Integrated control system for pressurized flotation tanks through multimodal ai-based image and water quality analysis
The AI-integrated DAF system addresses real-time total phosphorus control issues by predicting concentrations using water quality and image analysis, optimizing coagulant use and process parameters, ensuring stable discharge quality and cost-efficiency.
Patent Information
- Application Number
- KR1020260089084
- Authority / Receiving Office
- KR · KR
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2026-03-05
- Filing Date
- 2026-05-18
- Publication Date
- 2026-11-10
- Estimated Expiration
- 2046-05-18
Smart Images

Figure 112026059627988-PAT00001_ABST
Abstract
Description
Technology Field
[0001] The present invention relates to an integrated control system for a Dissolved Air Flotation (DAF) tank based on multimodal artificial intelligence-based image and water quality analysis. Specifically, it relates to a technology that fuses and analyzes water quality measurement information of wastewater flowing into the DAF with image information capturing the shapes of flocs and bubbles inside the DAF using a pre-stored artificial intelligence algorithm, thereby predicting the total phosphorus (TP) concentration of treated wastewater in real time based on the analysis results, and optimally controlling the amount of coagulant injected into the wastewater and the operating parameters of the process equipment according to the predicted total phosphorus concentration of the treated wastewater. Background Technology
[0002] With the recent strengthening of the Total Water Pollution Load Management System, maintaining the concentration of Total Phosphorus (TP) in effluent from sewage and wastewater treatment facilities below legal standards has emerged as a key challenge. Consequently, the Dissolved Air Flotation (DAF) process, which removes pollutants by attaching microbubbles dissolved in high-pressure air to them, is being widely utilized in the advanced treatment stage. To maximize the efficiency of this DAF process, the amount of coagulant injected must be precisely adjusted in accordance with changes in the pollution load of the incoming sewage. However, there is a technical limitation in that, for Total Phosphorus (TP), a key management parameter, a significant measurement delay of typically 30 minutes to over an hour occurs during the analysis process using a Total Measurement System (TMS).
[0003] Due to the time lag in water quality measurements, it is difficult to respond in real-time to rapid fluctuations in inflow load at actual operational sites. Consequently, chronic problems have persisted, such as increased operating costs resulting from the over-injection of chemicals to prevent water quality accidents, or conversely, the deterioration of effluent water quality due to inappropriate responses.
[0004] Accordingly, the industry is developing various technologies to solve the aforementioned problems.
[0005] As an example, Korean registered patent 10-2365665 (automatic control device and control method for coagulant injection concentration for advanced treatment of phosphorus in wastewater, and wastewater treatment system capable of automatic control of coagulant injection concentration) discloses a technology for automatically controlling equipment by collecting environmental information.
[0006] However, the aforementioned prior art discloses only a technology for calculating the amount of coagulant to be injected to remove total phosphorus from influent water based on the turbidity of the influent water flowing into the wastewater treatment facility. It does not disclose a technology for predicting the total phosphorus concentration of treated wastewater in real time based on the analysis results by fusion-analyzing water quality measurement information of the sewage flowing into the pressurized flotation tank and image information capturing the shape of flocs and bubbles on the surface and inside the pressurized flotation tank using a pre-stored artificial intelligence algorithm, and for optimally controlling the amount of coagulant injected into the sewage and the operating parameters of the process equipment according to the predicted total phosphorus concentration of the treated wastewater. Consequently, there is a growing need for a technology capable of addressing this issue. The problem to be solved
[0007] Accordingly, the present invention is designed to solve the problems of the aforementioned existing technologies. Its purpose is to simultaneously achieve the stability of discharge water quality and the maximization of operational economics by fusion-analyzing water quality measurement information of wastewater flowing into a pressurized flotation tank with image information capturing the shapes of flocs and bubbles on the surface and inside the pressurized flotation tank using a pre-stored artificial intelligence algorithm, thereby predicting the total phosphorus concentration of the discharged water in real time based on the analysis results, and optimally controlling the amount of coagulant injected into the wastewater and the operating parameters of the process equipment according to the predicted total phosphorus concentration of the discharged water. means of solving the problem
[0008] A multimodal artificial intelligence-based integrated control system for a pressurized flotation tank through image and water quality analysis, implemented by a computing device comprising one or more processors and one or more memories for storing instructions executable by said processors according to one embodiment of the present invention, comprising: a sensor unit that generates water quality measurement information by measuring the inflow load, flow rate, and water quality of sewage flowing into the pressurized flotation tank from a coagulation tank which is the upstream end of the pressurized flotation tank; a camera module that generates image information by capturing the surface scum state and the bubble and floc shapes on the inner side of the pressurized flotation tank; and a water quality state prediction unit that, upon receiving the water quality measurement information from the sensor unit and the image information from the camera module, analyzes the correlation between the water quality measurement information and the image information through a previously stored artificial intelligence algorithm, while also reflecting real-time process state information in the analysis to predict the total phosphorus (TP) concentration for the discharged treated water of the pressurized flotation tank where measurement delay occurs. The apparatus is characterized by including an actuator control unit that, when the function of the water quality state prediction unit is completed, adjusts the amount of coagulant injected for coagulating pollutants in the sewage and the operating parameters of the pressurized flotation tank so that the predicted total phosphorus concentration satisfies a preset target effluent water quality, and adjusts the amount of coagulant injected and the operating parameters of the pressurized flotation tank so that the ratio of the amount of coagulant injected to the amount of total phosphorus inflow is minimized.
[0009] The above water quality measurement information is information that includes detailed information generated by a plurality of sensors included in the sensor unit installed in the upstream coagulation tank, and preferably includes information such as the flow rate of wastewater from the upstream coagulation tank flowing into the pressurized flotation tank, the turbidity of the incoming wastewater, the pH of the incoming wastewater, the electrical conductivity (EC) of the incoming wastewater, and the phosphorus (P) concentration of the incoming wastewater.
[0010] The above water quality state prediction unit preferably includes: a visual indicator extraction unit that, upon receiving the water quality measurement information from the sensor unit and the image information from the camera module, analyzes the image information through a stored artificial intelligence algorithm to extract visual process indicators including the size and density of flocs on the surface and inside the pressurized flotation tank, the phase separation state between water and solids, and the rising speed of microbubbles through the analysis results; a pollution level derivation unit that, upon completion of the extraction of the visual process indicators, analyzes the water quality measurement information through the stored artificial intelligence algorithm to derive the water quality pollution level based on the correlation between the electrical conductivity (EC) and turbidity changes of the incoming sewage through the analysis results; and a concentration information generation unit that, upon completion of the extraction of the visual process indicators and the derivation of the water quality pollution level, analyzes the correlation between the real-time process state information, the visual process indicators, and the water quality pollution level through the stored artificial intelligence algorithm to time-sequence the total phosphorus concentration for the discharged treated water with a measurement delay to complete the generation of treated water concentration information.
[0011] The artificial intelligence algorithm stored above may include: a first model trained to derive numerical feature values for the size and density of flocs in a pressurized flotation tank, visual boundary patterns resulting from the layer separation of water and solids, and rising speeds resulting from time-series displacement changes of bubbles generated inside, using multiple image information and final visual process indicators corresponding to the multiple image information as training information; a second model trained to derive correlation patterns between changes in the electrical conductivity (EC) and turbidity of sewage and water quality pollution levels using detailed information included in multiple water quality measurement information and final water quality pollution levels corresponding to the multiple water quality measurement information as training information; and a third model that learns time-series prediction patterns of total phosphorus concentration for discharged treated water by analyzing integrated training information generated as the feature values and correlation patterns are matched with multiple real-time process status information.
[0012] The above third model makes it possible to generate the treated water concentration information in which the error caused by environmental changes is offset, by applying the deviation in the reaction rate of the coagulant due to temperature changes among the external environmental information as a correction weight of the time series prediction pattern when external environmental information is received from a linked server and the received external environmental information matches the integrated learning information.
[0013] The actuator control unit preferably includes: an input cost calculation unit that, when the generation of the treated water concentration information is completed, calculates the total phosphorus load of the inflowing sewage through the water quality measurement information and calculates the chemical input cost, which is the ratio of the current amount of coagulant input to the calculated total phosphorus load; an operation effect index calculation unit that, when the calculation of the chemical input cost is completed, compares and analyzes the calculated chemical input cost and preset input history information, and calculates the operation effect index as a real-time monitoring indicator by comparing and analyzing based on a preset operation effect index calculation formula; and a feedback loop control unit that, when the calculation of the operation effect index is completed, adjusts the amount of coagulant input and the operating parameters of the pressurized flotation tank so that the predicted total phosphorus concentration based on the treated water concentration information is located within the allowable water quality range based on the preset target effluent water quality, and adjusts the operating parameters of the pressurized flotation tank while reducing the amount of coagulant input based on the operation effect index, which is the real-time monitoring indicator.
[0014] Preferably, the actuator control unit comprises: a real-time load calculation unit that calculates a real-time inflow load through water quality measurement information received from the sensor unit while controlling the plurality of actuators by adjusting the coagulant injection amount and the operating parameters of the pressurized flotation tank in real time; and a feedforward control unit that, when the calculated real-time inflow load exceeds a preset standard load range, calculates the design residence time of the pressurized flotation tank and applies the design residence time as a delay time to preemptively adjust the coagulant injection amount in response to the point when the inflow load reaches the reaction zone within the pressurized flotation tank.
[0015] Preferably, the actuator control unit comprises: an opening parameter adjustment unit that controls the plurality of actuators by adjusting the coagulant input amount and the operating parameters of the pressurized flotation tank in real time, and adjusts a parameter corresponding to the opening of the air compressor valve among the operating parameters so as to adjust the ratio of air solids in the pressurized flotation tank based on the rising speed and density of microbubbles through a visual process indicator extracted by a visual indicator extraction unit of the water quality state prediction unit; and an RPM parameter adjustment unit that improves solid removal efficiency by accelerating the rotation speed of the scraper among the operating parameters when the predicted total phosphorus concentration reaches a threshold of the allowable water quality range based on the preset target discharge water quality as the coagulant input amount is reduced and adjusted. Effects of the invention
[0016] Through the integrated control system for pressurized flotation tanks based on multimodal artificial intelligence image and water quality analysis of the present invention, the problem of a measurement delay of about one hour in automatic total phosphorus meters can be overcome, and the target discharge water quality can be stably achieved by combining feedforward control for inflow load fluctuations and feedback loop control based on the operational effectiveness index.
