A water pollution monitoring method and system based on parameter automatic calibration

Abstract

The application discloses a water pollution monitoring method and system based on parameter automatic calibration, and belongs to the technical field of water quality optical detection. The method constructs a detection cycle function and a maintenance cycle function based on three monitoring parameters, generates a detection cycle and a maintenance cycle, then performs water pollution monitoring on sample water based on the detection cycle to generate sample physicochemical data, grades flushing of a sample channel based on the maintenance cycle, and then constructs three calibration parameters according to the sample physicochemical data, flow rate data, hydraulic data, external events and input signals, performs adaptive smoothing calibration on the monitoring parameters, and adjusts the detection cycle and the maintenance cycle based on the calibrated monitoring parameters. According to the application, the detection and maintenance cycles of the unmanned water station are adaptively adjusted according to real-time monitoring data, environmental data and manual intervention signals, so that the accuracy of the monitoring data can be effectively improved, and the operation and maintenance cost of the water station can be reduced.

Description

Technical Field

[0001] This invention relates to the field of optical detection technology for water quality, and in particular to a water pollution monitoring method and system based on automatic parameter calibration. Background Technology

[0002] Water quality monitoring stations can be deployed in specific watersheds to collect water quality parameters over long periods, providing continuous water source environmental parameters. Continuous operation of water quality monitoring stations requires regular maintenance. Chinese Patent Publication No. CN114167019A discloses a monitoring method and cleaning device for an automatically cleaning water quality monitor. This cleaning device extracts cleaning fluid after each sample of the monitored liquid is discharged, sets the rinsing height and rinsing force, and rinses the sensor. This method achieves automatic cleaning without supervision by preset rinsing height and rinsing force. While the cleaning parameters in this patent application are variable, its monitoring and cleaning cycles rely on manual instructions issued by a remote server. This approach is difficult to adapt to dynamic changes in the water quality environment. When water quality deteriorates rapidly, automatic monitoring to capture pollution peaks cannot be achieved, while cleaning when water quality is good can easily lead to unnecessary resource waste. Furthermore, the liquid in the sample channel is not an ideally stable fluid. In continuous water pollution monitoring, the periodic pulsation of the pump and transient bubbles cause instantaneous aggregation of suspended particles, resulting in fluctuations in the refracted signal unrelated to pollutants. Therefore, it is necessary to suppress these fluctuations to improve the accuracy of the automatic parameter calibration cycle. Summary of the Invention

[0003] To address the aforementioned problems, this invention provides a water pollution monitoring method and system based on automatic parameter calibration. Calibration parameters are constructed based on real-time monitoring data collected from unmanned water stations, external environmental data acquired from remote control terminals, and human signals. These parameters are then used to calibrate the monitoring parameters, adjusting the detection and maintenance cycles of the unmanned water station. This not only ensures the timeliness and accuracy of monitoring data during peak pollution periods but also significantly reduces the waste of reagents, pure water, and electricity, thereby lowering the operation and maintenance costs of the water station. Furthermore, a false positive suppression method based on hydraulic data is employed to avoid interference from hydraulic disturbances in water pollution monitoring, improving the accuracy and reliability of periodic calibration.

[0004] The objective of this invention can be achieved through the following technical means: A water pollution monitoring method based on automatic parameter calibration includes the following steps: Step 1: Construct the detection cycle function and maintenance cycle function based on the preset cycle generation function, initialize the first monitoring parameter, the second monitoring parameter and the third monitoring parameter, and generate the detection cycle and maintenance cycle; Step 2: Based on the detection cycle, enter the detection state, pump the sample water and pure water into the sample channel and control channel respectively, collect the hydraulic data of the sample channel, guide the incident light through the sample channel and control channel, collect the first refraction signal and the second refraction signal, and generate the sample physicochemical data; Step 3: Enter maintenance mode based on maintenance cycle, collect flow rate data of sample channel, calculate contamination index based on sample physicochemical data, calculate blockage index based on flow rate data, and calculate hydraulic disturbance index based on first refraction signal and hydraulic data; Step 4: Combine the pollution index, blockage index, and hydraulic disturbance index to generate the first calibration parameter, construct the second calibration parameter based on external events, and construct the third calibration parameter based on the input signal; Step 5: Calibrate the first monitoring parameter based on the first calibration parameter, calibrate the second monitoring parameter based on the second calibration parameter, and calibrate the third monitoring parameter based on the third calibration parameter; Step 6: Input the calibrated first monitoring parameter, second monitoring parameter and third monitoring parameter into the detection cycle function and maintenance cycle function respectively, regenerate the detection cycle and maintenance cycle, and return to step 2.

[0005] In this invention, in step 1, the period generation function is T=T0+(1-μW1)(1-λW2)W3, where T0 is the basic period, μ and λ are sensitivity coefficients, and W1, W2 and W3 are the first monitoring parameter, the second monitoring parameter and the third monitoring parameter, respectively.

[0006] In this invention, a detection periodic function T is constructed based on the periodic generation function T. a and maintain the periodic function T b T a =T 0,a (1-μ a W1)(1-λ a W2)W 3,a T b =T 0,b (1-μ b W1)(1-λ b W2)W 3,b , among which, T 0,a μ a , λ a To detect the preset constant parameter of the periodic function, T 0,b μ b , λ b To maintain the preset constant parameters of the periodic function, W 3,a W 3,b These are two independent variables in the third monitoring parameter, corresponding to the detection cycle and maintenance cycle.

[0007] In this invention, in step 2, the physicochemical data of the sample is the target component concentration. The first light intensity of the first refraction signal and the second light intensity of the second refraction signal are extracted. The absorbance of the sample channel is calculated based on the first light intensity and the second light intensity. The target component concentration is generated by combining the absorbance of different wavelengths.

