and radar signals are preprocessed. Based on the preprocessed raw acoustic and radar signals, the estimated flow rate and estimated liquid level are determined. Then, based on the estimated flow rate and estimated liquid level, the flow rate change rate and liquid level rise rate are determined.

[0070] The preprocessing involves performing lightweight denoising on the original acoustic and radar signals. For example, for the original acoustic signal, the amplitude thresholding method (removing extreme values ​​with amplitude < 0.1V) + 3-point moving average is used to initially suppress instantaneous bubble interference in order to retain the core time difference information. For the original radar signal, a 32nd-order FIR low-pass filter (cutoff frequency 10kHz, fast calculation) is used to remove pulse interference and retain the flow velocity time sequence trend.

[0071] In this embodiment, the steps for determining the estimated flow velocity and estimated liquid level based on the pre-processed acoustic wave original signal and radar original signal are as follows: First, determine the first flow velocity corresponding to the radar original signal based on the radar amplitude-flow velocity calibration curve, and locate the first liquid level based on the beam reflection position of the radar original signal; extract the time difference of the acoustic wave original signal, determine the second flow velocity based on the time difference, and determine the second liquid level based on the cross-section of the pipeline network; determine the estimated flow velocity based on the first flow velocity and the second flow velocity, determine the estimated liquid level based on the first liquid level and the second liquid level, determine the flow velocity change rate based on the estimated flow velocity, and determine the liquid level rise rate based on the estimated liquid level.

[0072] The amplitude-flow velocity calibration curve can be obtained by fitting the experimental calibration. The second flow velocity can be calculated based on the ratio between the probes and the time difference of the ultrasonic unit. The average value of the first liquid level and the second liquid level is used as the estimated liquid level, and the average value of the first flow velocity and the second flow velocity is used as the estimated flow velocity.

[0073] In this embodiment, the flow rate change rate and the liquid level rise rate are calculated based on the estimated liquid level and liquid level of the last 10 sampling cycles (0.1s). If any modal signal is interrupted, the flow rate or liquid level is estimated by using the data of another modal. If both modal signals are abnormal, the flow rate or liquid level of the previous cycle is used. If there are abnormalities for 3 consecutive cycles, an equipment fault warning is triggered.

[0074] Then, it is determined whether the flow velocity mutation rate reaches the preset mutation rate and whether the liquid level rise rate reaches the preset rise rate. If so, the scenario type is a rainstorm impact scenario under a non-standard scenario.

[0075] If not, extract the attenuation coefficient of the original acoustic signal and the peak offset of the original radar signal, and determine whether the attenuation coefficient exceeds the first preset coefficient and whether the peak offset exceeds the preset frequency. If so, the scenario type is a scaling scenario under non-standard scenarios.

[0076] If not, determine whether the estimated flow rate is less than the preset flow rate and whether the estimated liquid level is less than the preset liquid level. If yes, the scenario type is a low flow rate and low liquid level scenario under non-standard scenarios. If no, the scenario type is a standard scenario.

[0077] The specific principles for dividing the land are as follows:

[0078] Rainfall impact scenario: When the flow rate change rate is ≥50% / min (preset change rate) and the liquid level rise rate is ≥0.1m / min (preset rise rate), it is identified as this scenario;

[0079] Scaling scenario: When the attenuation coefficient is greater than 1.2 dB / cm (first preset coefficient) and the peak frequency offset is greater than 100 Hz (preset frequency), this scenario is identified.

[0080] Low flow rate and low liquid level scenario: When the estimated flow rate is ≤0.02m / s and the estimated liquid level is ≤0.1D (D is the inner diameter of the pipe), it is determined to be this scenario.

[0081] When all of the above conditions are met, it is determined to be a standard scenario.

[0082] Step S30: Create an adaptive filtering strategy according to the scene type, and filter the original acoustic signal and radar signal according to the adaptive filtering strategy to obtain ultrasonic monitoring value and radar monitoring value. The ultrasonic monitoring value and radar monitoring value both include liquid level and flow rate.

[0083] Specifically, the adaptive filtering strategy is as follows:

[0084] In a standard scenario, basic filtering rules are established, and these rules are used to filter the original acoustic and radar signals to obtain the ultrasonic and radar monitoring values ​​under the standard scenario.

[0085] In non-standard scenarios, the basic filtering rules are fine-tuned, and the original acoustic and radar signals are filtered using the fine-tuned basic filtering rules to obtain the ultrasonic and radar monitoring values ​​under non-standard scenarios.