[0017] In addition, based on compliance with discharge water quality standards, it is possible to realize the operation of a pressurized flotation tank that is economical and ensures water quality stability by reducing reliance on chemicals and saving on chemical and electricity costs. Brief explanation of the drawing
[0018] FIG. 1 is a drawing illustrating a pressurized flotation tank integrated control system through image and water quality analysis based on multimodal artificial intelligence according to one embodiment of the present invention. FIG. 2 is a block diagram illustrating a water quality state prediction unit of a pressurized flotation tank integrated control system through image and water quality analysis based on multimodal artificial intelligence according to an embodiment of the present invention. FIG. 3 is a block diagram illustrating an actuator control unit of a pressurized flotation tank integrated control system through image and water quality analysis based on multimodal artificial intelligence according to an embodiment of the present invention. FIG. 4 is another block diagram illustrating the actuator control unit of a pressurized flotation tank integrated control system through image and water quality analysis based on multimodal artificial intelligence according to one embodiment of the present invention. FIG. 5 is another block diagram illustrating the actuator control unit of a pressurized flotation tank integrated control system through multimodal artificial intelligence-based image and water quality analysis according to one embodiment of the present invention. FIG. 6 is a drawing for explaining an example of the internal configuration of a computing device according to an embodiment of the present invention. Specific details for implementing the invention
[0019] Hereinafter, various embodiments and / or aspects are disclosed with reference to the drawings. For illustrative purposes, numerous specific details are disclosed in the following description to aid in a general understanding of one or more aspects. However, it will also be recognized by those skilled in the art that these aspects may be practiced without such specific details. The following description and the accompanying drawings describe specific exemplary aspects of one or more aspects in detail. However, these aspects are exemplary, and some of the various methods in the principles of the various aspects may be used, and the description is intended to include all such aspects and their equivalents.
[0020] As used herein, terms such as "examples," "examples," "aspects," "examples," etc., may not be interpreted as implying that any aspect or design described is better or more advantageous than other aspects or designs.
[0021] Additionally, the terms “comprising” and / or “comprising” should be understood to mean that the relevant feature and / or component is present, but not to exclude the presence or addition of one or more other features, components and / or groups thereof.
[0022] Additionally, terms including ordinal numbers, such as first, second, etc., may be used to describe various components, but said components are not limited by said terms. Such terms are used solely for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be named the second component, and similarly, the second component may be named the first component. The term "and / or" includes a combination of a plurality of related described items or any of a plurality of related described items.
[0023] Furthermore, in the embodiments of the present invention, all terms used herein, including technical or scientific terms, unless otherwise defined, have the same meaning as generally understood by those skilled in the art to which the present invention pertains. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in the embodiments of the present invention.
[0024] In the following description, it will be understood that the method of monitoring the tooth alignment status of a patient and checking the treatment status according to each embodiment of the present invention is performed by the device for monitoring the tooth alignment status of a patient and checking the treatment status according to each embodiment of the present invention illustrated in FIG. 5, which will be described later, and / or the computing device illustrated in FIG. 6. That is, it will be understood that the device of the present invention is implemented by combining one or more of the computing devices of FIG. 6.
[0025] FIG. 1 is a drawing illustrating a pressurized flotation tank integrated control system through image and water quality analysis based on multimodal artificial intelligence according to one embodiment of the present invention.
[0026] Referring to FIG. 1, a multimodal artificial intelligence-based integrated control system for image and water quality analysis (100) implemented as a computing device comprising one or more processors and one or more memories for storing instructions that can be executed by said processors may include a sensor unit (101), a camera module (103), a water quality state prediction unit (105), and an actuator control unit (107).
[0027] According to one embodiment, the sensor unit (101) can generate water quality measurement information by measuring the inflow load, flow rate, and water quality of the sewage flowing into the pressurized flotation tank from the coagulation tank which is the upstream end of the pressurized flotation tank.
[0028] In relation to the above, the above-mentioned shear refers to a coagulation tank, which is a shear section where a coagulant is introduced and mixed before sewage flows into the pressurized flotation tank, and the above-mentioned sensor section (101) may be composed of a plurality of sensors (e.g., flow meter, turbidity meter, pH meter, electrical conductivity sensor, phosphate phosphorus measuring instrument, etc.) installed in the above-mentioned shear section.
[0029] According to one embodiment, the water quality measurement information may include detailed information generated by a plurality of sensors included in a sensor unit (101) installed in the upstream section, and may include the flow rate of sewage in the upstream coagulation tank flowing into the pressurized flotation tank, the turbidity of the incoming sewage, the pH of the incoming sewage, the electrical conductivity (EC) of the incoming sewage, and the phosphorus (P) concentration of the incoming sewage.
[0030] In relation to the above, the flow rate of wastewater from the upstream coagulation tank flowing into the pressurized flotation tank is a value obtained by measuring the volume of wastewater flowing into the pressurized flotation tank per unit time through a flow meter, and may be information used to calculate the inflow load and the hydraulic retention time (HRT) of the pressurized flotation tank.
[0031] In addition, the turbidity of the incoming sewage is an optical measurement value that indirectly indicates the concentration of suspended solids in the sewage through a turbidimeter, and may serve as an indicator for detecting fluctuations in the inflow load in real time.
[0032] In addition, the pH of the incoming sewage is the hydrogen ion concentration of the sewage measured through a pH meter, and may be chemical condition information that affects the reaction efficiency of the coagulant. The electrical conductivity (EC) of the incoming sewage is a measurement value that indirectly indicates the dissolved ion concentration in the sewage through an electrical conductivity sensor, and may be information used to derive the degree of water pollution.
[0033] Finally, the phosphorus (P) concentration of the incoming wastewater mentioned above is the concentration of phosphorus components in the wastewater measured by a phosphate phosphorus meter, and may be information used when calculating the total phosphorus (TP) load.
[0034] According to one embodiment, the sensor unit (101) can generate water quality measurement information by integrating each detailed information measured through the plurality of sensors and transmit the generated water quality measurement information to the water quality state prediction unit (105).
[0035] For example, the sensor unit (101) can generate water quality measurement information by integrating the following when a flow rate of 500 m³ / h, turbidity of 15 NTU, pH of 7.2, electrical conductivity of 450 μS / cm, and phosphorus (P) concentration of 3.5 mg / L are measured.
[0036] At this time, the inflow load can be calculated as the product of the flow rate and the phosphorus (P) concentration, resulting in 500 m³ / h × 3.5 mg / L = 1,750 g / h.
[0037] According to one embodiment, the camera module (103) can generate image information by capturing the surface scum state and the bubble and floc shapes on the inner side of the pressurized flotation tank.
[0038] According to one embodiment, the camera module (103) may include a surface camera installed on the upper part of the pressurized flotation tank to photograph the state of scum layer formation on the surface and the state of phase separation between the water layer and the solid, and a side camera installed on the inner side of the pressurized flotation tank to photograph the upward behavior of microbubbles and the state of formation and growth of flocs.
[0039] In relation to the above, the scum refers to a layer of solids that has risen to the surface of the water by microbubbles within the pressurized flotation tank, and the thickness and uniformity of the scum layer may be indicators that can visually determine the processing efficiency of the pressurized flotation process.
[0040] In addition, the above floc is a mass formed by the binding of fine particles in the sewage by a coagulant, and the size and density of the floc may be an indicator of the efficiency of the coagulation reaction.
[0041] According to one embodiment, the camera module (103) can integrate image data captured through a surface camera and a side camera, respectively, to generate image information and transmit the generated image information to a water quality state prediction unit (105).
[0042] According to another embodiment, the camera module (103) can simultaneously operate the surface camera and the side camera to simultaneously monitor the surface scum condition of the pressurized flotation tank, the rising behavior of internal air bubbles, and the formation condition of internal flocs, thereby enabling the complex identification of the three-dimensional process conditions inside the pressurized flotation tank that are difficult to capture with a single viewpoint image.
[0043] Accordingly, the camera module (103) can improve the prediction accuracy of the water quality state prediction unit (105) described later by supplementarily obtaining the size distribution and rising speed change of internal bubbles, which cannot be confirmed by surface images alone, through side images.
[0044] According to one embodiment, when the water quality state prediction unit (105) receives the water quality measurement information from the sensor unit (101) and the image information from the camera module (103), it analyzes the correlation between the water quality measurement information and the image information through a stored artificial intelligence algorithm (109), while also reflecting real-time process state information in the analysis, thereby predicting the total phosphorus (TP) concentration of the discharged treated water of the pressurized flotation tank where measurement delay occurs.
[0045] In relation to the above, the water quality state prediction unit (105) can be implemented as an AI hub.
[0046] In relation to the above, the AI hub is an artificial intelligence processing device that integrally performs the functions of collecting and analyzing water quality data, collecting and analyzing image data, and generating abnormal state notifications and control commands, and may include a computation server that performs computations of the artificial intelligence algorithm (109) that is stored, a DB server that stores and manages process data including water quality measurement information and image information, and a middleware server that relays data linkage between the computation server and the DB server and communication with an external system.
[0047] In addition, the above AI hub is linked with an operation server via a Human Machine Interface (HMI) to provide an operating environment in which an administrator can set target effluent water quality, monitor process status, and query control history.
[0048] According to one embodiment, the artificial intelligence algorithm (109) stored above may include a first model learned to extract visual process features from image information, a second model learned to derive water quality pollution levels from water quality measurement information, and a third model learned to predict the total phosphorus concentration of discharged water in a time series by integrating the outputs of the first and second models and real-time process status information, and the specific configuration and operation of each model will be described in detail in FIG. 2.
[0049] According to one embodiment, the real-time process status information may be mechanical operation data collected in real time during the operation of the pressurized flotation tank, and may include information such as the current value of the air pump, the air-solid ratio (A / S ratio), the rotational speed (RPM) of the scraper, the circulation flow rate of the pressurized water pump, and the stirring speed of the mixing tank.
[0050] In relation to the above, the measurement delay may refer to a time delay phenomenon in which the total phosphorus concentration of the effluent at the current time cannot be confirmed in real time, as the measurement time required for about one hour occurs during the analysis process of the automatic total phosphorus (TP) meter.