[0008] In this invention, in step 3, the hydraulic change coefficient is calculated based on the hydraulic data, the energy ratio is calculated based on the first refraction signal, the short-time correlation coefficient is calculated by combining the first refraction signal and the hydraulic data, and the hydraulic disturbance index is calculated by combining the hydraulic change coefficient, the energy ratio and the short-time correlation coefficient.

[0009] In this invention, in step 4, the external events include daily rainfall and ambient temperature. A rainfall stress index is calculated based on the daily rainfall, and a temperature stress index is calculated based on the ambient temperature. A second calibration parameter is generated by combining the rainfall stress index and the temperature stress index.

[0010] In this invention, in step 4, the input signals include detection signals, maintenance signals, and sample loading signals. The number of different input signals during the period from the end of the previous maintenance cycle to the end of the current maintenance cycle is counted. A first influence factor of the detection cycle is calculated based on the number of detection signals and sample loading signals. A second influence factor of the maintenance cycle is calculated based on the number of maintenance signals and sample loading signals. The first influence factor and the second influence factor constitute a third calibration parameter.

[0011] In this invention, in step 5, a smoothing coefficient is calculated based on the deviation between the first calibration parameter and the first monitoring parameter, and the first calibration parameter and the first monitoring parameter are fused based on the smoothing coefficient to obtain the calibrated first monitoring parameter.

[0012] A system for implementing the aforementioned parameter-based automatic calibration water pollution monitoring method includes: The detection optical path has a sample channel; The reference optical path has a reference channel; The analysis unit is configured to generate physicochemical data for the sample; The acquisition unit is configured to acquire flow velocity data and hydraulic pressure data; The first mother liquor tank is configured to connect to the sample channel; The second mother liquor tank is configured to be connected to the first mother liquor tank; The pure water tank is configured to connect the control channel and the sample channel; The air tank is configured to connect to the sample channel; The local monitoring terminal is configured to generate the first calibration parameters based on the sample's physicochemical data, flow rate data, and hydraulic data. The remote control terminal is configured to construct a second calibration parameter based on an external event and a third calibration parameter based on an input signal. The timing control unit is configured to calibrate the first monitoring parameter, the second monitoring parameter, and the third monitoring parameter, and to update the detection cycle and maintenance cycle. The timing control unit connects the pure water tank to the control channel and the first mother liquor tank to the sample channel based on the detection cycle, and sequentially connects the first mother liquor tank, the pure water tank, and the air tank to the sample channel based on the maintenance cycle. In response to the detection cycle and the maintenance cycle, it connects the first mother liquor tank to the second mother liquor tank.

[0013] In this invention, there is also a light generator for generating incident light, the detection optical path is configured to guide the incident light through the sample channel and generate a first refracted signal, and the reference optical path is configured to guide the incident light through the control channel and generate a second refracted signal.

[0014] The water pollution monitoring method and system based on parameter automatic calibration of the present invention have the following advantages: The present invention constructs a detection cycle function and a maintenance cycle function based on first, second, and third monitoring parameters to generate detection and maintenance cycles; it generates a pollution index, a blockage index, and a hydraulic disturbance index based on sample physicochemical data, flow velocity data, first refraction signal, and hydraulic data; then, it obtains the first calibration parameter by weighted fusion of the pollution index and the blockage index, and reduces the weight of the pollution index by the hydraulic disturbance index to suppress false positive interference; then, it generates second and third calibration parameters based on external events and input signals, calibrates the first, second, and third monitoring parameters according to the first, second, and third calibration parameters, and updates the detection and maintenance cycles. This allows the water station cycle to adaptively adjust from real-time monitoring data, environmental data, and manual intervention, realizing a shift from fixed cycles to on-demand operation. Furthermore, the present invention reduces the waste of pure water resources by using a three-stage rinsing mode of sample water cleaning, pure water rinsing, and air rinsing during the maintenance cycle, thereby reducing the operation and maintenance costs of unmanned water stations. Attached Figure Description

[0015] Figure 1 This is a flowchart of the water pollution monitoring method based on automatic parameter calibration according to the present invention; Figure 2 This is a schematic diagram of the unmanned water station of the present invention; Figure 3 This is a schematic diagram of the multi-parameter sample channel of the present invention; Figure 4 This is a schematic diagram of the reference optical path and the detection optical path of the present invention; Figure 5 This is a schematic diagram of the workflow of the unmanned water station of the present invention; Figure 6This is a schematic diagram of the component control for the detection cycle of the present invention; Figure 7 This is a schematic diagram of component control for the maintenance cycle of the present invention; Figure 8 This is a schematic diagram of component control during the sample loading state of the present invention; Figure 9 This is a block diagram of the system for implementing the water pollution monitoring method based on automatic parameter calibration according to the present invention.

[0016] The attached figures are labeled as follows: Sample channel 101, Control channel 102, Acquisition unit 103, First mother liquor tank 111, Second mother liquor tank 112, Pure water tank 113, Air tank 114, First pump 121, Second pump 122, Third pump 123, Fourth pump 124, First throttle valve 131, Second throttle valve 132, Third throttle valve 133, First shut-off valve 141, Second shut-off valve 142, Third shut-off valve 143, Fourth shut-off valve 144, First drain valve 151, Second drain valve 152, Reagent kit 160, Light generator 201, Monochromator 202, Beam spectroscope 203, First receiver 204, Second receiver 205. Detailed Implementation

[0017] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Example 1

[0018] like Figures 1 to 8 As shown, the present invention provides a water pollution monitoring method based on automatic parameter calibration, comprising the following steps.