[0086] The basic filtering rules for the original radar signal are as follows: the original radar signal is acquired at a preset sampling frequency, the beam scattering intensity and spectrum data are extracted, and the proportion of spectral clutter is calculated based on the spectrum data; it is determined whether the beam scattering intensity is greater than the preset intensity and whether the proportion of spectral clutter is greater than the preset proportion. If so, the original radar signal is filtered using a 128th-order FIR low-pass filter; if not, the original radar signal is filtered using a 32nd-order FIR low-pass filter. The original radar signal after filtering is then subjected to self-healing verification. After the self-healing verification is passed, the radar monitoring value under the standard scenario is obtained.

[0087] Specifically, the preset sampling frequency in this embodiment is 100Hz. Beam scattering intensity and spectrum data are extracted simultaneously. Based on the beam scattering intensity and spectrum data, sediment interference can be determined. For example, by comparing the scattering intensity I with the reference value I0 (clean water calibration value), if I ≥ 0.3I0 (preset intensity) and the proportion of spectral clutter is > 20%, it is determined to be sediment interference. A 128th-order FIR low-pass filter (Hanning window design, cutoff frequency 3kHz) is activated to filter the original radar signal to suppress sediment clutter interference. Otherwise, a 32nd-order FIR low-pass filter is used to filter the original radar signal to improve processing efficiency.

[0088] After filtering is completed, a self-healing verification is performed. The verification steps are as follows:

[0089] First, the volatility is calculated using a 3-point moving average. The functional expression for the volatility is as follows:

[0090] ;

[0091] In the formula, For fluctuations, Let be the amplitude of the radar signal at time i. This is the average of the radar signal amplitudes at three points.

[0092] If the ratio of the fluctuation to the average of the radar signal amplitudes at the three points is greater than the validity fluctuation threshold (preferably 5%), it indicates that the collected signal is invalid and requires self-healing re-collection (re-collection interval 0.02s, maximum re-collection 3 times). If the 3 re-collections are invalid, the data from the previous valid time point will be used as the replacement.

[0093] In non-standard scenarios, radar signals are prone to deviation from the flow velocity measurement point due to weak reflected signals and significant beam spread effects under low flow velocity and low liquid level conditions, making it difficult to guarantee accuracy. Therefore, to improve the accuracy of flow velocity and liquid level detection in non-standard scenarios, this embodiment fine-tunes and optimizes the basic filtering rules according to different scenarios to improve measurement accuracy. Specifically, as follows:

[0094] In low-flow-rate and low-liquid-level scenarios, the filtering parameters of the FIR low-pass filter in the standard scenario are adjusted. The radar monitoring value obtained under the adjusted standard scenario is used as the radar monitoring value under the low-flow-rate and low-liquid-level scenario. The filtering parameters of the FIR low-pass filter include at least the order and cutoff frequency. For example, the FIR filter order is reduced to 64th order, and the cutoff frequency is lowered to 1kHz to reduce distortion of the low-liquid-level reflected signal; the signal validity fluctuation threshold is relaxed to 8% to avoid data breakage caused by frequent re-sampling, and the re-sampling interval is maintained at 0.02s.

[0095] In the scaling scenario, the frequency offset is input into the pre-built frequency-thickness fitting equation to obtain the scaling thickness. The radar monitoring value is corrected based on the scaling thickness and the pre-built correction model. The self-healing verification is performed again on the corrected radar monitoring value. After the self-healing verification passes, the radar monitoring value in the scaling scenario is obtained.

[0096] Because the dielectric constant of the scale layer (mainly composed of calcium carbonate and rust) on the inner wall of the pipe differs significantly from that of the water, the radar beam is refracted after incident, deviating from the ideal propagation path. This leads to a shift in the flow velocity measurement point (the shift increases with the thickness of the scale). In this case, the beam refraction characteristic parameter (frequency shift) is extracted through spectral analysis. The frequency shift is the difference between the peak value of the spectrum and the peak value under clean operating conditions. The frequency shift is positively correlated with the scale thickness, and a frequency-thickness fitting equation can be established through prior laboratory calibration. In this embodiment, after extracting the frequency shift, it is input into the frequency-thickness fitting equation to calculate the scale thickness. When the scale thickness exceeds 0.25 cm, a correction model is activated to correct the radar monitoring value.