[0051] Accordingly, the water quality state prediction unit (105) can perform a virtual sensing function that predicts the total phosphorus concentration in minutes by analyzing water quality measurement information and image information through a stored artificial intelligence algorithm (109) to overcome the measurement delay.
[0052] For example, when the water quality state prediction unit (105) confirms that the turbidity is 15 NTU and the electrical conductivity is 450 μS / cm in the water quality measurement information received from the sensor unit (101), and confirms that the floc size is 0.8 mm and the scum layer thickness is 12 mm in the image information received from the camera module (103), it can analyze the correlation between these using a stored artificial intelligence algorithm (109) to predict the total phosphorus concentration of the discharged treated water at the current time as 0.18 mg / L.
[0053] According to one embodiment, when the function of the water quality state prediction unit (105) is completed, the actuator control unit (107) can be operated.
[0054] According to one embodiment, when the function of the water quality state prediction unit (105) is completed, the actuator control unit (107) can adjust the amount of coagulant to be injected for coagulating pollutants in sewage and the operating parameters of the pressurized flotation tank so that the predicted total phosphorus concentration satisfies a preset target discharge water quality, and can adjust the amount of coagulant to be injected and the operating parameters of the pressurized flotation tank so that the ratio of the amount of coagulant injected to the total phosphorus load inflow is minimized.
[0055] According to one embodiment, the previously set target effluent water quality is an allowable upper limit value of total phosphorus (TP) concentration set by the manager for the effluent treated water of the pressurized flotation tank, and may be a value set according to the effluent water quality standards and operational goals of the sewage treatment plant. For example, the target effluent water quality may be set to a total phosphorus concentration of 0.2 mg / L or less.
[0056] In relation to the above, the coagulant is a chemical agent added to coagulate fine particles and phosphorus components in sewage into flocs, and PAC (polyaluminum chloride) can be used as a representative example. The amount of the coagulant added can be precisely controlled by adjusting the stroke length of the coagulant pump or the inverter frequency.
[0057] According to one embodiment, the operating parameter may include a control variable that determines the mechanical operating conditions of the pressurized flotation tank, and a parameter that adjusts the opening degree of the air compressor valve, the circulation flow rate of the pressurized water pump, and the rotational speed (RPM) of the scraper.
[0058] According to one embodiment, the actuator control unit (107) can be implemented with a TM / TC (Telemetry / Telecommand) and a plurality of PLCs (Programmable Logic Controllers).
[0059] The above TM / TC may be a remote monitoring and control device that receives a command to adjust the amount of coagulant and a command to adjust operating parameters from the AI hub, which is the water quality condition prediction unit (105), and transmits them to the plurality of PLCs.
[0060] The above plurality of PLCs may include a field controller that changes the physical operating conditions of the pressurized flotation tank according to a control command received from the TM / TC, a first PLC that is positioned in the upstream coagulation tank to directly drive and control the stroke and valve opening of the coagulant pump and collect measurement data from the water quality meter, and a second PLC that is positioned in the pressurized flotation tank to directly drive and control the opening of the air compressor valve, the circulation flow rate of the pressurized water pump, and the rotational speed of the scum remover.
[0061] According to one embodiment, the actuator control unit (107) determines whether the predicted total phosphorus concentration satisfies the preset target effluent water quality, and when the preset target effluent water quality is satisfied, it can maintain treatment efficiency while reducing dependence on chemicals by gradually reducing the amount of coagulant injected to minimize the ratio of the amount of coagulant injected to the amount of total phosphorus inflow, while simultaneously adjusting the operating parameters of the pressurized flotation tank to increase the proportion of mechanical control.
[0062] For example, if the total phosphorus concentration predicted by the water quality state prediction unit (105) satisfies the preset target discharge water quality of 0.2 mg / L or less with a total phosphorus concentration of 0.18 mg / L, the actuator control unit (107) calculates the chemical injection ratio, which is the ratio of the amount of coagulant injected to the total phosphorus load at the current time, and determines whether to reduce the amount of coagulant injected and the extent of reduction based on the operational effect index calculated by comparing the chemical injection ratio with the reference value before the business application.
[0063] Accordingly, the actuator control unit (107) can reduce the current amount of coagulant (PAC) input from 50 L / h to 45 L / h when it confirms that the operating effect index has increased compared to the previous control cycle, but if the predicted total phosphorus concentration approaches the threshold of the target discharge water quality or the operating effect index decreases due to the reduction, instead of maintaining the amount of coagulant input, it can increase the opening of the air compressor valve by 5% and accelerate the rotation speed of the scraper from 3 RPM to 3.5 RPM.
[0064] As a result, the actuator control unit (107) can compensate for the decrease in processing efficiency due to the reduction in the amount of coagulant injected by increasing the proportion of mechanical control.
[0065] In relation to the above, the method for calculating the operational effectiveness index and the detailed operation of the feedback loop control based thereon are described in Fig. 3 below.
[0066] FIG. 2 is a block diagram illustrating a water quality state prediction unit of a pressurized flotation tank integrated control system through image and water quality analysis based on multimodal artificial intelligence according to an embodiment of the present invention.
[0067] Referring to FIG. 2, a pressurized flotation tank integrated control system through multimodal artificial intelligence-based image and water quality analysis (e.g., the pressurized flotation tank integrated control system through multimodal artificial intelligence-based image and water quality analysis of FIG. 1 (100)) (hereinafter referred to as the integrated control system) implemented as a computing device comprising one or more processors and one or more memories for storing instructions that can be executed by said processors may include a water quality state prediction unit (200) (e.g., the water quality state prediction unit (105) of FIG. 1).
[0068] According to one embodiment, when the water quality state prediction unit (200) receives water quality measurement information (e.g., water quality measurement information of FIG. 1) from a sensor unit (e.g., sensor unit (101) of FIG. 1) and image information (e.g., image information of FIG. 1) from a camera module (e.g., camera module (103) of FIG. 1), it analyzes the correlation between the water quality measurement information and the image information through a stored artificial intelligence algorithm (207), while also reflecting real-time process state information in the analysis, thereby predicting the total phosphorus (TP) concentration of the discharged treated water of a pressurized flotation tank where measurement delay occurs.
[0069] According to one embodiment, the water quality state prediction unit (200) may include a visual indicator extraction unit (201), a pollution level derivation unit (203), and a concentration information generation unit (205) as detailed configurations for performing the above-described function.
[0070] According to one embodiment, when the visual indicator extraction unit (201) receives the water quality measurement information (201b) from the sensor unit and the image information (201a) from the camera module, it analyzes the image information (201a) through a previously stored artificial intelligence algorithm (207) and can extract visual process indicators including the size and density of flocs in the pressurized flotation tank, the state of phase separation between water and solids, and the rising speed of microbubbles through the analysis results.
[0071] In relation to the above, the model used in the visual indicator extraction unit (201) among the previously stored artificial intelligence algorithms (207) may be a first model trained to derive numerical feature values for the size and density of flocs in the pressurized flotation tank, the visual boundary pattern due to the layer separation of water and solids, and the rising speed due to the time-series displacement change of bubbles occurring inside, using multiple image information and the final visual process indicator corresponding to the multiple image information as training information.
[0072] In relation to the above, the first model may be a Convolutional Neural Network (CNN)-based image analysis model, which is trained to hierarchically extract spatial features from an input image through multiple convolutional layers, reduce the dimensionality of the feature map through pooling layers, and finally output numerical feature values corresponding to the size, density, phase separation boundary pattern, and bubble rising speed of the floc through a fully connected layer.
[0073] According to one embodiment, the first model can extract features corresponding to the thickness and uniformity of the scum layer from an image obtained from a surface camera of a pressurized flotation tank, and extract features corresponding to the size distribution and rising trajectory of microbubbles from an image obtained from a side camera, and generate a visual process index that integrates these.
[0074] According to one embodiment, the visual process indicator is a set of process state feature values quantified from image information (201a), wherein the size and density of the floc are quantified as the area of the floc region within the image and the number of flocs per unit area, the phase separation state between water and solid is quantified as the clarity and positional change of the visual boundary line between the water layer and the scum layer, and the rising speed of the microbubble may be a value quantified as the time-series displacement change of the bubble region between consecutive frames.
[0075] For example, the visual indicator extraction unit (201) can extract visual process indicators including a floc size of 0.8 mm, a floc density of 120 / cm², a phase separation boundary clarity of 0.85 (0 to 1 scale), and a microbubble rise speed of 2.3 mm / s from image information (201a) through a first model.
[0076] According to another embodiment, the first model applies an object detection-based model in addition to a convolutional neural network to independently identify individual flocs and bubbles within an image, and by individually tracking the size, shape, and movement trajectory of each object, it is possible to additionally include the size distribution histogram of the flocs and the velocity distribution histogram of the bubbles in the visual process indicators.
[0077] According to one embodiment, when the extraction of the visual process indicator is completed, the pollution level derivation unit (203) can analyze the water quality measurement information (201b) through the previously stored artificial intelligence algorithm (207) and derive the water quality pollution level based on the correlation between the electrical conductivity (EC) and turbidity changes of the incoming sewage through the analysis results.
[0078] According to one embodiment, the water quality pollution level is a value derived by analyzing the change patterns of the electrical conductivity (EC) and turbidity of the influent wastewater, and may be an indicator that quantitatively represents the relative concentration level of pollutants in the influent wastewater.
[0079] In relation to the above, since electrical conductivity (EC) indirectly reflects the total concentration of dissolved ions in sewage and turbidity reflects the concentration of suspended solids, the pollution level derivation unit (203) can derive a complex pollution level that is difficult to identify with a single measurement value by analyzing the correlation between the two measurements through the previously stored artificial intelligence algorithm (207).
[0080] According to one embodiment, the model used in the pollution level derivation unit (203) among the previously stored artificial intelligence algorithms (207) may be a second model trained to derive a correlation pattern between the change in electrical conductivity (EC) and turbidity of sewage and the water quality pollution level using detailed information included in a plurality of water quality measurement information and the final water quality pollution level corresponding to the plurality of water quality measurement information as training information.
[0081] In relation to the above, the second model may be a water quality analysis model based on a Multiple Linear Regression model or a Multi-Layer Perceptron (MLP) model.
[0082] In this case, the multiple linear regression model may be a model that derives the water quality pollution level by performing linear regression analysis with multiple water quality measurement variables, such as electrical conductivity (EC), turbidity, and pH, as independent variables and the water quality pollution level as the dependent variable, and learning regression coefficients that represent the contribution of each variable.