[0019] Step 1: Construct a detection cycle function and a maintenance cycle function based on a preset cycle generation function, initialize the first monitoring parameter, the second monitoring parameter, and the third monitoring parameter, and generate the detection cycle and the maintenance cycle. Specifically, the cycle generation function is T = T0 + (1 - μW1)(1 - λW2)W3, where T0 is the basic cycle, μ and λ are sensitivity coefficients, and W1, W2, and W3 are the first monitoring parameter, the second monitoring parameter, and the third monitoring parameter, respectively. The first monitoring parameter is used to characterize the water quality status and the degree of contamination and blockage in the sample channel; the second monitoring parameter is used to characterize the stress intensity of the external environment on the water quality; and the third monitoring parameter is used to characterize the cumulative effect of remote human intervention. Further, a detection cycle function T is constructed based on the cycle generation function T. a and maintain the periodic function T b T a =T 0,a (1-μ a W1)(1-λ a W2)W 3,a Tb =T 0,b (1-μ b W1)(1-λ b W2)W 3,b Among them, T 0,a μ a , λ a To detect the preset constant parameters of the periodic function, for example, T 0,a =1h, μ a =0.7, λ a =0.3. T 0,b μ b , λ b To maintain the preset constant parameters of the periodic function, for example, T 0,b =3 days, μ b =0.6, λ b =0.4. W 3,a W 3,b The third monitoring parameter consists of two independent variables corresponding to the detection cycle and maintenance cycle, i.e., W3 = (W 3,a W 3,b In this invention, during system initialization, the first monitoring parameter W1 is set to 0, the third monitoring parameter W3 is set to (1,1), and the second monitoring parameter W2 is set according to the actual deployment environment of the water station.

[0020] Step 2: Based on the detection cycle, enter the detection state, pump the sample water and purified water into the sample channel and control channel respectively, collect the hydraulic pressure data of the sample channel, guide the incident light through the sample channel and control channel, collect the first refraction signal and the second refraction signal, and generate the sample physicochemical data. In this invention, the full-spectrum method in ultraviolet-visible spectroscopy water quality detection technology is used to test the concentration of multiple target components in the sample water, such as COD, nitrate nitrogen, turbidity, etc. (Refer to...) Figure 4 In the detection state, the light generator 201 emits incident light, which is a mixed light containing multiple frequencies. The incident light is separated into single-wavelength light by the monochromator 202, and then generated into first transmitted light and second reflected light by the beam splitter 203. The first transmitted light passes through the sample channel 101 via the detection optical path assembly and reaches the first receiver 204, generating a first refracted signal; the second reflected light passes through the reference optical path assembly through the control channel 102 and reaches the second receiver 205, generating a second refracted signal. The light generator 201 is a continuous spectrum light source, such as a deuterium lamp or a tungsten halogen lamp. The beam splitter 203 is a polarizing beam splitter, used to split the single-wavelength light into two linearly polarized beams with the same frequency and mutually perpendicular polarization directions.

[0021] The first light intensity G1 of the first refracted signal and the second light intensity G2 of the second refracted signal are extracted. The absorbance A of the sample channel is calculated based on the first light intensity G1 and the second light intensity G2, where A = lg(G2 / G1). The first and second light intensities are the average light intensities of the first and second refracted signals. The target component concentration is generated by combining the absorbance at different wavelengths according to the Lambert-Beer law A = eLC, where e is the molar absorptivity, L is the optical path length, and C is the concentration. Further, N sets of absorbance A at different wavelengths are extracted. n n=1,2,…,N. It is assumed that there are M target components in the sample water. The concentrations of the target components are represented by the concentration vector C=[C1,C2,…,C…]. m ,…,C M ] indicates that C m That is, the concentration of the target component m. For the target component m, let its molar absorptivity for a single wavelength of light corresponding to the nth wavelength group be e. mn According to the absorbance formula A n =e 1n L0C1+e 2n L0C2+…+e mn L0C m +…+e Mn L0C M N equations can be obtained. By solving these N equations simultaneously using the least squares algorithm, the concentration C of the target component can be obtained. Here, L0 is the optical path length of the detection optical path. The target component concentration is the physicochemical data of the sample.

[0022] Step 3: Based on the maintenance cycle, enter maintenance mode, collect flow rate data of the sample channel, calculate the contamination index based on the sample physicochemical data, calculate the blockage index based on the flow rate data, and calculate the hydraulic disturbance index based on the first refraction signal and hydraulic data. The specific calculation process of the contamination index, blockage index, and hydraulic disturbance index is as described in Example 2. As the number of tests increases, suspended particles, colloids, organic matter, and microorganisms in the water sample will gradually adhere to and accumulate on the inner wall of the sample channel, resulting in a decrease in incident light transmittance and enhanced light scattering, causing deviations in the measurement results of the first refraction signal. Therefore, the sample channel needs to be cleaned regularly. In this invention, when the unmanned water station enters maintenance mode, the flow rate data of the sample channel is collected immediately, usually for 5-10 seconds. Then, the first mother liquor tank, pure water tank, and air tank are sequentially connected to the sample channel. The sample channel is cleaned, rinsed, and emptied using sample water, pure water, and air, respectively, ensuring that the channel is clean and unobstructed while avoiding waste of pure water resources. In addition, to ensure that there is sufficient sample water in the second mother liquor tank for testing and cleaning, the present invention also collects hydraulic data at the output ends of the first and second mother liquor tanks and calculates the hydraulic difference. When the hydraulic difference is lower than a threshold, the first and second mother liquor tanks are connected to pump the clean sample water in the upper layer of the second mother liquor tank into the second mother liquor tank.

