High-density aquaculture water quality monitoring and sensor intelligent cleaning system and method

Abstract

The invention provides a high-density aquaculture water quality monitoring and sensor intelligent cleaning system and method, the system comprises a water quality monitoring device, a sensor intelligent cleaning device, a control module, a wireless transmission module, a power supply module and a bin body supporting device, and each device is connected with the control module through an RS485 bus to form a cooperative organic whole. The water quality monitoring and sensor intelligent cleaning method comprises the following steps: feeding culture water; detecting water quality and uploading water quality data; draining after detection is finished; detecting the pollution degree of the sensor; sensor cleaning decision and execution; uploading sensor cleaning data; carrying out comprehensive analysis and prediction based on multiple factors; and maintenance based on the monitoring period. The pollution degree of the sensor is monitored in real time through the light emitting and receiving sensor, the intelligent algorithm is combined, the cleaning strategy is dynamically regulated and controlled on the basis of pollution degree dynamic analysis, the accuracy and continuity of monitoring data are improved, the manual maintenance cost is reduced, and intelligent development of high-density breeding monitoring and maintenance is promoted.

Description

Technical Field

[0001] This invention belongs to the field of high-density aquaculture environment monitoring and sensor maintenance technology, and particularly relates to a high-density aquaculture water quality monitoring and sensor intelligent cleaning system and method. Background Technology

[0002] High-density aquaculture is an important development direction in modern aquaculture, increasing the yield per unit of water body through intensive management to meet the market's demand for large-scale aquatic products. In high-density aquaculture, farmed organisms such as fish and crustaceans have stringent and differentiated requirements for water quality throughout their growth cycles. Even slight fluctuations in water quality indicators such as temperature, pH, dissolved oxygen, ammonia nitrogen content, and turbidity directly affect the feeding appetite, growth rate, survival rate, and product quality of these organisms. During the aquaculture process, due to the high density of farmed organisms, large amounts of excrement, and rapid decomposition of uneaten feed, the concentrations of pollutants such as organic matter, ammonia nitrogen, and nitrite in the aquaculture water are much higher than in ordinary low-density aquaculture water. Water quality deteriorates extremely quickly, and if not detected and treated in time, it can lead to widespread disease and even death of farmed organisms within a short period, causing irreparable economic losses. Therefore, high-frequency, high-precision real-time monitoring of water quality in high-density aquaculture, and rapid response based on monitoring results, are core elements in ensuring the profitability of high-density aquaculture.

[0003] Meanwhile, the water quality in high-density aquaculture is worse than in other aquaculture environments. On the one hand, the water is rich in viscous substances such as biological metabolites and uneaten feed decomposition products, which easily form a dense adhesion layer on the sensor surface. On the other hand, the water in high-density aquaculture scenarios is frequently disturbed (such as the operation of aeration equipment and biological activity), and a large number of suspended particles adhere to the sensor probe surface with the water flow, which not only affects the accuracy of monitoring but also accelerates the electrochemical corrosion of the sensor. As the core component of monitoring, water quality sensors will experience decreased sensitivity and increased measurement deviation when immersed in such water for a long time. Their service life is shortened by 30% to 50% compared to ordinary aquaculture environments, seriously restricting the continuity and reliability of water quality monitoring in high-density aquaculture.

[0004] Traditional sensor maintenance methods rely heavily on manual cleaning and replacement, which not only requires significant labor costs but also makes it difficult to dynamically adjust the cleaning cycle according to the actual pollution level. If the cleaning interval is too long, the sensor may become severely contaminated, and the monitoring data will be distorted. If cleaning is too frequent, mechanical friction and the corrosive effects of chemical cleaning solutions will exacerbate sensor wear and tear, further increasing aquaculture costs. In addition, monitoring needs to be paused during manual cleaning, resulting in data gaps and preventing the formation of complete water quality change curves, which affects aquaculture managers' judgment of water quality trends.

[0005] Prior art 1 (CN106643910A) discloses a multi-point water quality monitoring and sensor cleaning and maintenance device for aquaculture. This device uses a monitoring tank system, a clean water tank system, a three-axis movement system, and an electronic control device to drive a single set of water quality monitoring sensors to switch between multiple monitoring tanks and a clean water tank, achieving multi-point monitoring and sensor cleaning (combined with ultrasonic waves and clean water circulation). The device also uses the clean water tank to calibrate the sensors and ensure accuracy. However, this device has significant drawbacks in high-density aquaculture scenarios: the cleaning intensity is not adjustable, making it unsuitable for dynamic pollution in high-density water bodies; the polling monitoring of a single set of sensors has a time lag, making it difficult to capture instantaneous changes in water quality; cleaning triggering relies on a fixed cycle and is not linked to real-time pollution status; the open structure and mechanical moving parts are prone to failure in complex aquaculture environments, making it difficult to meet the requirements for long-term unattended operation.

[0006] Furthermore, prior art 2 (CN117310114A) discloses a self-cleaning water quality monitoring system and detection method. Its core is to use a waterproof motor outside the casing to drive a rotating cleaning brush, actively cleaning the surface of the sensor to remove microorganisms. Simultaneously, it employs multi-sensor adaptive weighted fusion and a BP neural network algorithm to correct measurement errors for parameters such as dissolved oxygen. While this system enhances data accuracy and sensor self-cleaning capabilities, it still has shortcomings in high-density aquaculture scenarios: the cleaning brush is a fixed mechanical structure, and its cleaning effect drops sharply after the bristles wear down; it also cannot adjust the cleaning intensity for highly viscous dirt in high-density water bodies; the multi-sensor fusion algorithm relies on a large amount of data from similar devices, and in small- to medium-sized high-density aquaculture farms (where the number of devices deployed is limited), the data fusion effect is limited, making it difficult to fully correct errors; the system does not have optimized protective design for the high concentration of suspended particles in high-density water bodies, and long-term use can easily lead to problems such as cleaning brush jamming and sensor hole clogging.

[0007] Based on this, the present invention designs a high-density aquaculture water quality monitoring and sensor intelligent cleaning system and method, aiming to solve the problems of easy sensor contamination, low monitoring accuracy, high maintenance cost and poor adaptability in the prior art through an intelligent monitoring and cleaning collaborative mechanism, so as to provide stable, reliable and efficient water quality monitoring support for high-density aquaculture and promote the development of high-density aquaculture towards intelligence and precision. Summary of the Invention

[0008] To address the shortcomings of existing technologies, this invention provides a high-density aquaculture water quality monitoring and intelligent sensor cleaning system and method. This system enables automatic monitoring of water quality and intelligent cleaning of sensors in high-density aquaculture. By dynamically analyzing and adjusting cleaning strategies based on pollution levels, it improves the accuracy and continuity of monitoring data, reduces manual maintenance costs, and solves the problems of easy contamination of water quality sensors, low monitoring accuracy, and poor cleaning efficiency in high-density aquaculture. This promotes the intelligent development of high-density aquaculture monitoring and maintenance.

[0009] The present invention achieves the above-mentioned technical objectives through the following technical means.

[0010] A high-density aquaculture water quality monitoring and sensor intelligent cleaning system includes a water quality monitoring device, a sensor intelligent cleaning device, a control module, a wireless transmission module, and a power supply module;

[0011] The water quality monitoring device includes a water quality testing chamber and a water inlet unit. The water quality testing chamber is equipped with various types of sensors, a water level limiter on its inner wall, and a drain outlet at its bottom. The drain outlet connects to a drainage pipe, which is fitted with a PVC ball valve switch A controlled by a control module. The water inlet unit includes a water pump filter installed in the aquaculture pond. The water pump filter is connected to the inlet of the water pump via an inlet pipe, and the outlet of the water pump is connected to the water quality testing chamber via a pipe. The water pump is also connected to an inlet relay signal, which is connected to the control module via a signal control line.