[0083] In addition, the above multilayer perceptron model may be a forward neural network composed of an input layer, one or more hidden layers, and an output layer, wherein when water quality measurements such as electrical conductivity, turbidity, and pH are input into the input layer, the hidden layer learns nonlinear correlations between variables through a nonlinear activation function, and the output layer derives the water quality pollution level.
[0084] According to one embodiment, the pollution level derivation unit (203) can derive the water quality pollution level of the incoming sewage as a numerical value by using the electrical conductivity, turbidity, pH, and phosphorus (P) concentration included in the water quality measurement information (201b) through the second model as input variables.
[0085] For example, when the pollution level derivation unit (203) identifies that the electrical conductivity is 450 μS / cm, the turbidity is 15 NTU, the pH is 7.2, and the phosphorus (P) concentration is 3.5 mg / L through the water quality measurement information (201b), it can analyze the water quality measurement information (201b) through the second model and derive a water quality pollution level of 0.72 (e.g., 0 to 1 scale, the closer to 1, the higher the pollution level).
[0086] According to another embodiment, the second model combines a multiple linear regression model and a multilayer perceptron model in an ensemble manner, and by applying a weighted average to the output values of each model, it is possible to improve the accuracy of deriving water quality pollution levels compared to a single model.
[0087] According to one embodiment, when the extraction of the visual process indicator and the derivation of the water quality pollution level are completed, the concentration information generation unit (205) can analyze the correlation between the real-time process status information, the visual process indicator (201a), and the water quality pollution level through the previously stored artificial intelligence algorithm, and time-sequence the total phosphorus concentration for the discharged treated water with a measurement delay to complete the generation of treated water concentration information.
[0088] According to one embodiment, the treated water concentration information is information that predicts the total phosphorus (TP) concentration of the discharged treated water as time series data in minutes based on the current time point, and may be a time series prediction result that includes the trend of change in total phosphorus concentration over a certain period (e.g., the next hour) rather than a predicted value at a single time point.
[0089] According to one embodiment, among the previously stored artificial intelligence algorithms, the model utilized in the concentration information generation unit (205) may be a third model that learns a time series prediction pattern of total phosphorus concentration for discharged treated water by analyzing integrated learning information generated as the numerical feature value derived from the first model of the visual indicator extraction unit (201) and the correlation pattern derived from the second model of the pollution level derivation unit (203) are matched with a plurality of real-time process status information.
[0090] According to one embodiment, the integrated learning information may be a multivariate time series data set composed of a single input vector in which quantified feature values derived from the first model (flock size, density, phase separation boundary clarity, bubble rise speed, etc.), water quality pollution level derived from the second model and associated correlation patterns, and real-time process status information (air pump current value, A / S ratio, scraper RPM, pressurized water pump circulation flow rate, etc.) are matched based on the same point in time.
[0091] That is, the above integrated learning information is multimodal fusion data in which image-based visual process features, sensor-based water pollution features, and mechanical operation state features are synchronized and combined on the time axis, and can be used as input data for learning and inference for the third model to predict the total phosphorus concentration of the discharged water in a time series.
[0092] For example, the concentration information generation unit (205) can generate treated water concentration information that predicts the total phosphorus concentration of the discharged treated water in minutes from the current time to the next 60 minutes by inputting into the third model as integrated inputs the quantified feature values derived from the first model, such as floc size 0.8 mm, floc density 120 pieces / cm², phase separation boundary clarity 0.85, and bubble rise speed 2.3 mm / s, and the water quality pollution level derived from the second model, such as 0.72, air pump current 12 A, A / S ratio 0.03, and scraper RPM 3.0 included in the real-time process status information.
[0093] At this time, the third model can generate treated water concentration information including a time series prediction pattern in which the predicted total phosphorus concentration at the current time is calculated as 0.18 mg / L, and recovers to 0.19 mg / L at the next 30 minutes by reflecting the gradual increase trend of the inflow load, and to 0.17 mg / L at the next 60 minutes.
[0094] According to one embodiment, the third model may be a time series prediction model based on LSTM (Long Short-Term Memory), which learns the long-term dependency of integrated learning information input in chronological order and performs time series prediction that reflects the influence of the process state change pattern at past time points on the total phosphorus concentration at future time points.
[0095] The above LSTM model can simultaneously learn long-term patterns and short-term fluctuations of time-series data through a cell structure composed of an input gate, a forget gate, and an output gate, and thus can be suitable for predicting total phosphorus concentration in environments where rapid changes in inflow load and gradual trend changes are mixed, such as in a pressurized flotation process.
[0096] According to one embodiment, the third model can generate a time series prediction pattern of total phosphorus concentration for discharged treated water by receiving as integrated inputs quantified feature values such as floc size and density, phase separation boundary clarity, and bubble rise velocity derived from the first model, and air pump current values, A / S ratio, and scraper RPM included in water quality pollution levels and real-time process status information derived from the second model.
[0097] According to one embodiment, when the third model receives external environment information from a linked server and the received external environment information matches the integrated learning information, it can generate the treated water concentration information in which the error due to environmental change is offset by applying the deviation in the reaction rate of the coagulant due to temperature change among the external environment information as a correction weight of the time series prediction pattern.
[0098] According to one embodiment, the external environment information may be real-time weather data received from a weather agency API or an environment monitoring server, and may include information such as temperature, amount of precipitation, whether precipitation has occurred, and duration of precipitation.
[0099] In relation to the above, the variation in the reaction rate of the coagulant due to changes in temperature may be a correction factor to reflect the phenomenon in which the reaction rate of the coagulant decreases in a low-temperature environment, resulting in reduced coagulation efficiency relative to the same input amount.
[0100] For example, the concentration information generation unit (205) can predict the total phosphorus concentration of the discharged treated water in minutes from the current time to the next 60 minutes through the third model, and if external environmental information such as the current temperature of 5 degrees and precipitation of 10 mm / h is received from the linked meteorological agency server, the reduction in the reaction rate of the coagulant due to low temperature is applied as a correction weight to generate treated water concentration information that reflects the environmental change.
[0101] In relation to the above, the temperature correction weight is a correction factor to reflect the phenomenon in which the chemical reaction rate of the coagulant decreases as the temperature drops relative to a preset reference temperature (e.g., 20 degrees), thereby reducing the coagulation efficiency relative to the same input amount, and may be a temperature-efficiency attenuation coefficient derived by learning the correlation between temperature and coagulation efficiency from past operating data.
[0102] For example, if the coagulation efficiency at a reference temperature of 20 degrees is normalized to 1.0, the temperature correction weight is calculated as 1.15 at the current temperature of 5 degrees, so the predicted total phosphorus concentration in the time series prediction pattern of the third model can be corrected up by 15%.
[0103] According to another embodiment, the integrated control system calculates the error between the predicted value of the treated water concentration information generated by the third model and the actual value received from the total phosphorus (TP) automatic measuring instrument after about one hour, and if it is determined that the prediction accuracy has decreased because the error exceeds a preset model drift threshold, the weight parameters of the third model can be conditionally updated based on the actual data accumulated recently.
[0104] Accordingly, the integrated control system performs a conditional self-correction function that maintains prediction reliability without unnecessary computational load, thereby adaptively updating the model to maintain prediction accuracy over the long term even when a discrepancy occurs between the prediction pattern of the previously learned third model and the actual process state due to seasonal changes, fluctuations in inflow characteristics, etc.
[0105] FIG. 3 is a block diagram illustrating an actuator control unit of a pressurized flotation tank integrated control system through image and water quality analysis based on multimodal artificial intelligence according to an embodiment of the present invention.
[0106] Referring to FIG. 3, a pressurized flotation tank integrated control system through multimodal artificial intelligence-based image and water quality analysis (e.g., the pressurized flotation tank integrated control system through multimodal artificial intelligence-based image and water quality analysis of FIG. 1 (100)) (hereinafter referred to as the integrated control system) implemented as a computing device comprising one or more processors and one or more memories for storing commands that can be executed by said processors may include an actuator control unit (300) (e.g., the actuator control unit (107) of FIG. 1).
[0107] According to one embodiment, the actuator control unit (300) can adjust the amount of coagulant to be injected for coagulating pollutants in sewage and the operating parameters of the pressurized flotation tank so that when the function of the water quality state prediction unit (e.g., water quality state prediction unit (105) of FIG. 1) is completed, the amount of coagulant to be injected and the operating parameters of the pressurized flotation tank are adjusted so that the ratio of the amount of coagulant injected to the amount of total phosphorus inflow is minimized.
[0108] According to one embodiment, the actuator control unit (300) may include an input cost calculation unit (301), an operation effect index calculation unit (303), and a feedback loop control unit (305) as detailed configurations for performing the above-described function.
[0109] According to one embodiment, when the generation of the treatment water concentration information is completed, the input ratio calculation unit (301) can calculate the total phosphorus load of the inflowing sewage through the water quality measurement information (301a) and calculate the chemical input ratio, which is the ratio of the amount of coagulant currently being input to the calculated total phosphorus load.
[0110] In relation to the above, the total phosphorus load is a value calculated by multiplying the flow rate and phosphorus (P) concentration included in the water quality measurement information (301a), and may be an indicator representing the mass of total phosphorus flowing into the pressurized flotation tank per unit time.
[0111] In relation to the above, the above chemical input ratio is a ratio obtained by dividing the amount of coagulant (PAC) being input at the current time by the total phosphorus load at the current time, and may be an operational efficiency indicator representing the amount of coagulant consumed per unit total phosphorus load.
[0112] For example, the above input ratio calculation unit (301) can calculate the total phosphorus load as 1,750 g / h by confirming the flow rate of 500 m³ / h and the phosphorus (P) concentration of 3.5 mg / L from the above water quality measurement information (301a), and when the current amount of coagulant (PAC) input is 50 L / h, the chemical input ratio (301a) can be calculated as 50 L / h χ 1,750 g / h = 0.0286 L / g.
[0113] According to one embodiment, when the calculation of the drug input cost is completed, the operation effect index calculation unit (303) compares and analyzes the calculated drug input cost and the preset input history information, and calculates the operation effect index as a real-time monitoring indicator by comparing and analyzing based on the preset operation effect index calculation formula (303a).