[0097] The modified model in this embodiment is constructed using Snell's law, and the functional relationship of the modified model in this embodiment is as follows:

[0098]

[0099] In the formula, To correct the correction coefficients in the model output, L represents the ideal path length without scaling. s d is the actual propagation path length of the beam. s For scale thickness, The incident angle of the radar beam. Let h be the angle of refraction and h be the actual liquid level height, which can be determined based on the estimated liquid level. The relationship between the radar beam incident angle and the angle of refraction is:

[0100] ;

[0101] In the formula, n 12 The refractive index of the scale layer and the water surface is determined based on the dielectric constants of the scale layer and the water surface.

[0102] After correcting the radar monitoring values ​​using a correction factor, the fluctuation of the corrected radar monitoring values ​​is effectively verified using a 3-point sliding square. If the fluctuation is large, the refractive index n is adjusted. 12 To accommodate different scaling components.

[0103] In the scenario of rainstorm impact, the filtering parameters of the FIR low-pass filter and the parameters of the self-healing verification in the standard scenario are adjusted, and the radar monitoring values ​​obtained in the adjusted standard scenario are used as the radar monitoring values ​​in the rainstorm impact scenario. For example, the cutoff frequency is dynamically adjusted according to the flow velocity change rate, and is increased to 5kHz when the change rate is ≥50% / min to quickly track signal changes; the re-sampling interval in the self-healing verification is shortened to 0.01s to ensure data continuity during the impact, and the effectiveness fluctuation threshold is maintained at 5%.

[0104] The basic filtering rules for the original acoustic signal are as follows: The original ultrasonic signal is acquired at a preset sampling frequency, and the time-domain amplitude is extracted. The signal attenuation coefficient is determined based on the time-domain amplitude, and the interference type is determined based on the signal attenuation coefficient. The interference type is either bubble interference or flow field anomaly. When the interference type is bubble interference, a second-order Volterra filter is used to filter the original acoustic signal. When the interference type is flow field anomaly, a rejection-interpolation rule is used to process the original acoustic signal. The processed original acoustic signal is then used to adjust its stability based on the AGC closed-loop adjustment rule to obtain the adjusted original acoustic signal. Finally, the ultrasonic monitoring value under standard scenarios is determined based on the adjusted original acoustic signal.

[0105] In this embodiment, the original ultrasonic signal is also acquired at a preset sampling frequency of 100Hz, the time-domain amplitude is extracted, and then the attenuation coefficient is calculated based on the time-domain amplitude of the transmitting end and the time-domain amplitude of the receiving end. The expression for the attenuation coefficient is as follows:

[0106] ;

[0107] In the formula, A is the attenuation coefficient. in A0 represents the time-domain amplitude at the receiving end, and A0 represents the time-domain amplitude at the transmitting end.

[0108] When the attenuation coefficient is greater than 0.8 dB / cm, it is determined to be bubble interference; when the attenuation coefficient is less than or equal to 0.8 dB / cm, it is determined to be flow field anomaly (including vortices and backflow). For bubble interference, a second-order Volterra filter is used to correct the nonlinear signal distortion caused by bubbles.

[0109] The elimination-interpolation rule in this embodiment is to directly eliminate invalid time difference data (e.g., adjacent fluctuations > 0.1ms) and use linear interpolation to supplement valid time differences.

[0110] After filtering, the signal stability needs to be adjusted using AGC closed-loop control rules. The adjustment steps for AGC closed-loop control rules are as follows:

[0111] Step a1: Acquire the real-time amplitude of the filtered signal;

[0112] Step a2: Calculate the deviation between the real-time amplitude and the reference amplitude, and determine whether the deviation is less than the deviation threshold. If not, calculate the target gain based on the real-time amplitude and the reference amplitude. The expression for the target gain is:

[0113] ;

[0114] In the formula, G is the target gain, and V ref V is the reference amplitude. in This is the real-time amplitude.

[0115] Step a3: Constrain the gain adjustment range to 0~20dB, with an adjustment step size of 0.5dB, adjust the target gain, and after the target gain is adjusted, generate a control command to drive the AGC control circuit to dynamically adjust the amplifier's amplification factor.