[0012] The sensor-based intelligent cleaning device includes a clean water storage tank and a cleaning unit. A water level limiter is installed on the inner wall of the clean water storage tank, and a drain outlet is located at the bottom. The drain outlet is connected to a drain pipe, which is equipped with a manually controlled PVC ball valve switch B. A circumferential moving track is installed on the inner side wall of the water quality testing chamber, and a sliding mechanism is mounted on the moving track. The light emission and reception sensor of the cleaning unit is fixed to the sliding mechanism and is signal-connected to the control module. The cleaning unit also includes an adjustable pressure cleaning water pump. The inlet of the adjustable pressure cleaning water pump is connected to the clean water storage tank via a clean water pipe, and the outlet is connected to the water quality testing chamber via a pipe. Multiple cleaning spray guns are also installed on the pipe extending into the water quality testing chamber. The adjustable pressure cleaning water pump is also signal-connected to a cleaning relay, which is connected to the control module via a signal control line.

[0013] The high-density aquaculture water quality monitoring and sensor intelligent cleaning method utilizing the aforementioned high-density aquaculture water quality monitoring and sensor intelligent cleaning system includes the following processes:

[0014] Step 1: Fill the aquaculture tank with water;

[0015] Step 2: Water quality testing and data upload;

[0016] Step 3: Drain the aquaculture water after testing;

[0017] Step 4: Sensor contamination detection;

[0018] Step 5: Sensor cleaning decision and execution;

[0019] Step 6: Sensor cleaning data upload;

[0020] Step 7: Comprehensive analysis and prediction based on multiple factors.

[0021] Further, step 1 includes:

[0022] The control module triggers the inlet relay to start the inlet water pump. After the water sample in the aquaculture pond is filtered to remove suspended impurities by the water pump filter, it is transported to the water quality testing chamber through the inlet pipe. When the water level of the sample reaches the corresponding water level limiter position in the water quality testing chamber, the water level signal is fed back to the control module. The control module immediately issues a control command to stop the inlet water pump, ensuring that the water volume in the water quality testing chamber meets the sensor's immersion detection requirements, while avoiding water overflow and waste.

[0023] Further, step 2 includes:

[0024] All sensors in the water quality testing chamber start up synchronously to perform multi-parameter testing on the water sample. The real-time collected water quality data is transmitted to the control module through signal lines. The control module performs noise reduction and calibration on the raw data, generates a standardized water quality parameter report, and then uploads it to the remote monitoring platform via a wireless transmission module with encryption.

[0025] Further, step 3 includes:

[0026] After the water quality test is completed, the control module sends an opening command to the drain solenoid valve, and the water sample in the water quality test chamber is discharged through the drain outlet of the test chamber; when the water level drops to the corresponding water level limit position, the control module receives the signal and closes the drain solenoid valve to ensure that there is no residual water sample in the water quality test chamber.

[0027] Further, step 4 includes:

[0028] Step 4.1: Perform baseline value calibration for the sensors under clean conditions, and obtain the baseline value of reflected light intensity for each sensor under clean conditions;

[0029] Step 4.2: Perform contamination detection on the first sensor;

[0030] The sliding mechanism equipped with a light emission and reception sensor slides along the moving track under the command and control of the control module. The light emission and reception sensor first aligns with the side wall surface of the first sensor and emits infrared light of a specific wavelength in a directional manner. After the light is reflected by the sensor surface, it is captured by the receiver of the light emission and reception sensor. The control module calculates the current comprehensive contamination level and cleaning requirement of the sensor based on the attenuation of the reflected light intensity.

[0031] Among them, comprehensive pollution level The calculation is as follows:

[0032]

[0033] in, Indicates the current pollution level. , The reference value for the intensity of reflected light when the sensor is clean. This represents the current intensity of the reflected light. This represents the current pollution level weighting coefficient. =0.6~0.8; This is the average pollution level weighting coefficient. =0.2~0.4; The weighting coefficient for the rate of change in pollution level. =0.1~0.2; For the past Sub-average pollution level , This is the average number of pollution tests used in the past, adjusted according to the monitoring cycle. Indicates the first The current level of contamination detected in this test; The rate of change in pollution level. , Indicates the last calculation Subtract the calculation from the previous one The difference; Indicates the last calculation Compared to the calculation before last The time interval between;

[0034] Cleaning demand The calculation is as follows:

[0035]

[0036] in, Correction coefficients for each sensor type; This represents the cleaning interval time coefficient. ; This is the time since the last cleaning. Indicates the turbidity coefficient of water quality. , This is the real-time turbidity value;

[0037] Step 4.3: The control module continues to control the sliding mechanism to move and calculate the overall contamination level and cleaning requirement of the next sensor, and so on until the overall contamination level and cleaning requirement of all sensors are calculated.

[0038] Further, step 5 includes:

[0039] Step 5.1: The control module compares the cleaning requirements of each sensor calculated in Step 4 with the preset cleaning threshold;

[0040] Step 5.2: If the threshold is exceeded, the cleaning unit calculates cleaning parameters in real time, including cleaning pressure. Single sensor cleaning time If the threshold is not exceeded, proceed directly to step 6.

[0041] Cleaning pressure Calculated using the following piecewise function:

[0042]

[0043] Single sensor cleaning time The calculation formula is:

[0044]

[0045] Step 5.3: The control module starts the adjustable pressure cleaning water pump via the cleaning relay;

[0046] Step 5.4: The control module controls the corresponding cleaning spray gun to perform directional rinsing of the sensors that need to be cleaned within a preset time period according to the calculated pressure of the water flow;

[0047] Step 5.5: Calculate the effectiveness coefficient after cleaning is completed. :

[0048]

[0049] in, The overall level of contamination before cleaning. The overall level of contamination after cleaning;

[0050] Step 5.6: Cleaning Feedback:

[0051] when When the percentage is less than 60%, it indicates that the cleaning intensity needs to be increased. Increase the cleaning pressure by 10% or extend the cleaning time for a single sensor by 15%, then perform a second cleaning. After cleaning, recalculate the effectiveness coefficient to determine if a third cleaning is necessary. Repeat this process until the percentage is 60% or less. ≤90%;

[0052] when When the efficiency is >90%, it indicates that the cleaning intensity needs to be reduced. Reduce the cleaning pressure by 5% or shorten the cleaning time for a single sensor by 10%, then perform a second cleaning. After cleaning, recalculate the effectiveness coefficient to determine if a third cleaning is necessary. Repeat this process until 60% ≤ ≤90%;

[0053] When 60%≤ If the water content is ≤90%, no further cleaning is required; proceed directly to step 6.

[0054] Further, step 7 includes:

[0055] The remote monitoring platform uses its built-in data analysis model to perform multi-dimensional correlation analysis on water quality data and cleaning data uploaded by the wireless transmission module; assess the pollution trend of the sensors; determine the effectiveness of cleaning operations; and use machine learning algorithms based on historical data to predict the sensor pollution risk level in the next 24 hours; providing decision support for aquaculture managers to adjust cleaning frequency or water quality control measures in advance.