[0114] According to one embodiment, the pre-set input history information may be past data of the coagulant input amount and total phosphorus load accumulated in the existing operating method prior to the application of the integrated control system of the present invention, and may include information including the reference coagulant input amount (Baseline PAC) and the reference total phosphorus load (Baseline TP Load).
[0115] For example, the above-mentioned input history information may be set as an average value calculated based on operating data from a certain period prior to the application of the present invention (e.g., the previous 6 months).
[0116] According to one embodiment, the pre-set operational effect index calculation formula (303a) is a formula for quantitatively evaluating the coagulant reduction effect resulting from the application of the integrated control system of the present invention, and may be a formula set in the form of, for example, IBE = (1 - (Current PAC / Current TP Load) / (Baseline PAC / Baseline TP Load)) Х 100(%).
[0117] In relation to the above, Current PAC is the amount of coagulant being injected at the current time, Current TP Load is the total phosphorus load at the current time, Baseline PAC is the standard amount of coagulant injected before the application of the present invention, and Baseline TP Load may be the standard total phosphorus load before the application of the present invention.
[0118] According to one embodiment, the operational effectiveness index (IBE) may be a real-time operational efficiency indicator that indicates that the coagulant consumption per unit total load has decreased compared to the standard due to the application of the present invention as the positive (+) value increases, and indicates that the coagulant is being over-injected compared to the standard when it turns to a negative (-) value.
[0119] For example, the above-mentioned operational effect index calculation unit (303) can reflect the values in the above-mentioned operational effect index calculation formula (303a) when the current chemical input ratio calculated in the input ratio calculation unit (301) is 0.0286 L / g and the standard chemical input ratio based on the standard coagulant input amount (Baseline PAC) of 60 L / h and the standard total phosphorus load (Baseline TP Load) of 1,750 g / h in the pre-set input history information is 0.0343 L / g.
[0120] Accordingly, the above-mentioned operation effect index calculation unit (303) can calculate the operation effect index (IBE) as (1 - 0.0286 / 0.0343) Х 100 = 16.6 (%). This may mean that the coagulant consumption per unit total phosphorus load has decreased by 16.6% compared to the standard due to the application of the present invention.
[0121] According to one embodiment, the feedback loop control unit (305), when the calculation of the operation effect index is completed, adjusts the coagulant dosage and the operating parameters of the pressurized flotation tank so that the predicted total phosphorus concentration based on the treated water concentration information is within the allowable water quality range based on the preset target discharge water quality, and can adjust the operating parameters of the pressurized flotation tank while reducing the coagulant dosage based on the operation effect index, which is the real-time monitoring indicator.
[0122] According to one embodiment, the allowable water quality range may be a total phosphorus concentration range in which the upper limit is the target discharge water quality set by the manager, and the lower limit is the value obtained by subtracting a preset margin from the upper limit.
[0123] For example, if the target effluent water quality is 0.2 mg / L and the margin is 0.05 mg / L, the allowable water quality range can be set from 0.15 mg / L to 0.2 mg / L.
[0124] According to one embodiment, the feedback loop control unit (305) determines whether the predicted total phosphorus concentration satisfies the target effluent water quality, and if it does, can gradually reduce the amount of coagulant injected so that the operational effect index increases compared to the previous control cycle.
[0125] In relation to the above, the above-mentioned stepwise reduction may be a control process designed to minimize the risk of deterioration in effluent water quality due to reduction by gradually reducing the amount of coagulant input by a preset reduction unit (e.g., 5 to 10% of the current input amount) without drastically reducing the amount of coagulant input in a single control cycle.
[0126] According to one embodiment, the feedback loop control unit (305) can maintain the operating effect index in a maximized state by adjusting the operating parameters of the pressurized flotation tank to increase the mechanical control weight, instead of maintaining or increasing the amount of coagulant input, when the predicted total phosphorus concentration reaches the threshold of the target effluent water quality or the operating effect index decreases due to the reduction of the amount of coagulant input.
[0127] According to one embodiment, the threshold value may be a value close to the target effluent water quality, which is the upper limit of the allowable water quality range, and may be a boundary value indicating that if the predicted total phosphorus concentration reaches the value, there is a risk that additional reduction of coagulant will lead to exceeding the effluent water quality standard.
[0128] For example, if the target effluent water quality is 0.2 mg / L, the threshold can be set to 0.19 mg / L.
[0129] In relation to the above, increasing the mechanical control weight may mean increasing the air-solids ratio (A / S ratio) by increasing the opening of the air compressor valve instead of further reducing the amount of coagulant injected, or improving the solids removal efficiency by accelerating the rotational speed (RPM) of the scraper.
[0130] This allows for an operating method that maintains processing efficiency while reducing dependence on chemicals. The detailed operation of the above mechanical control is described in Fig. 5 below.
[0131] For example, the feedback loop control unit (305) can determine whether the reduction condition is satisfied when the currently predicted total phosphorus concentration is 0.17 mg / L, satisfying the target discharge water quality of 0.2 mg / L, and the operation effect index is 16.6%.
[0132] In relation to the above, the feedback loop control unit (305) may utilize the trend of the minimum margin and the operational effect index, which are pre-set as criteria for determining whether to reduce the amount of coagulant injected.
[0133] The above-mentioned minimum margin is a minimum safety distance between the predicted total phosphorus concentration and the threshold, and may be a discharge water quality protection standard value that allows for a reduction in the amount of coagulant added only when said margin is secured.
[0134] More specifically, the feedback loop control unit (305) checks whether the remaining margin between the predicted total phosphorus concentration and the threshold (0.19 mg / L) is greater than or equal to a preset minimum margin (e.g., 0.01 mg / L), and can confirm that the remaining margin between the current 0.17 mg / L and the threshold 0.19 mg / L is 0.02 mg / L, which exceeds the minimum margin of 0.01 mg / L.
[0135] In addition, the feedback loop control unit (305) can check whether the operating effect index has not decreased compared to the operating effect index of the previous control cycle, and can confirm that the current 16.6% has increased compared to the previous cycle's 16.0%.
[0136] Accordingly, when both of the two reduction conditions are satisfied, the feedback loop control unit (305) can reduce the amount of coagulant (PAC) input from 50 L / h to 45 L / h by 5 L / h according to a preset reduction unit (10% of the current input amount).
[0137] Subsequently, when the total phosphorus concentration predicted at the reduced input amount of 45 L / h reaches a threshold of 0.19 mg / L and the remaining margin becomes less than the minimum margin of 0 mg / L, the feedback loop control unit (305) determines that the additional reduction condition is not met and, instead of maintaining the coagulant input amount at 45 L / h, increases the opening of the air compressor valve by 5% and accelerates the scraper rotation speed from 3 RPM to 3.5 RPM to increase the mechanical control weight.
[0138] Accordingly, the feedback loop control unit (305) can restore the predicted total phosphorus concentration to 0.17 mg / L without further reducing the amount of coagulant added, thereby maintaining the operational effectiveness index in a maximized state.
[0139] According to one embodiment, the feedback loop control unit (305) can prevent an increase in chemical costs and energy waste caused by over-injection by issuing an efficiency reduction warning to the manager when the operating effect index turns to a negative (-) value, restoring the amount of coagulant injected to a preset reference value, and simultaneously automatically switching the mechanical operation parameters to an energy saving mode.
[0140] In relation to the above, the above-mentioned preset reference value may be a coagulant input amount calculated by multiplying the ratio of the baseline coagulant input amount (Baseline PAC) and the baseline total phosphorus load amount (Baseline TP Load) included in the above-mentioned preset input history information by the total phosphorus load amount at the current time.
[0141] That is, the above-mentioned reference value is the amount of coagulant added corresponding to an operational effectiveness index (IBE) of 0%, and when restored to that value, the chemical injection efficiency returns to the same level as before the application of the present invention, and from this point, the operational effectiveness index can be increased again by adjusting the mechanical operation parameters.
[0142] According to one embodiment, the energy saving mode is an operating mode that adjusts the operating rate of the air compressor to a minimum level proportional to the current load, reduces the circulation flow rate of the pressurized water pump to a preset lower limit, and lowers the rotational speed of the scraper to the minimum RPM required for scum removal, and may be an operating process for minimizing the consumption of chemical costs and electricity costs on the premise of compliance with discharge water quality standards.
[0143] FIG. 4 is another block diagram illustrating the actuator control unit of a pressurized flotation tank integrated control system through image and water quality analysis based on multimodal artificial intelligence according to one embodiment of the present invention.
[0144] Referring to FIG. 4, a pressurized flotation tank integrated control system through multimodal artificial intelligence-based image and water quality analysis (e.g., the pressurized flotation tank integrated control system through multimodal artificial intelligence-based image and water quality analysis of FIG. 1 (100)) (hereinafter referred to as the integrated control system) implemented as a computing device comprising one or more processors and one or more memories for storing instructions that can be executed by said processors may include an actuator control unit (400) (e.g., the actuator control unit (107) of FIG. 1).
[0145] According to one embodiment, when the function of the water quality state prediction unit (e.g., the water quality state prediction unit (105) of FIG. 1) is completed, the actuator control unit (400) can adjust the amount of coagulant to be injected for coagulating pollutants in sewage and the operating parameters of the pressurized flotation tank so that the predicted total phosphorus concentration satisfies the preset target discharge water quality, and can adjust the amount of coagulant to be injected and the operating parameters of the pressurized flotation tank so that the ratio of the amount of coagulant injected to the total phosphorus load inflow is minimized.
[0146] According to one embodiment, the actuator control unit (400) can perform feedforward control to preemptively respond to sudden changes in inflow load while controlling a plurality of actuators by adjusting the amount of coagulant injected and the operating parameters of the pressurized flotation tank in real time.
[0147] In relation to the above, the plurality of actuators may include a coagulant pump, an air compressor valve, a pressurized water pump, and a scraper drive motor, as a driving device that changes the physical operating conditions of the pressurized flotation tank according to the control of the actuator control unit (400).
[0148] According to one embodiment, the actuator control unit (400) may include a real-time load calculation unit (401) and a feedforward control unit (403) as detailed configurations for performing the above-described function.
[0149] According to one embodiment, the real-time load amount calculation unit (401) can calculate the real-time inflow load amount through water quality measurement information (401a) received from the sensor unit (e.g., sensor unit (101) of FIG. 1) while controlling the plurality of actuators by adjusting the coagulant input amount and the operating parameters of the pressurized flotation tank in real time.