[0116] Step a4: Reacquire the adjusted signal amplitude and determine whether the deviation between the adjusted signal amplitude and the reference amplitude is less than the deviation threshold. If yes, it means that the adjustment is stable; if not, repeat steps a3-a4 (Note that the maximum number of iterations can be repeated is 3. If it is still not stable after more than 3 iterations, the gain value of the previous cycle will be used and marked as "gain adjustment abnormal").

[0117] In non-standard scenarios, the method for fine-tuning the basic filtering rules of the original sound wave signal is as follows:

[0118] In low-flow-rate and low-liquid-level scenarios, the filtering parameters of the Volterra filter and the adjustment parameters of the AGC closed-loop control rules in the standard scenario are adjusted. The ultrasonic monitoring values ​​obtained in the adjusted standard scenario are used as the ultrasonic monitoring values ​​in the low-flow-rate and low-liquid-level scenario. The filtering parameters of the Volterra filter include at least the order. For example, the order of the first-order kernel function of the Volterra filter is increased to 24th order to enhance the signal distortion correction capability caused by static pressure interference; and the reference amplitude is reduced and the adjustment range is expanded to 0~30dB to target weak signals at low flow rates (V in <0.5V), the adjustment step size is adaptively reduced to 0.2dB to improve the amplitude compensation accuracy; the upper limit of the number of iterations is increased to 5 times to ensure that the weak signal amplitude is stable within the range of ±0.05V of the reference value.

[0119] In scaling scenarios, the threshold for judging interference types in standard scenarios is reconstructed, and the adjustment parameters of the AGC closed-loop regulation rule are adjusted to obtain ultrasonic monitoring values ​​under scaling conditions. For example, the threshold for the attenuation coefficient is adjusted to 1.2 dB / cm, and phase stability (phase deviation ≤ ±5°) is used to distinguish between scaling and bubble interference to avoid misjudgment; the adjustment step size is fixed at 0.2 dB to accurately compensate for the gradual signal attenuation caused by scaling, and gain trend analysis is performed. If the cumulative gain increase is >3 dB for 5 consecutive cycles, scaling is judged to be aggravated, and the scaling thickness is recalculated.

[0120] In rainstorm impact scenarios, the Volterra filter in the standard scenario is replaced with a Butterworth filter, and the adjustment parameters of the AGC closed-loop adjustment rule are adjusted to obtain ultrasonic monitoring values ​​under rainstorm impact scenarios. For example, the Volterra filter is paused (to avoid computational delay > 0.005s), and a 5th-order Butterworth low-pass filter (cutoff frequency 8kHz) is switched to balance speed and anti-interference; and the AGC closed-loop adjustment rule adopts a "prediction-adjustment" mode, that is, the target gain is predicted based on the amplitude change trend of the first two cycles, reducing the number of iterations (the upper limit of iteration is 2), while the adjustment range is kept at 0~20dB with a step size of 0.5dB to cope with the sudden change in signal amplitude caused by the impact, ensuring that the amplitude is stable within the range of 2V±0.2V after adjustment to meet real-time requirements.

[0121] Therefore, this embodiment creates an adaptive filtering strategy based on the scenario type. The adaptive filtering strategy can use different filtering methods for the original acoustic signal and radar signal under different scenarios to achieve differentiated filtering processing and improve anti-interference capability.

[0122] Step S40: Create a dynamic fusion strategy based on the scenario type, and fuse the ultrasonic monitoring values ​​and radar monitoring values ​​based on the dynamic fusion strategy to obtain the flow velocity and liquid level of the drainage pipe.

[0123] In this embodiment, the dynamic fusion strategy is as follows:

[0124] In the standard scenario, the ultrasonic monitoring values ​​and radar monitoring values ​​are fused according to the extended Kalman filter algorithm to obtain the flow velocity and liquid level of the drainage pipe in the standard scenario.

[0125] In non-standard scenarios, the extended Kalman filter algorithm is optimized, and the optimized extended Kalman filter algorithm is used to fuse ultrasonic monitoring values ​​and radar monitoring values ​​to obtain the flow velocity and liquid level of the drainage pipe in non-standard scenarios.