[0056] Further, step 8 includes:

[0057] To address the differences in sensor monitoring cycles under different aquaculture scenarios, differentiated maintenance strategies are designed, and targeted maintenance measures, including emptying and waiting, timed water spraying for humidification, or soaking in clean water, are adopted based on the differences in sensor monitoring cycles.

[0058] When monitoring cycle When the frequency is ≤1 hour / time, after completing step 6, keep the water quality detection chamber empty to avoid residual moisture causing secondary contamination of the sensor. When waiting for the next detection, directly start the water inlet process in step 1 to reduce unnecessary energy consumption. In this mode, the sensor is exposed to air for a short time and does not require additional humidification.

[0059] When monitoring period 1 < When the time is ≤6 hours / time, after completing step 6, the control module instructs the cleaning unit's spray gun to spray a small amount of water mist onto the sensor probe for 10 seconds every 30 minutes to moisturize the sensor probe. The water mist covers the probe surface to form a water film, preventing the probe from drying out and scaling due to long-term exposure, and ensuring the sensor's response speed and detection accuracy in the next test.

[0060] When monitoring cycle When the test is conducted every 6 hours or more, after completing step 6, the control module starts the cleaning unit to inject clean water into the test chamber until the water level reaches the corresponding water level limit position. The sensor probe is kept moist and active by soaking in clean water. Five minutes before the next test, the drain solenoid valve is controlled to drain the clean water to avoid diluting the new water sample with residual water and to ensure the accuracy of the test data.

[0061] The present invention has the following beneficial effects:

[0062] (1) In view of the poor water quality of high-density aquaculture, this invention uses a light emission and reception sensor to monitor the contamination level of the sensor in real time and combines it with an intelligent algorithm to dynamically adjust the cleaning strategy, which solves the problems of inaccurate timing and uneven effect of traditional manual cleaning, and extends the service life of the sensor by more than 30%.

[0063] (2) The system integrates water quality monitoring and intelligent cleaning functions to achieve closed-loop management of "monitoring-cleaning-maintenance", avoids data distortion caused by sensor contamination, and improves the accuracy of water quality parameter detection to over 95%.

[0064] (3) It adopts solar power supply module and wireless transmission technology, which is suitable for unattended scenarios in large-scale high-density farms, reducing labor costs and energy consumption. The energy consumption of a single monitoring and cleaning process is reduced by 40% compared with traditional equipment.

[0065] (4) By implementing a phased operation process and a maintenance strategy based on the monitoring cycle, it takes into account both monitoring continuity and sensor protection, meets the high-frequency or long-cycle monitoring needs under different breeding densities, and is highly adaptable.

[0066] (5) The system can link water quality data with cleanliness data for analysis, providing data support for water quality control in high-density aquaculture, helping aquaculture managers to implement precise policies and reduce the incidence of diseases in high-density aquaculture. Attached Figure Description

[0067] Figure 1 This is a schematic diagram of the water quality monitoring and sensor intelligent cleaning system framework described in this invention;

[0068] Figure 2 This is a schematic diagram of the overall structure of the water quality monitoring and sensor intelligent cleaning system described in this invention;

[0069] Figure 3 This is a flowchart of the water quality monitoring and intelligent sensor cleaning method described in this invention;

[0070] In the diagram: 1-Water quality monitoring device; 101-Water quality testing chamber; 102-Water quality sensor; 1021-Water temperature sensor; 1022-pH sensor; 1023-Ammonia nitrogen sensor; 1024-Turbidity sensor; 1025-Dissolved oxygen sensor; 103-Sensor fixing device; 1031-Suspension bracket; 1032-Fixing plate frame; 104-Water inlet unit; 1041-Water inlet pump; 1042-Water inlet relay; 1043-Water pump filter; 1044-Water inlet pipe; 105-Testing chamber water level limiter; 1051-Water level limiter A; 1052-Water level limiter B; 1053-Water level limiter C; 106-Testing chamber drainage unit; 1061-Testing chamber drain outlet; 1062-PVC ball valve switch A; 1063-Drainage solenoid valve; 2-Sensor-based intelligent cleaning device; 201-Clean water storage tank; 202-Cleaning unit; 2021-Adjustable pressure cleaning water pump; 2022-Cleaning relay; 2023-Cleaning spray gun; 2024-Clean water pipe; 2025-Light emission and reception sensor; 203-Storage tank water level limiter; 2031-Water level limiter D; 2032-Water level limiter E; 204-Storage tank drainage unit; 2041-Storage tank drain outlet; 2042-PVC ball valve switch B; 3-Control module; 4-Wireless transmission module; 5-Power supply module; 501-Solar panel; 502-Battery; 503-Support plate; 504-Support rod; 6-Tank support device; 601-Support frame; 602-Lockable casters. Detailed Implementation

[0071] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but the scope of protection of the present invention is not limited thereto.

[0072] The high-density aquaculture water quality monitoring and sensor intelligent cleaning system described in this invention integrates water quality monitoring, pollution identification, intelligent cleaning, and data transmission functions to achieve fully automated monitoring of water quality in high-density aquaculture and adaptive sensor maintenance. It is particularly suitable for high-density intensive aquaculture farms. Compared with traditional manual monitoring or general-purpose water quality equipment, this system significantly improves in terms of pollution resistance, cleaning efficiency, and data accuracy, effectively reducing aquaculture risks and maintenance costs.

[0073] like Figure 1 , 2As shown, the high-density aquaculture water quality monitoring and sensor intelligent cleaning system of the present invention includes a water quality monitoring device 1, a sensor intelligent cleaning device 2, a control module 3, a wireless transmission module 4, a power supply module 5, and a housing support device 6. Each device is connected to the control module 3 via an RS485 bus, forming a collaborative organic whole. The control module 3, as the core unit, is responsible for coordinating the operation of each device, executing intelligent cleaning algorithms, and processing data. The water quality monitoring device 1 is used to collect key water quality parameters such as water temperature, pH, ammonia nitrogen, turbidity, and dissolved oxygen in the aquaculture pond. The sensor intelligent cleaning device 2 performs targeted cleaning of pollutants on the surfaces of each sensor in the water quality monitoring device 1 to ensure monitoring accuracy. The wireless transmission module 4 encrypts and uploads water quality data and cleaning data to a remote platform, supporting real-time viewing and decision-making by aquaculture managers. The power supply module 5 uses a combination of solar panels 501 and batteries 502 to ensure continuous operation of the system in outdoor environments. The housing support device 6 is used to fix the components and facilitates the system's movement to different aquaculture ponds.

[0074] Control module 3, as the core control unit of the system, adopts an STM32H743 microcontroller with a main frequency of 400MHz and a Flash capacity of 2MB. It is connected to water quality monitoring device 1, sensor intelligent cleaning device 2, and wireless transmission module 4 via RS485 bus to realize the collaborative work of each device and the real-time calculation of intelligent cleaning algorithms. Wireless transmission module 4 uses a 4G DTU and supports MQTT protocol encrypted upload to transmit water quality monitoring data and sensor cleaning data to the remote monitoring platform in real time, ensuring that aquaculture managers can view the data at any time. Power supply module 5 uses a combination of a 200W monocrystalline silicon solar panel 501 and a 24V / 100Ah gel battery 502 to ensure that the system can operate stably for more than 72 hours under continuous cloudy and rainy weather. The silo support device 6 consists of a galvanized steel pipe support frame 601 (height 0.2-0.5m) and lockable universal wheels 602 with brakes (diameter 7-10cm), which not only meets the system's fixed requirements but also facilitates movement to different aquaculture ponds for monitoring.