[0023] Step 4: Generate a first calibration parameter by combining the pollution index, blockage index, and hydraulic disturbance index; construct a second calibration parameter based on external events; and construct a third calibration parameter based on the input signal. In this invention, to avoid the interference of false positives in water pollution detection caused by hydraulic disturbances such as bubbles and turbulence on pollution determination, a hydraulic disturbance index is introduced to reduce the weight of the pollution index when generating the first calibration parameter by weighted fusion of the pollution index and blockage index, so as to ensure the reliability of the first calibration parameter. The external events include daily rainfall and ambient temperature. A rainfall stress index is calculated based on daily rainfall, and a temperature stress index is calculated based on ambient temperature. The second calibration parameter Z2 is generated by combining the rainfall stress index and the temperature stress index. The input signal refers to the manual command signal issued by the remote control terminal, including detection signal, maintenance signal, and sample loading signal. The number of different input signals from the end of the previous maintenance cycle to the end of the current maintenance cycle is counted. A first influence factor of the detection cycle is calculated based on the number of detection signals and sample loading signals, and a second influence factor of the maintenance cycle is calculated based on the number of maintenance signals and sample loading signals. The first influence factor and the second influence factor constitute the third calibration parameter Z3. The specific steps for generating the first calibration parameter, the second calibration parameter, and the third calibration parameter are as described in Example 3.

[0024] Step 5: Calibrate the first monitoring parameter based on the first calibration parameter, calibrate the second monitoring parameter based on the second calibration parameter, and calibrate the third monitoring parameter based on the third calibration parameter. During the long-term operation of the unmanned water station, the three monitoring parameters (W1, W2, W3) need to be updated according to constantly changing operating conditions. If the corresponding monitoring parameters are directly replaced with the three calibration parameters (Z1, Z2, Z3), it can easily lead to drastic changes in the detection and maintenance cycle, or even unreasonable scheduling, such as shortening the detection cycle to a few minutes due to an accidental turbidity spike, resulting in a large waste of reagents. In this embodiment, a smoothing coefficient θ is generated based on the deviation between the first calibration parameter Z1 and the first monitoring parameter W1. The calculation formula is θ=min(0.5,max(0.2,|Z1-W1| / 0.5)), which limits θ to the range [0.2,0.5]. The larger the deviation between Z1 and W1, the larger θ, and the higher the weight given to the first calibration parameter, enabling the first monitoring parameter to quickly track changes in the actual state; when the deviation is small, smoothing is the primary function, suppressing noise. The first calibration parameter Z1 and the first monitoring parameter W1 are fused based on the smoothing coefficient θ to obtain the calibrated first monitoring parameter W1', where W1' = θZ1 + (1-θ)W1. The second and third calibrated monitoring parameters can be obtained using the same method. In a more preferred embodiment, Kalman filtering, Bayesian updating, or other methods can also be used to update the monitoring parameters based on the calibration parameters.

[0025] Step 6: Input the calibrated first, second, and third monitoring parameters into the detection cycle function and maintenance cycle function respectively, regenerate the detection cycle and maintenance cycle, and return to Step 2. In this invention, the detection cycle function and maintenance cycle function are constructed by using a cycle generation function that is weighted by multiplying the first, second, and third monitoring parameters. After each detection and maintenance of the unmanned water station, the first, second, and third calibration parameters are constructed using the latest sample physicochemical data, flow rate data, hydraulic data, external events, and input signals. The first, second, and third monitoring parameters are then smoothly calibrated, and the calibration results are re-input into the cycle function to form a closed-loop feedback mechanism, making the detection and maintenance cycle adjustment of the unmanned water station more stable and accurate. In a further embodiment, the constant parameters of the detection cycle function and maintenance cycle function can also be dynamically calibrated to more comprehensively adapt to the long-term operating characteristics and operating condition drift of the water station, thereby further improving the accuracy and robustness of the cycle adjustment. Example 2

[0026] This embodiment further discloses a preferred method for calculating the pollution index, the blockage index, and the hydraulic disturbance index.

[0027] The contamination index is calculated based on the physicochemical data of the sample. In this invention, the degree of contamination of the sample channel by the sample water is quantified by detecting the concentration of the target component. Specifically, the physicochemical data of the sample, C=[C1,C2,…,C…], are extracted. m ,…,C M ] Calculate the pollution index C s , C m,th Let m be the cleaning threshold concentration of the target component, and M be the number of target component species. The function min(x,y) represents the minimum value between x and y. s The value range of C is [0,1]. s The closer the value is to 1, the higher the pollution level of the sample water, and the greater the impact of the sample water on the optical window of the sample channel. In a more preferred embodiment, the optical path parameters can also be calibrated based on the pollution index to further improve the accuracy of water quality parameter monitoring.

[0028] The blockage index is calculated based on flow velocity data. Flow velocity data V1 and V2 are collected at the input and output ends of the sample channel. Let V1 = [v...]. i,1 ,v i,2 ,…,v i,k ,…,v i,K ], V2=[v o,1 ,v o,2 ,…, v o,k ,…,v o,K ], v i,k v o,k This represents the flow velocity at the input and output terminals at sampling point k. K is the number of sampling points, K=T. c / Δt,T c Where T is the sampling duration, and Δt is the sampling period. For example, T c =10s, Δt=0.02s. Calculate the average flow velocity v at the input and output ends respectively. in v out ,but , Next, calculate the overall average flow velocity v, then v = (v in +v out ) / 2. Calculate the congestion index D based on the overall average flow velocity v. s D s =(v ref -v) / (v ref -v min ), v ref v min These are the preset baseline flow rate and minimum flow rate, v ref v can be obtained through experimental calibration. minThe average flow rate is typically taken as 0.05 m / s. The lower the average flow rate, the greater the flow resistance in the sample channel, and the more severe the blockage in the sample channel, i.e., the higher the blockage index.