[0150] In relation to the above, the real-time inflow load is a value calculated by multiplying the flow rate and phosphorus (P) concentration included in the water quality measurement information (401a) received from the sensor unit, and may be an indicator representing the mass of total phosphorus per unit time flowing into the pressurized flotation tank at the current time. The real-time inflow load may be continuously updated according to the measurement cycle of the sensor unit (e.g., every minute).
[0151] According to one embodiment, the feedforward control unit (403) can calculate the design residence time (HRT) of the pressurized flotation tank when the calculated real-time inflow load exceeds a preset reference load range, and apply the design residence time as a delay time to preemptively adjust the amount of coagulant injected in response to the time when the inflow load reaches the reaction zone within the pressurized flotation tank.
[0152] In relation to the above, the previously set reference load range is an inflow load range in a normal operating state set based on the design processing capacity of the pressurized flotation tank and past operating data, wherein the upper limit of the reference load range refers to the maximum inflow load at which the pressurized flotation tank can maintain normal processing efficiency, and the lower limit refers to the minimum inflow load to secure the minimum efficiency of chemical injection. For example, the reference load range may be set to 1,000 g / h to 2,000 g / h.
[0153] For example, the feedforward control unit (403) can confirm that the real-time inflow load is calculated as 800 m³ / h × 4.0 mg / L = 3,200 g / h by the performance of the function of the real-time load amount calculation unit (401), and that the calculated 3,200 g / h exceeds the upper limit of 2,000 g / h of the preset standard load amount range.
[0154] According to one embodiment, the design retention time (HRT, Hydraulic Retention Time) is a value obtained by dividing the effective volume of the upstream coagulation tank by the flow rate of the sewage, and may represent the average time that the inflowed sewage remains inside the upstream coagulation tank.
[0155] Accordingly, the above design residence time is recalculated in real time based on fluctuations in the inflow rate, and the residence time may be shortened as the inflow rate increases and extended as the inflow rate decreases.
[0156] According to one embodiment, the delay time is the time required for the inflow load fluctuation detected in the shear section to actually reach the reaction area inside the pressurized flotation tank, and can be applied identically to the design residence time.
[0157] Accordingly, the feedforward control unit (403) can achieve a state in which an appropriate amount of coagulant is already injected at the time the inflow load reaches the reaction area by preemptively adjusting the amount of coagulant injected by the delay time when the real-time inflow load exceeds the preset reference load range.
[0158] In relation to the above, the reaction zone refers to a section in which a coagulant and sewage are mixed in the shear coagulation tank of the pressurized flotation tank to form flocs and separate flotation, and may be a zone reached by sewage flowing in from the shear section after passing through a delay time corresponding to the design retention time.
[0159] According to one embodiment, when the feedforward control unit (403) preemptively adjusts the amount of coagulant to be injected, it can calculate the increase in the amount of coagulant to be injected in proportion to the increase in the inflow load.
[0160] At this time, the above increase range may be a value calculated by multiplying the difference between the current real-time inflow load and the upper limit of the preset standard load range by a preset load-chemical proportionality coefficient.
[0161] In relation to the above, the pre-set load-chemical proportionality coefficient is a coefficient derived by regression analysis of the proportional relationship between the increase in the inflow load amount and the corresponding increase in the coagulant injected from past operating data, and may be a value representing the additional amount of coagulant required per unit load excess.
[0162] For example, if the load-chemical proportionality coefficient is 0.005 L / g, it may mean that the amount of coagulant added is increased by 0.005 L / h whenever the inflow load exceeds the upper limit of the standard load range by 1 g / h.
[0163] For example, the feedforward control unit (403) can calculate the design retention time (HRT) as 0.125 hours (7.5 minutes) by dividing the effective volume of the pressurized flotation tank 100 m³ by the current flow rate 800 m³ / h when the real-time inflow load of 3,200 g / h exceeds the upper limit of the preset standard load range of 2,000 g / h by 1,200 g / h.
[0164] Accordingly, the feedforward control unit (403) predicts that the increased inflow load will reach the reaction zone 7.5 minutes after the current time, and by applying the load-chemical proportionality coefficient 0.005 L / g to calculate the increase range as 1,200 g / h × 0.005 L / g = 6 L / h, the current coagulant input amount of 50 L / h is immediately and preemptively increased to 56 L / h, which is the sum of the increase range of 6 L / h, thereby enabling the implementation of a state in which an appropriate amount of coagulant is pre-injected for the increased inflow load that reaches the reaction zone 7.5 minutes later.
[0165] According to another embodiment, the feedforward control unit (403) can simultaneously adjust the operating parameters of the pressurized flotation tank and the amount of coagulant injected.
[0166] For example, when the feedforward control unit (403) determines that it is difficult to secure treatment efficiency by increasing the amount of coagulant alone when the excess amount of the inflow load exceeds a preset rapid change threshold (e.g., 200% or more of the upper limit of the reference load range), it can increase the amount of microbubbles generated by increasing the amount of coagulant and increasing the circulation flow rate of the pressurized water pump, and increase the opening of the air compressor valve to adjust the air solid ratio (A / S ratio) upward.
[0167] That is, the feedforward control unit (403) can respond to rapid fluctuations in inflow load by simultaneously strengthening chemical control and mechanical control.
[0168] FIG. 5 is another block diagram illustrating the actuator control unit of a pressurized flotation tank integrated control system through multimodal artificial intelligence-based image and water quality analysis according to one embodiment of the present invention.
[0169] Referring to FIG. 5, a pressurized flotation tank integrated control system through multimodal artificial intelligence-based image and water quality analysis (e.g., the pressurized flotation tank integrated control system through multimodal artificial intelligence-based image and water quality analysis of FIG. 1 (100)) (hereinafter referred to as the integrated control system) implemented as a computing device comprising one or more processors and one or more memories for storing instructions that can be executed by said processors may include an actuator control unit (500) (e.g., the actuator control unit (107) of FIG. 1).
[0170] According to one embodiment, when the function of the water quality state prediction unit (e.g., the water quality state prediction unit (105) of FIG. 1) is completed, the actuator control unit (500) can adjust the amount of coagulant to be injected for coagulating pollutants in sewage and the operating parameters of the pressurized flotation tank so that the predicted total phosphorus concentration satisfies the preset target discharge water quality, and can adjust the amount of coagulant to be injected and the operating parameters of the pressurized flotation tank so that the ratio of the amount of coagulant injected to the total phosphorus load inflow is minimized.
[0171] In relation to the above, the actuator control unit (500) can perform mechanical control to compensate for the decrease in processing efficiency resulting from the reduction in the amount of coagulant injected by adjusting mechanical operating variables while controlling multiple actuators by adjusting the amount of coagulant injected and the operating parameters of the pressurized flotation tank in real time.
[0172] According to one embodiment, the actuator control unit (500) may include an opening parameter adjustment unit (501) and an RPM parameter adjustment unit (503) as detailed configurations for performing the above-described function.
[0173] According to one embodiment, the opening parameter adjustment unit (501) can adjust a parameter corresponding to the opening of the air compressor valve among the operating parameters so that the air solid ratio (A / S ratio) in the pressurized flotation tank is adjusted based on the rising speed and density of microbubbles through a visual process indicator (501a) extracted from a visual indicator extraction unit (e.g., visual indicator extraction unit (201) of FIG. 2) while controlling the plurality of actuators by adjusting the amount of coagulant input and the operating parameters of the pressurized flotation tank in real time.
[0174] According to one embodiment, the air-solids ratio (A / S ratio) is the ratio between the total amount of microbubbles generated in the pressurized flotation tank and the total amount of solids to be treated, and may be a key operating indicator for determining flotation separation efficiency.
[0175] Accordingly, if the above A / S ratio is excessively low, the microbubbles are insufficient to bring solid matter to the surface, resulting in reduced treatment efficiency, and if it is excessively high, energy may be wasted due to unnecessary air supply.
[0176] According to one embodiment, the opening degree of the air compressor valve is the degree of valve opening that determines the flow rate of pressurized air supplied to the pressurized flotation tank; as the opening degree increases, the amount of pressurized air supplied increases, thereby increasing the amount of microbubbles generated and raising the A / S ratio, and as the opening degree decreases, the amount of pressurized air supplied decreases, thereby lowering the A / S ratio.
[0177] According to one embodiment, the opening parameter adjustment unit (501) can determine the opening adjustment direction and adjustment range of the air compressor valve by comparing the rising speed and density of the microbubbles included in the visual process indicator (501a) with a preset bubble state reference range.
[0178] In relation to the above, the aforementioned pre-set bubble state standard range is a standard value for the rising speed and density ranges of microbubbles in which the flotation separation efficiency of the pressurized flotation tank is optimally maintained, and may be a value set by statistically analyzing the bubble state of the range in which the treatment efficiency stably achieved the target effluent water quality in past operating data.
[0179] For example, the above bubble state standard range can be set to an upward speed of 2.0 to 3.0 mm / s and a density of 150 to 250 bubbles / cm².
[0180] According to one embodiment, the opening parameter adjustment unit (501) can increase the opening of the air compressor valve by determining that the amount of microbubbles generated is insufficient when the rising speed of the microbubbles is below the lower limit of a preset bubble state reference range or the density is below the lower limit.
[0181] According to one embodiment, the opening parameter adjustment unit (501) can reduce the opening of the air compressor valve by determining that the microbubbles are being oversupplied when the rising speed of the microbubbles exceeds the upper limit of a preset bubble state reference range or when the density exceeds the upper limit.
[0182] For example, the above-mentioned opening parameter adjustment unit (501) can confirm that the rising speed of microbubbles is 1.5 mm / s and the density is 100 bubbles / cm² in the visual process indicator (501a) received from the visual indicator extraction unit, and that each falls short of the lower limit of the preset bubble state standard range (rising speed 2.0~3.0 mm / s, density 150~250 bubbles / cm²).
[0183] Accordingly, the opening parameter adjustment unit (501) can determine the opening adjustment range by calculating the deviation rate between the current bubble state and the lower limit of the preset bubble state reference range.