[0075] like Figure 1 , 2 As shown, the water quality monitoring device 1 includes a water quality testing chamber 101 (40L volume, made of 304 stainless steel), a water quality sensor 102, a sensor fixing device 103, a water inlet unit 104, a testing chamber water level limiter 105, and a testing chamber drainage unit 106.

[0076] like Figure 1 , 2As shown, a support frame 601 is fixed at the bottom of the water quality testing chamber 101, and a lockable caster wheel 602 is installed at the bottom of the support frame 601. The water inlet unit 104 includes a water inlet pump 1041 (flow rate 15L / min), a water inlet relay 1042, a water pump filter 1043 (100-mesh stainless steel filter screen), and a water inlet pipe 1044 (20mm diameter PVC pipe). The water pump filter 1043 is located in the high-density aquaculture pond and can filter out large particulate impurities such as uneaten feed and biological excrement in the aquaculture water. The water pump filter 1043 is connected to the inlet of the water inlet pump 1041 through the water inlet pipe 1044, and the outlet of the water inlet pump 1041 is connected to the water quality detection chamber 101 through a pipe. The water inlet pump 1041 is also connected to the water inlet relay 1042 for signal connection. The water inlet relay 1042 is connected to the I / O interface of the control module 3 through a signal control line and receives high and low level control from it to inject aquaculture water into the water quality detection chamber 101 for detection by the water quality sensor 102.

[0077] like Figure 1 , 2 As shown, the water quality sensor 102 is connected to the control module 3 via an RS485 bus to transmit monitoring data. The water quality sensor 102 includes a water temperature sensor 1021 (measurement range 0~50℃, accuracy ±0.1℃), a pH sensor 1022 (measurement range 0~14, accuracy ±0.01), an ammonia nitrogen sensor 1023 (measurement range 0~10mg / L, accuracy ±0.01mg / L), a turbidity sensor 1024 (measurement range 0~1000NTU, accuracy ±2%FS), and a dissolved oxygen sensor 1025 (measurement range 0~20mg / L, accuracy ±0.1mg / L). The water quality sensor 102 is vertically installed in the middle of the water quality testing chamber 101 via the sensor fixing device 103. Specifically, the sensor fixing device 103 includes a suspension bracket 1031 and a fixing plate 1032. The top of the suspension bracket 1031 is fixed to the top wall of the water quality testing chamber 101, and the fixing plate 1032 is installed at the bottom of the suspension bracket 1031. The fixing plate 1032 is provided with multiple mounting holes for installing each sensor. The sensors are installed at the same height, and the probes are completely immersed in the water sample.

[0078] like Figure 1 , 2As shown, the water level limiter 105 in the detection chamber includes water level limiter A1051 (for detecting high water level, 20cm from the top of the chamber), water level limiter B1052 (for detecting low water level, 1cm from the bottom of the chamber), and water level limiter C1053 (for detecting moisturizing water level, covering only 5cm of the sensor probe). It uses infrared liquid level detection to achieve accurate water level detection. The detection data is transmitted to the control module 3 via an RS485 bus. After analysis, the control module 3 sends control commands to the inlet water pump 1041 to achieve precise water level control. The drainage unit 106 in the detection chamber includes a drainage outlet 1061 (25mm in diameter), a PVC ball valve switch A1062 (for safety control), and a drainage solenoid valve 1063. The drainage outlet 1061 is located at the bottom of the water quality detection chamber 101 and is connected to a drainage pipe. The PVC ball valve switch A1062 is installed on this drainage pipe, and its operation is controlled by the control module 3.

[0079] like Figure 1 , 2 As shown, the sensor-based intelligent cleaning device 2 includes a clean water storage tank 201 (capacity 100L, made of food-grade PE plastic), a cleaning unit 202, a storage tank water level limiter 203, and a storage tank drainage unit 204. A support frame 601 is fixed to the bottom of the clean water storage tank 201, and lockable casters 602 are installed at the bottom of the support frame 601. The cleaning unit 202 is the core execution component, including an adjustable pressure cleaning water pump 2021 (pressure adjustment range 0.1~1MPa), a cleaning relay 2022, a cleaning spray gun 2023 (rotation angle 0~360°, spray range diameter 5~15cm), a clean water pipe 2024 (16mm diameter high-pressure hose), and a light emission and reception sensor 2025 (infrared wavelength 850nm, detection distance 10~20cm). A moving track is arranged circumferentially on the inner side wall of the water quality detection chamber 101. A sliding mechanism controlled by the control module 3 is installed on the moving track. The light emission and reception sensor 2025 is fixed on the sliding mechanism and connected to the control module 3 through an RS485 bus. It can emit infrared light of a specific wavelength to the sensor. The light is reflected by the sensor surface and then captured. During this process, the control module 3 calculates the current contamination level of the sensor based on the attenuation of the reflected light intensity using an intelligent cleaning algorithm. The sliding mechanism uses existing technology, which only needs to enable sliding along the circular track, and is not the focus of this invention. Therefore, its specific structure and principle will not be described in detail.

[0080] like Figure 1 , 2As shown, the inlet of the adjustable pressure cleaning water pump 2021 is connected to the clean water storage tank 201 via a clean water pipe 2024, and the outlet of the adjustable pressure cleaning water pump 2021 is connected to the water quality testing tank 101 via a pipe. Multiple cleaning spray guns 2023 (the number matching the number of sensors, each controlled individually by the control module 3 for directional rinsing) are also installed on the pipe extending into the water quality testing tank 101. The adjustable pressure cleaning water pump 2021 is also signal-connected to the cleaning relay 2022, which is connected to the control module 3 via a signal control line, receiving high and low level control from the relay. This drives the adjustable pressure cleaning water pump 2021 to draw clean water from the clean water storage tank 201, which is then used to clean the water quality sensor 102 via the cleaning spray guns 2023. The water level limiter 203 for the storage tank includes a water level limiter D2031 (for low water level detection, 10cm from the bottom of the tank) and a water level limiter E2032 (for high water level detection, 5cm from the top of the tank). It uses infrared liquid level detection to achieve accurate water level detection. The detection data is transmitted to the control module 3 via an RS485 bus for analysis. When the water level is lower than the water level limiter D2031, the remote platform sends a water replenishment reminder. The storage tank drainage unit 204 includes a storage tank drain outlet 2041 and a PVC ball valve switch B2042. The storage tank drain outlet 2041 is located at the bottom of the clean water storage tank 201 and is connected to a drainage pipe. The PVC ball valve switch B2042 is installed on this drainage pipe. The PVC ball valve switch B2042 is manually controlled for easy periodic emptying of long-stagnant clean water.

[0081] like Figure 1 , 2 As shown, the power supply module 5 includes a solar panel 501, a battery 502, a support plate 503, and a support rod 504. The power supply module 5 converts solar energy through the solar panel 501 and stores it in the battery 502 to provide continuous power to the system. The support plate 503 is installed on the top of the water quality testing chamber 101. The waterproof box with the battery 502, the control module 3, and the wireless transmission module 4 are all installed on the support plate 503. The four corners of the support plate 503 are equipped with height-adjustable support rods 504, and the solar panel 501 is fixed by the support rods 504.

[0082] The control module 3 is equipped with an intelligent cleaning algorithm that can calculate the overall pollution level based on data from the 2025 light emission and reception sensor. and cleaning requirements Dynamically adjust cleaning pressure Single sensor cleaning time And through the cleaning effect coefficient By correcting subsequent cleaning parameters and achieving adaptive cleaning of the sensor, this invention achieves accurate monitoring of water quality in high-density aquaculture and adaptive cleaning of the sensor through phased automated operation and intelligent algorithm collaboration.