[0029] The hydraulic pressure variation coefficient is calculated based on hydraulic data. Hydraulic pressure data P is collected at the input and output ends of the sample channel. in and P out Let P in =[P i,1 ,P i,2 ,…,P i,k ,…,P i,K ], P out =[P o,1 ,P o,2 ,…, P o,k ,…,P o,K ], P i,k P o,k This represents the hydraulic pressure at sampling point k between the input and output terminals. Calculate the hydraulic pressure difference ΔP at sampling point k in the sample channel. k ΔP k = P i,k - P o,k Let k = 1, 2, ..., K. A larger hydraulic pressure difference indicates a more severe blockage in the sample channel. The hydraulic pressure change rate P is calculated based on the hydraulic pressure difference. k P k =(ΔP k -ΔP k-1 ) / Δt, k=2,3,…,K. Then calculate the hydraulic change coefficient P based on the hydraulic change rate. RMS , The hydraulic variation coefficient P RMS P reflects the intensity of water flow fluctuations within the sample channel. RMS The larger the value, the more severe the disturbances such as bubbles, turbulence, or pressure pulsations in the water flow.

[0030] The energy ratio is calculated based on the first refracted signal. Let the first refracted signal be Q = [q1, q2, ..., q]. k ,…,q K ], q k This represents the instantaneous light intensity at sampling point k. Calculate the energy ratio E of the first refracted signal. r , Among them, q filt,k This represents the filtered light intensity signal. G1 is the first light intensity, which is the average instantaneous light intensity of the K sampling points. ε is a very small positive number used to prevent the denominator from being 0; for example, it can be 10. -6 The energy ratio E r This reflects the proportion of wave energy in the first refracted signal that is in the same frequency band as the hydraulic disturbance to the total wave energy, E rThe larger the value, the more likely hydraulic disturbance is to be the main factor causing fluctuations in the first refracted signal.

[0031] The short-time correlation coefficient ρ is calculated by combining the first refraction signal and hydraulic data. Wherein, ΔP avg The mean of the hydraulic pressure difference across K sampling points is given. The short-time correlation coefficient ρ reflects the waveform similarity between the first refraction signal and the hydraulic data within the same time window; the larger ρ is, the more synchronized the waveforms are. It is important to emphasize that, to ensure the physical meaning and accuracy of the above energy ratio and short-time correlation coefficient calculations, the first refraction signal and hydraulic data must be collected synchronously within the same time window under the same time reference.

[0032] Combined with hydraulic variation coefficient P RMS Energy ratio E r The hydraulic disturbance index P is calculated using the short-time correlation coefficient ρ. s P s =ω1P RMS +ω2E r +ω3ρ, where ω1, ω2, and ω3 are weighting coefficients, and ω1+ω2+ω3=1, which can be set according to the actual situation. Since the short-time correlation coefficient is the most direct evidence of the synchronization between hydraulic disturbance and optical measurement, the weight value of ω3 is usually higher than that of ω1 and ω2, for example, ω1=ω2=0.3, ω3=0.4. The higher the hydraulic disturbance index, the greater the influence of hydraulic disturbance on the optical path, and the greater the possibility of false positives in water pollution detection caused by hydraulic disturbance. Therefore, the pollution index should be downweighted when fusing the first calibration parameter to suppress the risk of false positives. Example 3

[0033] This embodiment further discloses a preferred method for generating the first calibration parameter, the second calibration parameter, and the third calibration parameter.

[0034] Generate the first calibration parameters. Referring to Example 2, calculate the contamination index based on the sample's physicochemical data; calculate the blockage index based on the flow rate data; calculate the hydraulic pressure change coefficient based on the hydraulic pressure data; calculate the energy ratio based on the first refraction signal; calculate the short-time cross-correlation coefficient by combining the first refraction signal and the hydraulic pressure data; and then calculate the hydraulic pressure disturbance index by combining the hydraulic pressure change coefficient, energy ratio, and short-time cross-correlation coefficient. Based on the contamination index C... s Congestion Index D s and hydraulic disturbance index P s Generate the first calibration parameter Z1, Z1=a1C s (1-P) s )+a2D sa1 and a2 are weighting coefficients, satisfying a1 + a2 = 1, and can be adjusted according to actual conditions. For example, in this embodiment, a1 = 0.4 and a2 = 0.6. Where (1 - P s This is used to downweight the pollution index, i.e., when hydraulic disturbance is severe (P... s When the value approaches 1), the contribution of the pollution index to the first calibration parameter approaches zero, avoiding interference from false positives. The first calibration parameter reflects the health of the water station's operation; the higher its value, the more severe the pollution and blockage at the water station.

[0035] A second calibration parameter is constructed based on external events, including daily rainfall R and ambient temperature T. The rainfall stress index R is calculated based on the daily rainfall R. s R s =min(1,R / R max ), R max To significantly influence the threshold, such as 50 mm, the temperature stress index T is calculated based on the ambient temperature T. s T s =(TT opt ) / (T max -T opt ), T opt T max These are the activation and inhibition temperatures for algae or microorganisms, respectively. It should be noted that when T < Tactive... opt Or T > T max At that time, T s =0. Normally, T opt =20℃, T max =35℃. Furthermore, combined with the rainfall stress index R... s and temperature stress index T s Generate the second calibration parameter Z2, Z2=b1R s +b2T s b1 and b2 are weighting coefficients, satisfying b1 + b2 = 1. These values ​​can be set according to environmental factors. For example, when there is more rainfall but lower temperature, b1 and b2 can be set to 0.7 and 0.3 respectively. The second calibration parameter reflects the potential stress of external environmental factors on water quality and equipment operation. The larger the value, the more serious the threat of environmental pollution to the detection pool. The system should significantly shorten the detection and maintenance cycle to increase the monitoring frequency and intervene with cleaning in advance.