[0184] In relation to the above, the deviation rate is a value expressed as a percentage indicating the degree to which the current measured value falls short of the lower limit of the bubble state standard range, and the deviation rate of the rising speed can be calculated as (2.0 - 1.5) / 2.0 X 100 = 25%, and the deviation rate of the density can be calculated as (150 - 100) / 150 X 100 = 33.3%.
[0185] Accordingly, the opening parameter adjustment unit (501) can calculate the opening adjustment range by multiplying the larger value between the rising speed deviation rate and the density deviation rate, which is 33.3%, by a preset opening-deviation proportionality coefficient.
[0186] In relation to the above, the above-mentioned opening-deviation proportionality coefficient is the ratio at which the opening of the air compressor valve must be adjusted per 1% deviation rate of the bubble state, and may be a coefficient derived by analyzing the correlation between the amount of change in valve opening and the resulting change in bubble state in past operating data. For example, if the above-mentioned opening-deviation proportionality coefficient is 0.3%p / %, it may mean adjusting the valve opening by 0.3% points per 1% deviation rate.
[0187] Accordingly, the above-mentioned opening parameter adjustment unit (501) calculates the opening adjustment range as 10 percentage points by multiplying the maximum deviation rate of 33.3% by the opening-deviation proportionality coefficient of 0.3%p / %, and increases the current opening of the air compressor valve from 40% to 50%, thereby increasing the pressurized air supply amount and upgrading the A / S ratio.
[0188] According to another embodiment, the opening parameter adjustment unit (501) can adjust the circulation flow rate of the pressurized water pump in parallel with the opening adjustment of the air compressor valve.
[0189] In relation to the above, the circulation flow rate of the pressurized water pump is a flow rate that recirculates a portion of the treated water from the pressurized flotation tank back to the pressurized flotation tank via the pressurized dissolution tank, and as the circulation flow rate increases, the total amount of microbubbles generated as the pressurized dissolved air is depressurized inside the pressurized flotation tank may increase.
[0190] Accordingly, the above-mentioned opening parameter adjustment unit (501) can additionally increase the circulation flow rate of the pressurized water pump to supplementarily secure the amount of microbubbles generated when it is difficult to achieve the target level of the A / S ratio by only increasing the opening of the air compressor valve.
[0191] According to one embodiment, the RPM parameter adjustment unit (503) can improve solid removal efficiency by adjusting the parameter corresponding to the scraper installed in the pressurized flotation tank among the operating parameters when the predicted total phosphorus concentration reaches a threshold of the allowable water quality range based on the preset target discharge water quality as the amount of coagulant injected is reduced and adjusted, thereby accelerating the rotation speed of the scraper.
[0192] According to one embodiment, the scraper is a device for mechanically removing scum (a layer of solids) floating on the water surface of a pressurized flotation tank, and can operate in such a way that a rotating blade pushes the scum on the water surface into a scum collection tank.
[0193] In relation to the above, as the rotational speed (RPM) of the scraper increases, the amount of scum removed per unit time increases, thereby improving the efficiency of solid removal; however, if it is excessively fast, a counterproductive effect may occur in which scum that has risen due to disturbance of the water surface is mixed back into the lower layer.
[0194] According to one embodiment, the RPM parameter adjustment unit (503) may be configured to compensate for the limitations of chemical control in the section where the total phosphorus concentration of the treated water increases due to a decrease in floc formation efficiency caused by a reduction in the amount of coagulant injected, by improving the mechanical removal efficiency of the scraper.
[0195] According to one embodiment, when the RPM parameter adjustment unit (503) confirms that the predicted total phosphorus concentration has reached the threshold of the allowable water quality range, it can compare the current rotational speed of the scraper with a preset maximum allowable RPM and accelerate the rotational speed within the maximum allowable RPM.
[0196] In relation to the above, the above-mentioned maximum allowable RPM is a critical rotational speed at which processing efficiency is actually reduced due to scum mixing caused by water surface disturbance when the scraper's rotational speed exceeds the set RPM value, and may be a value set according to the design specifications of the pressurized flotation tank and the shape of the scraper blade. For example, the above-mentioned maximum allowable RPM may be set to 5.0 RPM.
[0197] For example, the RPM parameter adjustment unit (503) can determine the RPM adjustment range by calculating the risk of water quality exceeding the predicted total phosphorus concentration and the target discharge water quality when the predicted total phosphorus concentration reaches the threshold of the allowable water quality range at 0.19 mg / L after the coagulant dosage is reduced and adjusted from 50 L / h to 45 L / h by the feedback loop control unit (e.g., the feedback loop control unit (305) of FIG. 3).
[0198] In relation to the above, the risk of exceeding the water quality can be calculated as the ratio of the deviation between the predicted total phosphorus concentration and the lower limit of the allowable water quality range to the width of the allowable water quality range, which is (0.19 - 0.15) / (0.20 - 0.15) Х 100 = 80%.
[0199] That is, the RPM parameter adjustment unit (503) can calculate the RPM adjustment range by multiplying the water quality excess risk by a preset RPM-risk proportionality coefficient.
[0200] In relation to the above, the above-mentioned pre-set RPM-risk proportionality coefficient is a ratio at which the scraper rotation speed must be adjusted per 1% of water quality excess risk, and may be a coefficient derived by analyzing the correlation between the amount of change in scraper RPM and the resulting change in solid removal efficiency in past operation data.
[0201] For example, if the above RPM-risk proportionality coefficient is 0.00625 RPM / %, it may mean adjusting the scraper rotation speed by 0.00625 RPM for every 1% of water quality excess risk.
[0202] Accordingly, the RPM parameter adjustment unit (503) calculates an RPM adjustment range of 0.5 RPM by multiplying the water quality excess risk of 80% by an RPM-risk proportionality coefficient of 0.00625 RPM / %, and accelerates the current scraper rotation speed of 3.0 RPM to 3.5 RPM, which is within the preset maximum allowable RPM of 5.0 RPM, thereby increasing the amount of scum removed per unit time and improving the efficiency of removing solid matter.
[0203] According to another embodiment, if the predicted total phosphorus concentration continues to exceed the threshold of the allowable water quality range even after the RPM parameter adjustment unit (503) accelerates the rotation speed of the scraper, the opening of the air compressor valve is further increased in conjunction with the opening parameter adjustment unit (501), thereby simultaneously achieving improved solid removal efficiency by the scraper and improved flotation separation efficiency due to increased microbubble generation.
[0204] According to another embodiment, the RPM parameter adjustment unit (503) can determine the rotation speed adjustment range of the scraper by additionally referring to the thickness information of the scum layer included in the visual process indicator (501a).
[0205] At this time, the RPM parameter adjustment unit (503) can calculate the excess rate between the current scum layer thickness and the upper limit of the preset normal thickness range, and determine the RPM adjustment range by multiplying the excess rate by a preset RPM-scum proportionality coefficient.
[0206] In relation to the above, the above excess rate may be a value representing the degree to which the current scum layer thickness exceeds the upper limit of a preset normal thickness range as a percentage relative to the upper limit.
[0207] In relation to the above, the above-mentioned pre-set RPM-scum proportionality coefficient is a ratio at which the scraper rotation speed must be adjusted per 1% excess rate of the scum layer thickness, and may be a coefficient derived by analyzing the correlation between the amount of change in scum layer thickness and the amount of change in scraper RPM in past operation data.
[0208] For example, if the above RPM-scum proportionality coefficient is 0.045 RPM / %, it may mean adjusting the scraper rotation speed by 0.045 RPM per 1% of excess rate.
[0209] For example, if the RPM parameter adjustment unit (503) determines that the scum layer thickness in the visual process indicator (501a) is 20 mm, which is the upper limit of the preset normal thickness range (8 to 15 mm), the excess rate is calculated as (20 - 15) / 15 Х 100 = 33.3%, and the RPM adjustment range is calculated as 1.5 RPM by multiplying by the RPM-scum proportionality coefficient 0.045 RPM / %.
[0210] Accordingly, the RPM parameter adjustment unit (503) can accelerate the removal of excessively accumulated scum by adding an adjustment range of 1.5 RPM to the current scraper rotation speed of 3.0 RPM to accelerate it to 4.5 RPM.
[0211] On the other hand, if the RPM parameter adjustment unit (503) confirms that the scum layer thickness is within the normal thickness range of 10 mm, the excess rate is calculated as 0%, so RPM adjustment due to scum accumulation is unnecessary, but only an RPM adjustment range of 0.5 RPM based on the water quality excess risk is applied, thereby slightly accelerating the current rotation speed of 3.0 RPM to 3.5 RPM, thereby enabling adaptive RPM adjustment proportional to the scum accumulation state.
[0212] FIG. 6 is a drawing for explaining an example of the internal configuration of a computing device according to an embodiment of the present invention.
[0213] FIG. 6 illustrates an example of the internal configuration of a computing device according to an embodiment of the present invention. In the following description, descriptions of unnecessary embodiments that overlap with the descriptions of FIG. 1 to 6 described above will be omitted.
[0214] As illustrated in FIG. 6, the computing device (10000) may include at least one processor (11100), memory (11200), peripheral interface (11300), input / output subsystem (I / O subsystem) (11400), power circuit (11500), and communication circuit (11600). In this case, the computing device (10000) may correspond to a user terminal (A) connected to a haptic interface device or the aforementioned computing device (B).
[0215] The memory (11200) may include, for example, high-speed random access memory, a magnetic disk, SRAM, DRAM, ROM, flash memory, or non-volatile memory. The memory (11200) may include software modules, instruction sets, or various other data required for the operation of the computing device (10000).
[0216] At this time, access to memory (11200) from other components, such as the processor (11100) or peripheral device interface (11300), can be controlled by the processor (11100).
[0217] The peripheral device interface (11300) can connect input and / or output peripheral devices of the computing device (10000) to the processor (11100) and memory (11200). The processor (11100) can perform various functions for the computing device (10000) and process data by executing software modules or instruction sets stored in the memory (11200).
[0218] The input / output subsystem (11400) can connect various input / output peripherals to the peripheral interface (11300). For example, the input / output subsystem (11400) may include a controller for connecting peripherals such as a monitor, keyboard, mouse, printer, or, if necessary, a touchscreen or sensor to the peripheral interface (11300). According to another aspect, input / output peripherals may be connected to the peripheral interface (11300) without passing through the input / output subsystem (11400).