[0083] The water quality monitoring and intelligent sensor cleaning method of the high-density aquaculture water quality monitoring and sensor intelligent cleaning system described in this invention is as follows: Figure 3 As shown, the specific process includes the following:

[0084] Step 1: Water intake stage for aquaculture;

[0085] Control module 3 triggers inlet relay 1042 to start inlet pump 1041. Water samples from the aquaculture pond, after being filtered by pump filter 1043 to remove suspended impurities, are transported to water quality testing chamber 101 via inlet pipe 1044. When the water level reaches water level limiter A1051 in water quality testing chamber 101, a water level signal is fed back to control module 3. Control module 3 immediately issues a control command to stop inlet pump 1041, ensuring that the water volume in water quality testing chamber 101 meets the sensor's immersion detection requirements while preventing water overflow and waste. This stage, through pre-filtration and precise water level control, lays the foundation for the accuracy of subsequent water quality testing.

[0086] Step 2: Water quality testing and data upload stage;

[0087] The water temperature sensor 1021, pH sensor 1022, ammonia nitrogen sensor 1023, turbidity sensor 1024, and dissolved oxygen sensor 1025 in the water quality testing chamber 101 are activated simultaneously to perform multi-parameter detection on the water sample. The real-time collected water quality data is transmitted to the control module 3 via signal lines. After noise reduction and calibration of the raw data, the control module 3 generates a standardized water quality parameter report, which is then encrypted and uploaded to the remote monitoring platform via the wireless transmission module 4. This enables aquaculture managers to monitor the water quality status of the aquaculture ponds in real time, providing data support for water quality control.

[0088] Step 3: Drainage stage after water testing is completed;

[0089] After water quality testing is completed, control module 3 sends an opening command to drain solenoid valve 1063, and the water sample in water quality testing chamber 101 is discharged through the testing chamber drain outlet 1061. When the water level drops to the monitoring position of water level limiter B1052, control module 3 receives the signal and closes drain solenoid valve 1063 to ensure that there is no residual water sample in water quality testing chamber 101, avoid residual water from interfering with subsequent sensor contamination detection, and provide an interference-free working environment for sensor cleaning operations.

[0090] Step 4: Sensor contamination detection stage;

[0091] Step 4.1: First, calibrate the reference value of the sensor in a clean state. The specific operation is as follows:

[0092] First, the water quality testing chamber 101 is filled with purified water (simulating a clean water environment), ensuring that each sensor probe is completely submerged and free of air bubbles. Then, the purified water is drained. Next, the "benchmark calibration" program is initiated through the human-machine interface of the control module 3. The control module 3 drives the sliding mechanism to move along the moving track, causing the light emission and reception sensor 2025 mounted on it to sequentially emit infrared light (wavelength 850nm) 10 times to each sensor probe, with a 1-second interval between each emission, while simultaneously recording the reflected light intensity value. The program automatically discards one maximum and one minimum value, calculating the average of the remaining eight values ​​as the benchmark value for the reflected light intensity of the sensor in its clean state. After calibration, it will automatically... The value is stored in the memory of control module 3 and serves as the benchmark for subsequent pollution degree calculations. If the sensor is replaced or its status changes after cleaning, it needs to be recalibrated.

[0093] Step 4.2: Perform contamination detection on the first sensor (in this embodiment, water temperature sensor 1021 is used as the first sensor as an example);

[0094] The sliding mechanism equipped with the light emission and reception sensor 2025 slides along the moving track under the command control of the control module 3. The light emission and reception sensor 2025 first aligns with the side wall surface of the water temperature sensor 1021 and emits infrared light of a specific wavelength in a directional manner. After being reflected by the surface of the water temperature sensor 1021, the light is captured by the receiver of the light emission and reception sensor 2025. The control module 3 calculates the current comprehensive degree of contamination and cleaning requirement of the water temperature sensor 1021 based on the degree of attenuation of the reflected light intensity (compared with the reference value in the clean state of the sensor) through an intelligent cleaning algorithm.

[0095] Among them, comprehensive pollution level The calculation is as follows:

[0096]

[0097] in, Indicates the current pollution level. , The reference value for the intensity of reflected light when the sensor is clean. This represents the current intensity of the reflected light. This represents the current pollution level weighting coefficient. =0.6~0.8; This is the average pollution level weighting coefficient. =0.2~0.4; The weighting coefficient for the rate of change in pollution level. =0.1~0.2; For the past Sub-average pollution level , The default value in this embodiment is the number of pollution detections that have been used in the past to calculate the average pollution level. = 5 (can be adjusted according to the monitoring cycle, such as high-frequency monitoring settings) =3. Low-frequency monitoring equipment = 8), Indicates the first The current level of contamination detected in this test; The rate of change in pollution level. , Indicates the last calculation Subtract the calculation from the previous one The difference; Indicates the last calculation Compared to the calculation before last The time interval between them is in hours.

[0098] Cleaning demand The calculation is as follows:

[0099]

[0100] in, These are correction coefficients for each sensor type, with the water temperature sensor having a correction coefficient of 1021. =1. pH sensor 1022 correction factor =1.3, Correction coefficient for ammonia nitrogen sensor 1023 =1.4, Turbidity sensor 1024 correction factor =1.1, Dissolved oxygen sensor 1025 correction factor =1.2; This represents the cleaning interval time coefficient. t represents the time since the last cleaning, in hours. Indicates the turbidity coefficient of water quality. , The turbidity value is obtained in real time through a turbidity sensor 1024, and the unit is NTU.

[0101] Step 4.3: Control module 3 continues to control the sliding mechanism to move and calculate the overall contamination level and cleaning requirement of the next sensor, and so on until the overall contamination level and cleaning requirement of all sensors are calculated. This process does not require manual intervention and can accurately quantify the degree of contaminants attached to the surface of each sensor, providing an objective basis for cleaning decisions.

[0102] Step 5: Sensor cleaning decision and execution phase;

[0103] Step 5.1: Control module 3 compares the calculated cleaning requirements of each sensor with the preset cleaning threshold;

[0104] Step 5.2: If the threshold is exceeded, the cleaning unit 202 calculates the cleaning parameters in real time, including the cleaning pressure. (Water pressure parameters matching the level of contamination), single sensor cleaning time per cycle (Working time is dynamically adjusted based on the level of pollution); if the threshold is not exceeded, proceed directly to step 6;

[0105] The cleaning pressure Calculated using the following piecewise function:

[0106]

[0107] The unit is MPa.

[0108] The duration of a single sensor cleaning cycle The calculation formula is:

[0109]

[0110] The unit is seconds.

[0111] Step 5.3: Control module 3 starts adjustable pressure cleaning water pump 2021 via cleaning relay 2022;

[0112] Step 5.4: Control module 3 controls the corresponding cleaning spray gun 2023 to perform directional rinsing of the sensors to be cleaned within a preset time period according to the calculated pressure of the water flow;

[0113] Step 5.5: Calculate the effectiveness coefficient after cleaning is completed. :

[0114]

[0115] in, The overall level of contamination before cleaning. The overall level of contamination after cleaning.