[0036] A third calibration parameter is constructed based on the input signal. In this invention, the unmanned water station is normally in an automatic control state, but when the system detects an anomaly, it will intervene manually through a remote control terminal, sending an input signal to the local water station. This manual intervention reflects, to some extent, a lag in response or insufficient threshold in the current detection or maintenance cycle. Therefore, in this invention, a third calibration parameter is generated based on the input signal sent by the remote control terminal. The input signal includes a detection signal, a maintenance signal, and a sample loading signal. The frequency (f1, f2, f3) of the detection signal, maintenance signal, and sample loading signal from the end of the previous maintenance cycle to the end of the current maintenance cycle is counted. The more frequent the occurrence of the detection and maintenance signals, the longer the current detection and maintenance cycle may be; conversely, the more frequent the occurrence of the sample loading signal, the shorter the current detection and maintenance cycle may be. The first influence factor f of the detection cycle is calculated based on the frequency (f1, f3) of the detection and sample loading signals. a f a =clamp(1-αf1 / (f 1,th +f1)+γ a f3 / (f 3,th +f3),f min ,f max Then, based on the number of maintenance signals and sample loading signals f2 and f3, calculate the second influencing factor f of the maintenance cycle. b f b =clamp(1-βf2 / (f 2,th +f2)+γ b f3 / (f 3,th +f3),f b,min ,f b,max Finally, the first impact factor f... a Second Influence Factor f b Generate the third calibration parameter Z3, i.e., Z3=(f a ,f b ).

[0037] Where α is the shortening effect of the detection signal on the detection cycle, β is the shortening effect of the maintenance signal on the maintenance cycle, and γ a γ b These represent the intensity of the extension of the detection cycle and maintenance cycle by the sample loading signal, respectively. α, β, γ a γ b The values ​​are all between 0 and 1, and in this embodiment, they are set to 0.6, 0.8, 0.4, and 0.5 respectively. 1,th f 2,th f 3,th These are the half-saturation constants for the detection signal, maintenance signal, and sample loading signal, respectively; for example, they can be set to 3, 2, and 1. min fmax These are the minimum and maximum allowable values ​​for the first and second impact factors, respectively, used to prevent the testing and maintenance cycles from being too short or too long. For example, f min =0.5, f max =2. The function `clamp()` is used to limit the numerical range of the first and second impact factors to between the minimum and maximum allowable values. Example 4

[0038] like Figures 2 to 9 As shown, the system for implementing the water pollution monitoring method based on parameter automatic calibration according to the present invention includes a detection optical path, a reference optical path, an analysis unit, an acquisition unit 103, a first mother liquor tank 111, a second mother liquor tank 112, a pure water tank 113, an air tank 114, a local monitoring terminal, a remote control terminal, and a timing control unit.

[0039] The detection optical path has a sample channel 101. The reference optical path has a control channel 102. The analysis unit is configured to generate physicochemical data of the sample. The acquisition unit 103 is configured to acquire flow rate data and hydraulic pressure data. (Reference) Figure 2 The input end of sample channel 101 is connected to a first throttle valve 131, and the output end of sample channel 101 is connected to a first shut-off valve 141. The input end of control channel 102 is connected to a second throttle valve 132, and the output end of control channel 102 is connected to a second shut-off valve 142. (Refer to...) Figure 3 Multiple parallel sample channels 101 can be deployed in the unmanned water station. Different sample channels 101 are used to detect different water quality parameters. Reagents for detecting water quality parameters can be added to the sample channels 101 using a reagent kit 160. In addition, the output terminals of the first throttle valve 131 and the first shut-off valve 141 are respectively connected to a data acquisition unit 103.

[0040] The first mother liquor tank 111 is configured to connect to the sample channel 101, and is responsible for providing sample water to the sample channel 101 for detection, analysis, and cleaning. The second mother liquor tank 112 is configured to connect to the first mother liquor tank 111, and is responsible for providing sample water to the first mother liquor tank 111 after sedimentation and filtration to remove impurities. The pure water tank 113 is configured to connect the sample channel 101 and the control channel 102, and is responsible for providing pure water to the sample channel 101 for rinsing and to the control channel 102 for reference analysis. The air tank 114 is configured to connect to the sample channel 101, and is responsible for providing dry air to the sample channel 101 for air rinsing. Figure 2As shown, the first mother liquor tank 111 is connected to a first pump 121, which pumps sample water into the sample channel 101. The pure water tank 113 is connected to a second pump 122, which pumps pure water into the sample channel 101 or the control channel 102. The air tank 114 is connected to a third pump 123, which pumps air into the sample channel 101. The second mother liquor tank 112 is connected to a fourth pump 124, which pumps clean sample water into the first mother liquor tank 111. Furthermore, the outputs of the first pump 121 and the fourth pump 124 are each connected to a data acquisition unit to collect hydraulic data from the outputs of the first mother liquor tank 111 and the second mother liquor tank 112.

[0041] The local monitoring terminal is configured to generate a first calibration parameter based on sample physicochemical data, flow rate data, and hydraulic data. Specifically, the local monitoring terminal calculates a pollution index based on sample physicochemical data, a blockage index based on flow rate data, and a hydraulic disturbance index based on a first refraction signal and hydraulic data. The first monitoring parameter is generated by combining the pollution index, blockage index, and hydraulic disturbance index. The remote control terminal is configured to construct a second calibration parameter based on external events and a third calibration parameter based on input signals. In this invention, external events, including daily rainfall and ambient temperature, are monitored by the remote control terminal. A rainfall stress index is calculated based on daily rainfall to generate the second monitoring parameter. Furthermore, when the remote control terminal identifies an abnormal situation based on water pollution monitoring data, it sends an input signal to the water station to remotely intervene in the water station's status control. The input signals include detection signals, maintenance signals, and sample loading signals, which are used to control the water station to enter detection, maintenance, and sample loading states, respectively. A first influence factor for the detection cycle is calculated based on the number of detection signals and sample loading signals, and a second influence factor for the maintenance cycle is calculated based on the number of maintenance signals and sample loading signals. The third calibration parameter is generated from the first and second influence factors.