[0126] Specifically, the steps for fusing ultrasonic monitoring values ​​and radar monitoring values ​​in a standard scenario are as follows:

[0127] Step b1: Construct the state vector of the extended Kalman filter algorithm based on the average values ​​of the flow velocity in the ultrasonic monitoring values, the flow velocity in the radar monitoring values, and the liquid level in the ultrasonic monitoring values ​​and the liquid level in the radar monitoring values; and initialize the process noise covariance matrix, the observation noise covariance matrix, and the state transition matrix;

[0128] The state vector is: ,in, The flow velocity in the ultrasonic monitoring value. , where is the flow velocity in the radar monitoring value, h is the average of the liquid level in the ultrasonic monitoring value and the liquid level in the radar monitoring value, and T is the transpose sign;

[0129] The process noise covariance matrix Q is:

[0130] ;

[0131] The observation noise covariance matrix R is:

[0132] ;

[0133] The state transition matrix A is:

[0134] ;

[0135] In the formula, Δt is the time variable, with a value of 0.01s, and τ is the time constant, with a value of 1s.

[0136] Step b2: Construct a dynamic weight matrix. Determine the full pipe state based on the ratio of liquid level h to pipe inner diameter D. A full pipe state is defined as h / D ≥ 0.95. Assign weights w accordingly. r =0.6、w u =0.4; h / D < 0.95 indicates a non-full pipe, and the weight w is assigned accordingly. r =0.4、w u =0.6, and the liquid level weight is fixed at 0.5; therefore, the weight matrix W is:

[0137] .

[0138] Step b3: State prediction. Based on the optimal estimate from the previous time step, the state at time k is predicted using the state equation, and the prediction error covariance is calculated simultaneously.

[0139] The functional expression for the state at time k is:

[0140] ;

[0141] In the formula, The state at time k, This is the optimal estimate from the previous moment.

[0142] The functional expression for the prediction error covariance is:

[0143] ;

[0144] In the formula, Let k be the prediction error covariance. This represents the covariance of the prediction error at the previous moment.

[0145] Step b4: Observation update. First, calculate the Kalman gain, and then update the optimal state estimate to obtain the optimal flow velocity and liquid level. Use the optimal flow velocity and liquid level as the drainage pipe flow velocity and liquid level in the standard scenario.

[0146] Kalman gain K k The function expression is:

[0147] ;

[0148] In the formula, H is a 3rd order identity matrix.

[0149] The functional expression for updating the optimal state estimate is:

[0150] .

[0151] In non-standard scenarios, the extended Kalman filter algorithm is optimized using the following methods.

[0152] In low flow rate and low liquid level scenarios, the weight is adjusted to w u =0.7、w r =0.3 (ultrasonic signal is more stable); the process noise matrix is ​​optimized as follows:

[0153]

[0154] In the scaling scenario, a scaling error compensation term is constructed based on the attenuation coefficient and the spectral peak offset, and the optimal state estimate is corrected based on the scaling error compensation term; and the observation noise covariance matrix R is dynamically adjusted according to the scaling thickness, for example, the elements in the observation noise covariance matrix R are increased by 10% for every 0.1cm increase.

[0155] Among them, the scaling error compensation term △X is:

[0156] ;

[0157] In the formula, Δf is the spectral peak offset. This is the attenuation coefficient.

[0158] The corrected functional expression for the optimal state estimate is:

[0159] .

[0160] In the scenario of rainstorm impact, an update rule for the weight matrix is ​​constructed. The update period for the weight matrix is ​​shortened to 0.005s, and a 10ms sliding window is used to iterate the weights. At the same time, the time constant τ in the state transition matrix A is reduced to 0.5s to improve the ability of state prediction to track sudden changes in flow. The process noise covariance matrix Q is adjusted as follows:

[0161]

[0162] The process noise covariance matrix Q is adjusted to accommodate the high volatility of the data.

[0163] This embodiment establishes a scene recognition mechanism to dynamically adapt filtering parameters, fusion weights, and noise matrices for three special scenarios: low flow rate, scaling, and rainwater runoff. This breaks through the "one-size-fits-all" limitation of traditional algorithms and balances accuracy and real-time performance under different scenarios.

[0164] As a further optimization of this embodiment, the method further includes: constructing a lightweight prediction model, which can be a hybrid model of LSTM (Long Short-Term Memory) and CNN (Convolutional Neural Network); using the lightweight prediction model to predict the flow rate and liquid level in future periods; identifying the corresponding scene type based on the flow rate and liquid level in future periods; and issuing an early warning to the monitoring terminal based on the flow rate and liquid level in future periods and the corresponding scene type.

[0165] Example 3

[0166] This embodiment also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the monitoring method in Embodiment 2.

[0167] This embodiment also provides two computer-readable storage media on which a computer program is stored. When the program is executed by a processor, it implements the dual-modal monitoring method in Embodiment 1.