[0116] Step 5.6: Cleaning Feedback:

[0117] when When the percentage is less than 60%, it indicates that the cleaning intensity needs to be increased. Increase the cleaning pressure by 10% or extend the cleaning time for a single sensor by 15%, then perform a second cleaning. After cleaning, recalculate the effectiveness coefficient to determine if a third cleaning is necessary. Repeat this process until the percentage is 60% or less. ≤90%;

[0118] when When the efficiency is >90%, it indicates that the cleaning intensity needs to be reduced. Reduce the cleaning pressure by 5% or shorten the cleaning time for a single sensor by 10%, then perform a second cleaning. After cleaning, recalculate the effectiveness coefficient to determine if a third cleaning is necessary. Repeat this process until 60% ≤ ≤90%;

[0119] When 60%≤ If the water content is ≤90%, no further cleaning is required; proceed directly to step 6.

[0120] Step 6: Sensor cleaning data upload stage;

[0121] After the cleaning operation is completed, control module 3 automatically records key data of this cleaning, including the degree of contamination before cleaning, the actual cleaning water pressure during the cleaning process, and the cleaning time of a single sensor per cleaning session. The data is transmitted via wireless transmission module 4 and then linked with water quality data to a remote platform. This data can not only be used to trace the pollution change patterns of the sensors, but also provide historical data for the optimization of subsequent cleaning parameters, enabling continuous iteration of cleaning strategies.

[0122] Step 7: Comprehensive Analysis and Prediction Based on Multiple Factors; This step provides decision support for aquaculture managers to adjust cleaning frequency or water quality control measures in advance, promoting a shift from "passive cleaning" to "proactive prevention." Step 7 specifically includes:

[0123] Step 7.1: The remote monitoring platform calls the built-in data analysis model to perform multi-dimensional correlation analysis on the uploaded water quality data (such as ammonia nitrogen concentration and turbidity changes) and cleaning data (such as pollution growth rate and cleaning effect);

[0124] Step 7.2: Assess the contamination trend of the sensor (whether it is aggravated in tandem with water quality deterioration).

[0125] Step 7.3: Determine the effectiveness of the cleaning operation (whether the decrease in contamination level after a cleaning meets the standard);

[0126] Step 7.4: Using machine learning algorithms based on historical data, predict the sensor contamination risk level for the next 24 hours;

[0127] Step 7.5: Provide decision support for aquaculture managers to adjust cleaning frequency or water quality control measures in advance.

[0128] Step 8: Maintenance phase based on monitoring cycle;

[0129] To address the differences in sensor monitoring cycles under different aquaculture scenarios, differentiated maintenance strategies are designed, and targeted maintenance measures such as emptying and waiting, periodic water spraying for humidification, or soaking in clean water are adopted according to the differences in sensor monitoring cycles.

[0130] When monitoring cycle When the frequency is ≤1 hour / time (high frequency monitoring), after completing step 6, keep the water quality detection chamber 101 empty to avoid residual moisture causing secondary contamination of the sensor. When waiting for the next detection, directly start the water inlet process of step 1 to reduce unnecessary energy consumption. In this mode, the sensor is exposed to the air for a short time (usually <1 hour) and no additional humidification is required.

[0131] When monitoring period 1 < When the frequency is ≤6 hours / time (medium frequency monitoring), after completing step 6, the control module 3 instructs the spray gun of the cleaning unit 202 to spray a small amount of water mist (pressure 0.1MPa) onto the sensor probe for 10 seconds every 30 minutes to moisturize the sensor probe. The water mist covers the probe surface to form a water film, preventing the probe from drying out and scaling due to long-term exposure (1~6 hours), and ensuring the sensor's response speed and detection accuracy in the next test.

[0132] When monitoring cycle When the frequency is >6 hours / time (low-frequency monitoring), after completing step 6, control module 3 activates cleaning unit 202 to inject clean water into the detection chamber until the water level reaches water level limiter C1053. The probe is kept moist and active by soaking in clean water. Five minutes before the next detection, drain solenoid valve 1063 is controlled to drain the clean water, preventing residual water from diluting the new water sample and ensuring the accuracy of the detection data. This mode is suitable for the overwintering period (…). =12 hours / time), according to the test, the detection error of the sensor in the water immersion state after 12 hours (compared with the standard value) is ≤3%, while the error of the unimmersed sensor can reach 10% to 15%.

[0133] Before actual use, the high-density aquaculture water quality monitoring and sensor intelligent cleaning system described in this invention needs to be properly installed and arranged. The on-site installation process is as follows:

[0134] The on-site installation of the system needs to be combined with the layout characteristics of the high-density aquaculture ponds, and representative monitoring points should be selected (such as the middle of the aquaculture pond or a place with gentle water flow) to ensure that the monitoring data can reflect the overall water quality. The installation steps are as follows: First, place the silo support device 6 on a flat ground at the edge of the aquaculture pond and lock the lockable casters 602 to prevent the equipment from shaking; if the ground is uneven, the height of the support frame 601 can be adjusted to keep the equipment level. Then, connect the water inlet pipe 1044 to the aquaculture pond: the water inlet end of the water inlet pipe 1044 should be installed 30cm below the water surface of the aquaculture pond, connected to the water pump filter 1043, and fixed with a fixing frame to prevent the pipe from shifting due to water flow impact; the end of the water inlet pipe 1044 is connected to the water inlet of the water pump 1041, and the interface is sealed with Teflon tape to prevent water leakage. The replenishment of cleaning water needs to be prepared in advance: First, close the PVC ball valve switch B2042, open the cover of the clean water storage tank 201, and manually pour tap water or filtered well water into the tank until it reaches approximately the water level limiter E2032 (about 90L). Then close the cover of the clean water storage tank 201 to ensure that the cleaning water is clean and free of impurities. The installation of the power supply module 5 must ensure that the solar panel 501 receives sufficient sunlight: Fix the support plate 503 above the water quality testing tank 101, and adjust the height of the support rod 504 to form an angle of 30° (for areas around 30° North latitude, slight adjustments can be made according to the local latitude) so that the solar panel 501 can receive sunlight vertically at noon; place the battery 502 in a waterproof box, connect the charge / discharge controller to the solar panel 501 and the main unit of the equipment, ensuring that the positive and negative terminals are correctly connected to avoid short circuits. The antenna of the wireless transmission module 4 needs to be installed on an unobstructed area at the top of the water quality testing tank 101. Open the equipment casing and tighten the antenna connection cable to the module interface. Finally, check that all pipe connections are secure and the circuits are unobstructed. After confirming that everything is in order, perform a power-on test.

[0135] The system initialization and parameter settings are as follows:

[0136] Before the system is powered on for the first time, initialization settings must be completed to ensure that all modules work together. Before powering on, the water quality sensor 102 must be manually checked for cleanliness: use a lint-free cloth dampened with pure water to gently wipe each sensor probe to remove dust or impurities that may have been picked up during transportation, so as not to affect the initial detection accuracy. Then connect the main power supply, and the system will automatically enter the self-test mode. The control module 3 will sequentially check the communication status of the water quality monitoring device 1, the intelligent sensor cleaning device 2, the wireless transmission module 4, etc. If any module is abnormal, the indicator light on the device casing will flash the corresponding fault code (e.g., E01 indicates that the water temperature sensor 1021 has failed to communicate, and E21 indicates that the adjustable pressure cleaning water pump 2021 is not responding). At this time, the power must be turned off to check whether the wiring is loose, and the fault must be eliminated before powering on again.

[0137] After passing the self-inspection, parameter initialization settings are required, including the detection cycle, cleaning threshold, and data upload frequency. The detection cycle is set remotely via the platform: for juvenile frog ponds (where water quality changes slowly), it can be set to 3 hours / time; for adult frog ponds (where water quality changes rapidly), it can be set to 1 hour / time; and for the overwintering period, it can be set to 12 hours / time. The cleaning threshold, which is the starting value for cleaning demand, is set to 40 by default (this can be adjusted according to the actual pollution level; for severely polluted ponds, it can be reduced to 30). The data upload frequency is usually consistent with the detection cycle to ensure that each detection data point is uploaded in real time.