[0042] The timing control unit is configured to calibrate the first monitoring parameter, the second monitoring parameter, and the third monitoring parameter, and update the detection cycle and maintenance cycle. In this invention, the timing control unit establishes connections with the local monitoring terminal and the remote control terminal through a wired communication unit and a wireless communication unit, respectively, to obtain the first calibration parameter, the second calibration parameter, and the third calibration parameter. It then calibrates the first monitoring parameter, the second monitoring parameter, and the third monitoring parameter, and updates the detection cycle and maintenance cycle based on the calibrated first monitoring parameter, second monitoring parameter, and third monitoring parameter. Based on the detection cycle, the timing control unit connects the pure water tank 113 to the control channel 102 and the first mother liquor tank 111 to the sample channel 101. Based on the maintenance cycle, it sequentially connects the first mother liquor tank 111, the pure water tank 113, and the air tank 114 to the sample channel 101. In response to the detection cycle and the maintenance cycle, it connects the first mother liquor tank 111 to the second mother liquor tank 112. (Refer to...) Figure 6 At each detection cycle interval, the system enters the detection state. The timing control unit controls the corresponding components to connect the pure water tank 113 to the control channel 102 and the first mother liquor tank 111 to the sample channel 101, respectively, for reference and sampling analysis. Figure 7 At each maintenance cycle, the system enters maintenance mode. The timing control unit controls the corresponding components to sequentially connect the first mother liquor tank 111, the pure water tank 113, and the air tank 114 to the sample channel 101, performing cleaning, rinsing, and empty rinsing of the sample channel 101. (Refer to...) Figure 8 In response to the aforementioned detection and maintenance cycles, the system enters the sample loading state, connecting the first mother liquor tank 111 and the second mother liquor tank 112 to replenish the sample water in the first mother liquor tank 111. The specific component control of the water station in different states is described in Example 5.

[0043] In this invention, a light generator 201 for generating incident light is also included. The detection optical path is configured to guide the incident light through the sample channel 101 and generate a first refracted signal, and the reference optical path is configured to guide the incident light through the control channel 102 and generate a second refracted signal. (Refer to...) Figure 4 In the detection state, the light generator 201 emits incident light, which is a mixed light containing multiple frequencies. The incident light is separated into single-wavelength light by the monochromator 202, and then generated into first transmitted light and second reflected light by the beam splitter 203. The first transmitted light passes through the sample channel 101 via the detection optical path assembly and reaches the first receiver 204, generating a first refracted signal; the second reflected light passes through the reference optical path assembly through the control channel 102 and reaches the second receiver 205, generating a second refracted signal. The light generator 201 is a continuous spectrum light source, such as a deuterium lamp or a tungsten halogen lamp. The beam splitter 203 is a polarizing beam splitter, used to split the single-wavelength light into two linearly polarized beams with the same frequency and mutually perpendicular polarization directions. Example 5

[0044] like Figure 5 As shown, this embodiment further discloses the workflow of the unmanned water station. In this invention, before the unmanned water station is put into use, the detection cycle and maintenance cycle are initialized. Then, water pollution monitoring and equipment maintenance are carried out regularly according to the detection cycle and maintenance cycle, and the detection cycle and maintenance cycle are updated after each maintenance. In addition, in order to ensure that there is enough sample water in the first mother liquor tank for detection and maintenance, the unmanned water station will periodically enter the sample loading state to replenish the sample water.

[0045] like Figure 2 and Figure 6As shown, when the unmanned water station is in a detection cycle, the timing control unit activates the first mother liquor tank 111, the sample channel 101, the pure water tank 113, and the control channel 102, entering the sampling and reference state. Specifically, in the sampling state, the first pump 121 starts, the first throttle valve 131 is activated, and the first pump 121 pumps the sample water from the first mother liquor tank 111 into the sample channel 101 through the first throttle valve 131. In the reference state, the second pump 122 starts, the second throttle valve 132 and the fourth shut-off valve 144 are activated, and the second pump 122 pumps the pure water from the pure water tank 113 into the control channel 102 sequentially through the fourth shut-off valve 144 and the second throttle valve 132.

[0046] like Figure 2 and Figure 7 As shown, when the unmanned water station is in a maintenance cycle, the timing control unit sequentially connects the first mother liquor tank 111, the pure water tank 113, and the air tank 114 to the sample channel 101, respectively, to enter the cleaning, rinsing, and empty rinsing states. Specifically, in the cleaning state, the first pump 121 starts, and the first throttle valve 131 and the first shut-off valve 141 are opened. The first pump 121 pumps the sample water from the first mother liquor tank 111 into the sample channel 101 through the first throttle valve 131 for the first cleaning, and then discharges it through the first shut-off valve 141. In the rinsing state, the second pump 122 starts, and the third shut-off valve 143, the first throttle valve 131, and the first shut-off valve 141 are opened. The second pump 122 pumps the pure water from the pure water tank 113 into the sample channel 101 through the third shut-off valve 143 and the first throttle valve 131 for the second cleaning, and then discharges it through the first shut-off valve 141. In the empty washing state, the third pump 123 is started, and the third shut-off valve 143, the first throttle valve 131 and the first shut-off valve 141 are opened. The third pump 123 pumps the air in the air tank 114 into the sample channel 101 through the third shut-off valve 143 and the first throttle valve 131 for air washing. The residual sewage in the sample channel is discharged through the first shut-off valve 141, and finally a clean and dry sample channel 101 is obtained.

[0047] like Figure 2 and Figure 8 As shown, when the unmanned water station is in the sample loading state, the first drain valve 151 is first opened to drain the remaining sample water in the first mother liquor tank 111, and then closed. Next, the fourth pump 124 is started and the third throttle valve 133 is opened to pump the clean sample water from the upper layer of the second mother liquor tank 112 into the first mother liquor tank 111. Further, since the remaining sample water at the bottom of the second mother liquor tank 112 contains impurities such as mud and sand, the second drain valve 152 needs to be opened to re-extract external water to rinse the second mother liquor tank 112 for a certain period. Then, the second drain valve 152 is closed, and the second mother liquor tank 112 enters the sedimentation stage, awaiting the next sample loading.