[0219] The power circuit (11500) can supply power to all or part of the components of the terminal. For example, the power circuit (11500) may include one or more power sources such as a power management system, a battery or alternating current (AC), a charging system, a power failure detection circuit, a power converter or inverter, a power status indicator, or any other components for power generation, management, and distribution.
[0220] The communication circuit (11600) can enable communication with another computing device using at least one external port.
[0221] Alternatively, as described above, the communication circuit (11600) may enable communication with other computing devices by including an RF circuit and transmitting and receiving an RF signal, also known as an electromagnetic signal.
[0222] The embodiment of FIG. 6 is merely an example of a computing device (10000), and the computing device (11000) may have some components shown in FIG. 6 omitted, additional components not shown in FIG. 6 added, or a configuration or arrangement that combines two or more components. For example, a computing device for a communication terminal in a mobile environment may include, in addition to the components shown in FIG. 6, a touchscreen or a sensor, etc., and the communication circuit (1160) may include a circuit for RF communication of various communication methods (WiFi, 3G, LTE, Bluetooth, NFC, Zigbee, etc.). The components that can be included in the computing device (10000) may be implemented as hardware, software, or a combination of both hardware and software, including one or more integrated circuits specialized for signal processing or applications.
[0223] Methods according to embodiments of the present invention may be implemented in the form of program instructions that can be executed through various computing devices and recorded on a computer-readable medium. In particular, the program according to the present embodiment may be configured as a PC-based program or an application dedicated to a mobile terminal. An application to which the present invention is applied may be installed on a user terminal through a file provided by a file distribution system. For example, the file distribution system may include a file transmission unit (not shown) that transmits the file upon a request from the user terminal.
[0224] The device described above may be implemented as a hardware component, a software component, and / or a combination of a hardware component and a software component. For example, the device and components described in the embodiments may be implemented using one or more general-purpose or special-purpose computers, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing unit may execute an operating system (OS) and one or more software applications executed on said operating system. Additionally, the processing unit may access, store, manipulate, process, and generate data in response to the execution of the software. For ease of understanding, the processing unit may be described as being used as a single unit, but those skilled in the art will understand that the processing unit may include a plurality of processing elements and / or a plurality of types of processing elements. For example, the processing unit may include multiple processors or one processor and one controller. Additionally, other processing configurations, such as parallel processors, are also possible.
[0225] Software may include computer programs, code, instructions, or a combination of one or more of these, and may configure a processing unit to operate as desired or command the processing unit independently or collectively. Software and / or data may be permanently or temporarily embodied in any type of machine, component, physical device, virtual equipment, computer storage medium, or device so as to be interpreted by the processing unit or to provide instructions or data to the processing unit. Software may be distributed across networked computing devices and stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.
[0226] The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc., either alone or in combination. The program instructions recorded on the medium may be those specifically designed and configured for the embodiment, or they may be those known and available to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical recording media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, and flash memory. Examples of program instructions include machine code, such as that generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc. The hardware devices described above may be configured to operate as one or more software modules to perform the operation of the embodiment, and vice versa.
[0227] Although the embodiments have been described above with reference to limited embodiments and drawings, those skilled in the art can make various modifications and variations from the description above. For example, appropriate results may be achieved even if the described techniques are performed in a different order than described, and / or if the components of the described system, structure, device, circuit, etc. are combined or assembled in a form different from described, or replaced or substituted by other components or equivalents. Therefore, other implementations, other embodiments, and equivalents to the claims below also fall within the scope of the claims.
Claims
Claim 1 A multimodal artificial intelligence-based integrated control system for a pressurized flotation tank through image and water quality analysis, implemented by a computing device comprising one or more processors and one or more memories for storing instructions executable by said processors, comprising: a sensor unit that generates water quality measurement information by measuring the inflow load, flow rate, and water quality of sewage flowing into the pressurized flotation tank from a coagulation tank which is the upstream end of the pressurized flotation tank; a camera module that generates image information by capturing the surface scum state and the bubble and floc shapes on the inner side of the pressurized flotation tank; and a water quality state prediction unit that, upon receiving the water quality measurement information from the sensor unit and the image information from the camera module, analyzes the correlation between the water quality measurement information and the image information through a previously stored artificial intelligence algorithm, while also reflecting real-time process state information in the analysis to predict the total phosphorus (TP) concentration for the effluent treated water of the pressurized flotation tank where measurement delay occurs. The apparatus includes an actuator control unit that, when the function of the water quality state prediction unit is completed, adjusts the amount of coagulant injected for coagulating pollutants in sewage and the operating parameters of the pressurized flotation tank so that the predicted total phosphorus concentration satisfies a preset target discharge water quality, and adjusts the amount of coagulant injected and the operating parameters of the pressurized flotation tank so that the ratio of the amount of coagulant injected to the amount of total phosphorus inflow is minimized; wherein the water quality state prediction unit includes a visual indicator extraction unit that, upon receiving the water quality measurement information from the sensor unit and the image information from the camera module, analyzes the image information through a preset artificial intelligence algorithm and extracts visual process indicators including the size and density of flocs in the pressurized flotation tank, the phase separation state between water and solids, and the rising speed of microbubbles through the analysis results.A multimodal artificial intelligence-based integrated control system for a pressurized flotation tank through image and water quality analysis, characterized by comprising: a pollution level derivation unit that, when the extraction of the visual process indicators is completed, analyzes the water quality measurement information through the previously stored artificial intelligence algorithm to derive the water quality pollution level based on the correlation between the electrical conductivity (EC) and turbidity changes of the incoming sewage through the analysis results; and a concentration information generation unit that, when the extraction of the visual process indicators and the derivation of the water quality pollution level are completed, analyzes the correlation between the real-time process status information, the visual process indicators, and the water quality pollution level through the previously stored artificial intelligence algorithm to time-sequence the total phosphorus concentration for the discharged treated water with a measurement delay to complete the generation of treated water concentration information. Claim 2 A pressurized flotation tank integrated control system through multimodal artificial intelligence-based image and water quality analysis, characterized in that, in claim 1, the water quality measurement information is information that includes detailed information generated by a plurality of sensors included in a sensor unit installed in the shear coagulation tank, and includes information such as the flow rate of wastewater from the shear coagulation tank flowing into the pressurized flotation tank, the turbidity of the incoming wastewater, the pH of the incoming wastewater, the electrical conductivity (EC) of the incoming wastewater, and the phosphorus (P) concentration of the incoming wastewater. Claim 3 delete Claim 4 In claim 1, the artificial intelligence algorithm stored above comprises: a first model trained to derive numerical feature values for the size and density of flocs in the pressurized flotation tank, visual boundary patterns due to layer separation of water and solids, and rising speeds due to time-series displacement changes of bubbles generated inside, using multiple image information and final visual process indicators corresponding to the multiple image information as learning information; a second model trained to derive correlation patterns between changes in the electrical conductivity (EC) and turbidity of sewage and water quality pollution levels using detailed information included in multiple water quality measurement information and final water quality pollution levels corresponding to the multiple water quality measurement information as learning information; and a third model that learns time-series prediction patterns of total phosphorus concentration for discharged treated water by analyzing integrated learning information generated as the feature values and correlation patterns are matched with multiple real-time process status information. Claim 5 In claim 4, the third model is characterized by generating the treated water concentration information in which the error due to environmental change is offset by applying the deviation in the reaction rate of the coagulant due to temperature change among the external environmental information as a correction weight of the time series prediction pattern when the external environmental information is received from a linked server and the received external environmental information matches the integrated learning information. Claim 6 In claim 1, the actuator control unit comprises: an input cost calculation unit that, when the generation of the treated water concentration information is completed, calculates the total phosphorus load of the inflowing sewage through the water quality measurement information and calculates a chemical input cost, which is the ratio of the currently injected coagulant input amount to the calculated total phosphorus load; and an operation effectiveness index calculation unit that, when the calculation of the chemical input cost is completed, compares and analyzes the calculated chemical input cost and pre-set input history information, and calculates the operation effectiveness index as a real-time monitoring indicator by comparing and analyzing based on a pre-set operation effectiveness index calculation formula. A multimodal artificial intelligence-based integrated control system for a pressurized flotation tank through image and water quality analysis, characterized by including: a feedback loop control unit that, when the calculation of the above-mentioned operational effect index is completed, adjusts the coagulant dosage and the operating parameters of the pressurized flotation tank so that the predicted total phosphorus concentration based on the above-mentioned treated water concentration information is located within the allowable water quality range based on the above-mentioned preset target effluent water quality, while adjusting the operating parameters of the pressurized flotation tank while reducing the coagulant dosage based on the above-mentioned operational effect index, which is a real-time monitoring indicator. Claim 7 The pressurized flotation tank integrated control system based on multimodal artificial intelligence image and water quality analysis according to claim 1, wherein the actuator control unit calculates a real-time inflow load amount through water quality measurement information received from the sensor unit while controlling a plurality of actuators by adjusting the coagulant input amount and the operating parameters of the pressurized flotation tank in real time; and a feedforward control unit that, when the calculated real-time inflow load amount exceeds a preset standard load range, calculates the design retention time of the upstream coagulant tank and applies the design retention time as a delay time to preemptively adjust the coagulant input amount in response to the time when the inflow load reaches the reaction zone within the pressurized flotation tank. Claim 8 In claim 1, the actuator control unit controls a plurality of actuators by adjusting the amount of coagulant injected and the operating parameters of the pressurized flotation tank in real time, and in a state where the ratio of air solids in the pressurized flotation tank is adjusted based on the rising speed and density of microbubbles through a visual process indicator extracted from a visual indicator extraction unit of a water quality state prediction unit, and the opening parameter adjustment unit adjusts a parameter corresponding to the opening of the air compressor valve among the operating parameters so as to adjust the ratio of air solids in the pressurized flotation tank; and when the predicted total phosphorus concentration reaches a threshold value of the allowable water quality range based on the preset target discharge water quality as the amount of coagulant injected is reduced, the RPM parameter adjustment unit adjusts a parameter corresponding to the scraper installed in the pressurized flotation tank among the operating parameters to accelerate the rotation speed of the scraper and improve the solid removal efficiency.
Citation Information
Patent Citations
Estuary sewage treatment system
CN116621305A
A Water Treatment System using Mixing Coagulation and Dissolved Air Floatation
KR100292428B1
A coagulative separation device of wastewater using ai-based vision analysis and the method thereof
KR102714108B1