[0138] System daily operation instructions:

[0139] Water Quality Monitoring Process: The system automatically starts water quality monitoring according to a preset detection cycle. The entire process requires no manual intervention. The specific process is as follows: When the set detection time is reached, the control module 3 first checks the status of the water quality detection chamber 101 (if it is in a state of soaking in clean water after low-frequency monitoring, it will first control the drain solenoid valve 1063 to drain the clean water), and then starts the water intake stage. During the water intake stage, the control module 3 triggers the water intake relay 1042 to start the water intake pump 1041. The water sample in the aquaculture pond is filtered by the water pump filter 1043 and then injected into the water quality detection chamber 101 through the water intake pipe 1044 at a flow rate of 15L / min. At this time, the water level in the water quality detection chamber 101 will gradually rise. When the water level reaches the water level limiter A1051, the limiter will send a high-level signal to the control module 3. The control module 3 will immediately cut off the water intake relay 1042 and stop the water intake. At this time, the water volume in the chamber is about 20L, which just meets the requirement of completely submerging all sensor probes.

[0140] After the water intake is completed, the system enters the water quality detection stage. The control module 3 will first control each sensor to preheat for 30 seconds (to ensure detection accuracy), and then start data acquisition: the water temperature sensor 1021, pH sensor 1022, ammonia nitrogen sensor 1023, turbidity sensor 1024, and dissolved oxygen sensor 1025 work synchronously, collecting a set of data every 10 seconds. After collecting 3 sets of data, the average value is taken as the final detection result. These data will be transmitted to the control module 3 in real time, and after calibration, a standardized report will be generated.

[0141] After the test is completed, the drainage stage begins. Control module 3 sends an opening command to drainage solenoid valve 1063, and the water sample in the chamber is discharged through the test chamber drain outlet 1061 at a drainage rate of approximately 20 L / min. When the water level drops to the water level limit switch B1052, the limit switch sends a high-level signal, and control module 3 closes drainage solenoid valve 1063. At this point, there is no residual water sample in the chamber, providing a clean environment for subsequent contamination testing.

[0142] Contamination Detection and Cleaning Execution: After drainage is completed, the system immediately initiates contamination detection, a crucial step in determining whether the sensors require cleaning. The light-emitting and receiving sensors 2025 sequentially align themselves with the probe sidewall of each sensor, emitting infrared light of a specific wavelength (850nm). The light is reflected by the probe sidewall and captured by the receiver. The control module 3 calculates the overall contamination level based on the reflected light intensity. Taking the ammonia nitrogen sensor 1023 as an example: the control module 3 first retrieves the baseline value of the reflected light intensity for the sensor's cleanliness status. =4500 lux, combined with the current reflected light intensity =2700 lux, calculate the current pollution level. =(4500-2700) / 4500×100%=40%; then retrieve the past 5 times (35%, 38%, 42%, 36%, 39%), calculate the average pollution level. =(35+38+42+36+39) / 5=38%; Simultaneously calculate the pollution change rate over the past hour. (If this is the first test, set this value to 0). = 5% / h; then the overall pollution level =0.6×40%+0.3×38%+0.1×5%=24%+11.4%+0.5%=35.9%. Therefore, Then, the system will further calculate the cleaning requirement. =35.9%×1.4×0.259×1.45≈35.9%×0.53≈19.0%. Because =19.0% < 40 (preset threshold), the system determines that no further cleaning is needed and proceeds directly to the data upload stage.

[0143] Data Upload and Comprehensive Analysis: Both water quality testing data and cleaning data are uploaded to the remote platform in real time via wireless transmission module 4. Uploaded data includes device ID, testing time, values ​​of various water quality parameters (water temperature, pH, ammonia nitrogen, turbidity, dissolved oxygen), comprehensive pollution index, cleaning demand (if cleaning is initiated), cleaning pressure, duration, and effectiveness coefficient. The remote platform stores the data (retaining one year of historical data), plots curves, and provides anomaly alarms: When a water quality parameter exceeds a preset threshold (e.g., ammonia nitrogen > 1.5 mg / L, dissolved oxygen < 3 mg / L), the platform sends an alert to the aquaculture manager via WeChat official account or SMS, simultaneously displaying the specific value of the abnormal parameter and the testing time.

[0144] The comprehensive analysis capabilities of the remote platform can help managers understand the correlation trends between water quality and sensor contamination. For example, analyzing data from a fishpond over seven consecutive days revealed a correlation between turbidity and ammonia nitrogen levels at sensor 1023. Positive correlation (correlation coefficient R) 2 =0.82), when the turbidity exceeds 60 NTU, the ammonia nitrogen sensor 1023 within 2 hours. An average increase of 15% indicates that increased suspended particles in the water accelerate sensor contamination. In such cases, the cleaning cycle can be adjusted in advance to prevent sensor performance degradation. Furthermore, the remote platform predicts the contamination risk for the next 24 hours based on historical data. Using an LSTM neural network model, it takes water quality parameters, cleaning data, and ambient temperature from the past three days as input and outputs data from each sensor. When the predicted value exceeds 60%, a decision support message is sent suggesting "cleaning in advance".

[0145] The embodiments described above are preferred embodiments of the present invention, but the present invention is not limited to the above embodiments. Any obvious improvements, substitutions or modifications that can be made by those skilled in the art without departing from the essence of the present invention shall fall within the protection scope of the present invention.

Claims

1. A high-density aquaculture water quality monitoring and sensor intelligent cleaning system, characterized in that, It comprises a water quality monitoring device (1), a sensor intelligent cleaning device (2), a control module (3), a wireless transmission module (4), and a power supply module (5). The water quality monitoring device (1) comprises a water quality detection bin (101) and a water inlet unit (104). The water quality detection bin (101) is internally provided with multiple types of sensors, and the inner wall is provided with a detection bin water level limiter (105). The bottom is provided with a detection bin drain (1061), and the detection bin drain (1061) is connected with a drain pipe. The drain pipe is provided with a PVC ball valve switch A (1062) controlled by the control module (3). The water inlet unit (104) comprises a water pump filter (1043) arranged in the breeding pond. The water pump filter (1043) is connected with the inlet of a water inlet pump (1041) through a water inlet pipeline (1044). The outlet of the water inlet pump (1041) is connected with the water quality detection bin (101) through a pipeline. The water inlet pump (1041) is also signal connected with a water inlet relay (1042), and the water inlet relay (1042) is connected with the control module (3) through a signal control line. The sensor intelligent cleaning device (2) comprises a clean water storage bin (201) and a cleaning unit (202). The inner wall of the clean water storage bin (201) is provided with a storage bin water level limiter (203), and the bottom is provided with a storage bin drain (2041). The storage bin drain (2041) is connected with a drain pipeline, and the drain pipeline is provided with a PVC ball valve switch B (2042) controlled manually. The side wall of the water quality detection bin (101) is provided with a moving track in the circumferential direction. The moving track is provided with a sliding mechanism. The light emitting and receiving sensor (2025) of the cleaning unit (202) is fixed on the sliding mechanism and is signal connected with the control module (3). The cleaning unit (202) further comprises an adjustable pressure cleaning water pump (2021). The inlet of the adjustable pressure cleaning water pump (2021) is connected with the clean water storage bin (201) through a clean water pipeline (2024). The outlet is connected with the water quality detection bin (101) through a pipeline. A plurality of cleaning lances (2023) are further arranged on the pipeline extending into the water quality detection bin (101). The adjustable pressure cleaning water pump (2021) is also signal connected with a cleaning relay (2022), and the cleaning relay (2022) is connected with the control module (3) through a signal control line.