[0168] This embodiment utilizes an adaptive filtering strategy, which can employ different filtering methods for the original acoustic and radar signals under different scenarios to achieve differentiated filtering processing and improve anti-interference capabilities; and utilizes a dynamic fusion strategy to fuse ultrasonic monitoring values ​​and radar monitoring values, which can adapt to different fusion scenarios and improve fusion accuracy.

[0169] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0170] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A system that specifies functions in one or more boxes.

[0171] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A dual-modal drainage network monitoring sensor, characterized in that, The sensor includes an ultrasonic unit, a radar unit, and a processing unit, wherein the ultrasonic unit and the radar unit are both electrically connected to the processing unit. The processing unit is used to: acquire the raw acoustic signal from the ultrasonic unit and the raw radar signal from the radar unit; determine the current scene type of the pipeline network according to the preset rapid scene prediction rules; create an adaptive filtering strategy according to the scene type; filter the raw acoustic signal and the raw radar signal according to the adaptive filtering strategy to obtain ultrasonic monitoring values ​​and radar monitoring values, both of which include liquid level and flow velocity; create a dynamic fusion strategy according to the scene type; and fuse the ultrasonic monitoring values ​​and the radar monitoring values ​​based on the dynamic fusion strategy to obtain the flow velocity and liquid level of the drainage pipe.

2. The drainage network monitoring sensor based on dual-mode as described in claim 1, characterized in that, The sensor further includes a communication unit, which is electrically connected to the processing unit; The processing unit is equipped with a lightweight prediction model, which is used to predict the flow rate and liquid level in future periods using the flow rate and liquid level observation sequence as input, and obtain the predicted values ​​of flow rate and liquid level. The flow rate and liquid level observation sequence is created by fusing the flow rate and liquid level of the drainage pipe using a dynamic fusion strategy. The communication unit is used to upload the flow velocity and liquid level of the drainage pipe and the predicted values ​​of flow velocity and liquid level to the monitoring terminal, which includes at least a cloud service terminal and a user terminal.

3. A monitoring method applied to the dual-modal drainage network monitoring sensor described in claim 1 or 2, characterized in that, The method includes: Acquire the raw acoustic signals from the ultrasonic unit and the raw radar signals from the radar unit; The current scenario type of the pipeline network is determined based on the preset rapid scenario prediction rules; An adaptive filtering strategy is created based on the scenario type. The original acoustic signal and radar signal are filtered according to the adaptive filtering strategy to obtain ultrasonic monitoring values ​​and radar monitoring values. Both ultrasonic monitoring values ​​and radar monitoring values ​​include liquid level and flow rate. A dynamic fusion strategy is created based on the scenario type. The ultrasonic monitoring values ​​and radar monitoring values ​​are fused based on the dynamic fusion strategy to obtain the flow velocity and liquid level of the drainage pipe.

4. The monitoring method according to claim 3, characterized in that, The scenario types include: standard scenarios and non-standard scenarios. Non-standard scenarios are further divided into low flow rate and low liquid level scenarios, scaling scenarios, and rainstorm impact scenarios. The rapid scenario prediction rule is as follows: The raw acoustic and radar signals are preprocessed. Based on the preprocessed raw acoustic and radar signals, the estimated flow rate and estimated liquid level are determined. Based on the estimated flow rate and liquid level, the flow rate change rate and liquid level rise rate are determined. Determine whether the flow velocity mutation rate reaches the preset mutation rate and whether the liquid level rise rate reaches the preset rise rate. If so, the scenario type is a rainstorm impact scenario under non-standard scenarios. If not, extract the attenuation coefficient of the original acoustic signal and the peak offset of the original radar signal, and determine whether the attenuation coefficient exceeds the first preset coefficient and whether the peak offset exceeds the preset frequency. If so, the scenario type is a scaling scenario under non-standard scenarios. If not, determine whether the estimated flow rate is less than the preset flow rate and whether the estimated liquid level is less than the preset liquid level. If yes, the scenario type is a low flow rate and low liquid level scenario under non-standard scenarios. If no, the scenario type is a standard scenario.