[0048] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A water pollution monitoring method based on automatic parameter calibration, characterized in that, Includes the following steps: Step 1: Construct the detection cycle function and maintenance cycle function based on the preset cycle generation function, initialize the first monitoring parameter, the second monitoring parameter and the third monitoring parameter, and generate the detection cycle and maintenance cycle; Step 2: Based on the detection cycle, enter the detection state, pump the sample water and pure water into the sample channel and control channel respectively, collect the hydraulic data of the sample channel, guide the incident light through the sample channel and control channel, collect the first refraction signal and the second refraction signal, and generate the sample physicochemical data; Step 3: Enter maintenance mode based on maintenance cycle, collect flow rate data of sample channel, calculate contamination index based on sample physicochemical data, calculate blockage index based on flow rate data, and calculate hydraulic disturbance index based on first refraction signal and hydraulic data; Step 4: Combine the pollution index, blockage index, and hydraulic disturbance index to generate the first calibration parameter, construct the second calibration parameter based on external events, and construct the third calibration parameter based on the input signal; Step 5: Calibrate the first monitoring parameter based on the first calibration parameter, calibrate the second monitoring parameter based on the second calibration parameter, and calibrate the third monitoring parameter based on the third calibration parameter; Step 6: Input the calibrated first, second, and third monitoring parameters into the detection cycle function and maintenance cycle function respectively to regenerate the detection cycle and maintenance cycle, then return to Step 2. In step 1, the period generation function is T = T0 + (1 - μW1)(1 - λW2)W3, where T0 is the basic period, μ and λ are sensitivity coefficients, and W1, W2, and W3 are the first monitoring parameter, the second monitoring parameter, and the third monitoring parameter, respectively. Based on the periodic generation function T, a detection periodic function T is constructed respectively. a and maintain the periodic function T b T a =T 0,a (1-μ a W1)(1-λ a W2)W 3,a T b =T 0,b (1-μ b W1)(1-λ b W2)W 3,b , among which, T 0,a μ a , λ a To detect the preset constant parameter of the periodic function, T 0,b μ b , λ b To maintain the preset constant parameters of the periodic function, W 3,a W 3,b These are two independent variables in the third monitoring parameter, corresponding to the detection cycle and maintenance cycle.

2. The water pollution monitoring method based on automatic parameter calibration according to claim 1, characterized in that, In step 2, the physicochemical data of the sample is the target component concentration. The first light intensity of the first refraction signal and the second light intensity of the second refraction signal are extracted. The absorbance of the sample channel is calculated based on the first light intensity and the second light intensity. The target component concentration is generated by combining the absorbance of different wavelengths.

3. The water pollution monitoring method based on automatic parameter calibration according to claim 1, characterized in that, In step 3, the hydraulic change coefficient is calculated based on the hydraulic data, the energy ratio is calculated based on the first refraction signal, the short-time correlation coefficient is calculated by combining the first refraction signal and the hydraulic data, and the hydraulic disturbance index is calculated by combining the hydraulic change coefficient, the energy ratio and the short-time correlation coefficient.

4. The water pollution monitoring method based on automatic parameter calibration according to claim 1, characterized in that, In step 4, the external events include daily rainfall and ambient temperature. A rainfall stress index is calculated based on the daily rainfall, and a temperature stress index is calculated based on the ambient temperature. The rainfall stress index and the temperature stress index are combined to generate a second calibration parameter.

5. The water pollution monitoring method based on automatic parameter calibration according to claim 1, characterized in that, In step 4, the input signals include detection signals, maintenance signals, and sample loading signals. The number of different input signals during the period from the end of the previous maintenance cycle to the end of the current maintenance cycle is counted. A first influence factor of the detection cycle is calculated based on the number of detection signals and sample loading signals. A second influence factor of the maintenance cycle is calculated based on the number of maintenance signals and sample loading signals. The first influence factor and the second influence factor constitute a third calibration parameter.

6. The water pollution monitoring method based on automatic parameter calibration according to claim 1, characterized in that, In step 5, a smoothing coefficient is calculated based on the deviation between the first calibration parameter and the first monitoring parameter. The first calibration parameter and the first monitoring parameter are then fused based on the smoothing coefficient to obtain the calibrated first monitoring parameter.

7. A system for implementing the water pollution monitoring method based on automatic parameter calibration as described in claim 1, characterized in that, include: The detection optical path has a sample channel; The reference optical path has a reference channel; The analysis unit is configured to generate physicochemical data for the sample; The acquisition unit is configured to acquire flow velocity data and hydraulic pressure data; The first mother liquor tank is configured to connect to the sample channel; The second mother liquor tank is configured to be connected to the first mother liquor tank; The pure water tank is configured to connect the control channel and the sample channel; The air tank is configured to connect to the sample channel; The local monitoring terminal is configured to generate the first calibration parameters based on the sample's physicochemical data, flow rate data, and hydraulic data. The remote control terminal is configured to construct a second calibration parameter based on an external event and a third calibration parameter based on an input signal. The timing control unit is configured to calibrate the first monitoring parameter, the second monitoring parameter, and the third monitoring parameter, and to update the detection cycle and maintenance cycle. The timing control unit connects the pure water tank to the control channel and the first mother liquor tank to the sample channel based on the detection cycle, and sequentially connects the first mother liquor tank, the pure water tank, and the air tank to the sample channel based on the maintenance cycle. In response to the detection cycle and the maintenance cycle, it connects the first mother liquor tank to the second mother liquor tank.

8. The system according to claim 7, characterized in that, It also has a light generator for generating incident light, the detection optical path is configured to guide the incident light through the sample channel and generate a first refracted signal, and the reference optical path is configured to guide the incident light through the control channel and generate a second refracted signal.

Images