2. The high-density aquaculture water quality monitoring and sensor intelligent cleaning method using the high-density aquaculture water quality monitoring and sensor intelligent cleaning system of claim 1, characterized in that, The process comprises the following steps: Step 1: water inlet; Step 2: water quality detection and data uploading; Step 3: water discharge after detection; Step 4: sensor contamination detection; Step 5: sensor cleaning decision and execution; Step 6: sensor cleaning data uploading; Step 7: comprehensive analysis and prediction based on multiple factors. 3.The method of claim 2, wherein, The step 1 comprises: the control module (3) starts the water inlet pump (1041) by triggering the water inlet relay (1042), and the water sample in the culture tank is filtered by the water pump filter (1043) to remove suspended impurities, and then is transported to the water quality detection bin (101) through the water inlet pipeline (1044); when the water level of the water sample reaches the position of the corresponding water level limiter in the water quality detection bin (101), the water level signal is fed back to the control module (3), and the control module (3) controls the water inlet pump (1041) to stop running, so that the water amount in the water quality detection bin (101) meets the sensor immersion detection requirement, and water overflow and waste are avoided. 4.The method of claim 2, wherein, The step 2 comprises: the sensors in the water quality detection bin (101) are started synchronously, the water sample is detected in multiple parameters, the real-time collected water quality data is transmitted to the control module (3) through a signal line, the control module (3) generates a standardized water quality parameter report after noise reduction and calibration processing of the original data, and then the report is uploaded to the remote monitoring platform through the wireless transmission module (4). 5.The method of claim 2, wherein, The step 3 comprises: after the water quality detection is completed, the control module (3) sends an opening instruction to the drain electromagnetic valve (1063), and the water sample in the water quality detection bin (101) is discharged through the detection bin drain (1061); when the water level drops to the position of the corresponding water level limiter, the control module (3) receives the signal and closes the drain electromagnetic valve (1063), so that there is no residual water sample in the water quality detection bin (101). 6.The method of claim 2, wherein, The step 4 comprises: Step 4.1: the reference value calibration in the sensor cleaning state is performed, and the reference value of the reflected light intensity of each sensor in the cleaning state is obtained; Step 4.2: the contamination degree detection of the first sensor is performed; The sliding mechanism carrying the light emitting and receiving sensor (2025) slides along the moving track under the instruction control of the control module (3), the light emitting and receiving sensor (2025) is first aligned with the sidewall surface of the first sensor, directional infrared light of a specific wavelength is emitted, the light is reflected after passing through the sensor surface and is captured by the receiver of the light emitting and receiving sensor (2025); the control module (3) calculates the current comprehensive contamination degree and cleaning requirement degree of the sensor according to the attenuation degree of the reflected light intensity; wherein the degree of overall pollution is calculated as follows: ； wherein, represents the current pollution degree, , is a reference value of the reflected light intensity of the sensor cleaning state, is the current reflected light intensity; is a weight coefficient of the current pollution degree, = 0.6-0.8; is a weight coefficient of the average pollution degree, = 0.2-0.4; is a weight coefficient of the pollution degree change rate, = 0.1-0.2; is the average pollution degree of the past times, , is the number of times of pollution degree detection participating in the average calculation in the past, which is adjusted according to the monitoring period, represents the current pollution degree of the th detection; is the pollution degree change rate, , represents the difference between the calculated last time and the calculated the time before last; represents the time interval between the calculated last time and the calculated the time before last; cleaning need The calculation is as follows: ； wherein, is a correction factor for each sensor type; represents a cleaning interval time factor, is the time since the last cleaning; represents a water quality turbidity factor, , is a real-time turbidity value;​ Step 4.3: the control module (3) continues to control the sliding mechanism to move, and the comprehensive contamination degree and cleaning requirement degree of the next sensor are calculated, and the calculation is performed on all sensors in this way. 7.The method of claim 2, wherein, The step 5 comprises: Step 5.1: the control module (3) compares the cleaning requirement degree of each sensor calculated in step 4 with the preset cleaning threshold value; Step 5.2: If threshold is exceeded, the cleaning unit (202) calculates cleaning parameters in real time, including cleaning pressure , single sensor cleaning duration ; if threshold is not exceeded, go directly to step 6; Cleaning pressure Is calculated as a piecewise function as follows: ； Single sensor single cleaning duration The calculation formula is: ； Step 5.3: the control module (3) starts the adjustable pressure cleaning water pump (2021) through the cleaning relay (2022); Step 5.4: the control module (3) controls the corresponding cleaning lance (2023) to perform directional flushing on the sensor needing cleaning in a preset time length at a calculated pressure. Step 5.5: Calculate the effect coefficient after cleaning : ； wherein, is the overall pollution degree before cleaning, is the overall pollution degree after cleaning; Step 5.6: When <60% indicates the need to enhance the cleaning strength, increase the cleaning pressure by 10% or extend the single sensor single cleaning time by 15%, then clean again, and calculate the effect coefficient again after cleaning is completed, judge whether three times of cleaning is needed, and so on until 60%≤ ≤90%. When > 90% indicates the need to reduce the cleaning intensity, reduce the cleaning pressure by 5% or shorten the single sensor cleaning time by 10%, then clean again, and calculate the effect coefficient again after cleaning is completed. Determine whether three cleanings are needed, and so on until 60% ≤ ≤ 90%. When 60%≤ ≤90%, no need to clean again, go to step 6 directly. 8.The method of claim 2, wherein, The step 7 comprises: the remote monitoring platform calls the built-in data analysis model to perform multi-dimensional correlation analysis on the water quality data uploaded by the wireless transmission module (4) and the cleaning data; evaluate the pollution trend of the sensor; judge the effectiveness of the cleaning operation; predict the sensor pollution risk level within the next 24 hours based on the machine learning algorithm of historical data; provide decision support for the aquaculture manager to adjust the cleaning frequency or water quality control measures in advance. 9.The method of claim 2, wherein, The step 8 comprises: Different maintenance strategies are designed for different aquaculture scenarios and sensor monitoring cycle differences. When the monitoring period When the monitoring period is ≤1 hour / time, the water quality detection chamber (101) is kept empty after step 6 is completed to avoid secondary pollution of the sensor caused by residual moisture, and step 1 is directly started when the next detection is performed to reduce invalid energy consumption. In this mode, the sensor is exposed to air for a short time and does not need additional moisturizing. When monitoring period 1 < When the time is ≤6 hours / time, after completing step 6, the control module (3) controls the cleaning spray gun (2023) of the cleaning unit (202) to spray a small amount of water mist onto the sensor probe for 10 seconds every 30 minutes to moisturize the sensor probe and ensure the response speed and detection accuracy of the sensor in the next detection. When the monitoring period When the monitoring period is 6 hours per time, after step 6 is completed, the control module (3) starts the cleaning unit (202) to inject clean water into the detection bin until the water level reaches the position of the corresponding water level limiter, and the wet state and activity of the sensor probe are maintained through clean water soaking. 5 minutes before the next detection, the drain electromagnetic valve (1063) is controlled to empty the clean water to avoid dilution of the new water sample by residual clean water and ensure the accuracy of the detection data.

Images