5. The monitoring method according to claim 4, characterized in that, The adaptive filtering strategy is as follows: In a standard scenario, basic filtering rules are established, and the original acoustic and radar signals are filtered using these rules to obtain the ultrasonic and radar monitoring values ​​in the standard scenario. In non-standard scenarios, the basic filtering rules are fine-tuned, and the original acoustic and radar signals are filtered using the fine-tuned basic filtering rules to obtain the ultrasonic and radar monitoring values ​​under non-standard scenarios.

6. The monitoring method according to claim 5, characterized in that, The basic filtering rules for the raw radar signal are as follows: The original radar signal is acquired at a preset sampling frequency, and the beam scattering intensity and spectrum data are extracted. The proportion of spectral clutter is calculated based on the spectrum data. It is determined whether the beam scattering intensity is greater than the preset intensity and whether the proportion of spectral clutter is greater than the preset proportion. If so, the original radar signal is filtered using a 128th-order FIR low-pass filter. If not, the original radar signal is filtered using a 32nd-order FIR low-pass filter. The original radar signal after filtering is self-healing verified. After the self-healing verification is passed, the radar monitoring value under the standard scenario is obtained.

7. The monitoring method according to claim 6, characterized in that, Fine-tuning of the basic filtering rules for the raw radar signal includes: In low flow rate and low liquid level scenarios, the filtering parameters of the FIR low-pass filter in the standard scenario are adjusted, and the radar monitoring value obtained in the adjusted standard scenario is used as the radar monitoring value in the low flow rate and low liquid level scenario. The filtering parameters of the FIR low-pass filter include at least the order and the cutoff frequency. In the scaling scenario, the frequency offset is input into the pre-built frequency-thickness fitting equation to obtain the scaling thickness. The radar monitoring value is corrected according to the scaling thickness and the pre-built correction model. The self-healing verification is performed again on the corrected radar monitoring value. After the self-healing verification passes, the radar monitoring value in the scaling scenario is obtained. In the rainstorm impact scenario, the filtering parameters of the FIR low-pass filter and the self-healing verification parameters in the standard scenario are adjusted, and the radar monitoring value obtained in the adjusted standard scenario is used as the radar monitoring value in the rainstorm impact scenario.

8. The monitoring method according to claim 5, characterized in that, The basic filtering rule for the original acoustic signal is as follows: The raw ultrasonic signal is acquired at a preset sampling frequency, and the time-domain amplitude is extracted. The signal attenuation coefficient is determined based on the time-domain amplitude, and the interference type is determined based on the signal attenuation coefficient. The interference type is either bubble interference or flow field anomaly. When the interference type is bubble interference, a second-order Volterra filter is used to filter the original acoustic signal. When the interference type is flow field anomaly, a rejection-interpolation rule is used to process the original acoustic signal to obtain the processed original acoustic signal. The stability of the processed original acoustic signal is adjusted based on the AGC closed-loop adjustment rule to obtain the adjusted original acoustic signal. The ultrasonic monitoring value under the standard scenario is determined based on the adjusted original acoustic signal.

9. The monitoring method according to claim 8, characterized in that, The fine-tuning of the basic filtering rules for the original sound wave signal is as follows: In low flow rate and low liquid level scenarios, the filtering parameters of the Volterra filter and the adjustment parameters of the AGC closed-loop adjustment rule in the standard scenario are adjusted, and the ultrasonic monitoring value obtained in the adjusted standard scenario is used as the ultrasonic monitoring value in the low flow rate and low liquid level scenario. The filtering parameters of the Volterra filter include at least the order. In the scaling scenario, the threshold for judging the interference type in the standard scenario is reconstructed and the adjustment parameters of the AGC closed-loop regulation rule are adjusted to obtain the ultrasonic monitoring value in the scaling scenario. In the rainstorm impact scenario, the Volterra filter in the standard scenario is replaced with the Butterworth filter, and the adjustment parameters of the AGC closed-loop regulation rule are adjusted to obtain the ultrasonic monitoring values ​​under the rainstorm impact scenario.

10. The monitoring method according to claim 4, characterized in that, The dynamic fusion strategy is as follows: In the standard scenario, the ultrasonic monitoring values ​​and radar monitoring values ​​are fused according to the extended Kalman filter algorithm to obtain the flow velocity and liquid level of the drainage pipe in the standard scenario. In non-standard scenarios, the extended Kalman filter algorithm is optimized, and the optimized extended Kalman filter algorithm is used to fuse ultrasonic monitoring values ​​and radar monitoring values ​​to obtain the flow velocity and liquid level of the drainage pipe in non-standard scenarios.

Images