A method and system for adjusting the taste of direct drinking water in a pipeline

Abstract

The application discloses a kind of for pipe direct drinking water taste adjusting control method and system, for the business scene problem that water body quality is not consistent in the process of water treatment due to uneven distribution of adjusting agent, by real-time monitoring of sensor in pipe flow rate, pressure and pipe diameter change, obtain water flow mixing state parameter, using clustering algorithm to identify the area of large deviation of adjusting agent concentration, and when exceeding preset threshold, by injection device dynamic adjustment of dosing ratio, build optimized adjusting agent distribution model.The model combines water temperature change data, fusion dissolution efficiency index and seasonal environmental data, through prediction model to generate dynamic adjustment scheme, ensure system stability, finally through cyclic iteration optimization whole system taste consistency index, determine mixed diffusion process control parameter.The application improves water body quality uniformity and system operation stability through real-time monitoring, data fusion and dynamic adjustment, realizes the automation control of water treatment process.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of water treatment, in particular to a water quality adjusting and controlling method for a direct drinking water system and a corresponding system. BACKGROUND

[0002] With the development of social economy and the improvement of people's living standards, the public's demand for drinking water has shifted from basic safety to higher pursuit of taste, health and convenience. Direct drinking water systems can provide high-quality drinking water for immediate use, and are increasingly used in high-end residential areas, commercial complexes, public service facilities and other scenarios. An ideal direct drinking water system should not only ensure pure water quality and meet health standards, but also ensure that users at every water intake point in the pipe network can enjoy the experience of uniform quality and refreshing taste.

[0003] However, in existing direct drinking water systems, maintaining consistent taste throughout the pipe network faces technical challenges. Most existing technical solutions focus on the deep purification of water quality, such as removing impurities, residual chlorine and harmful substances through reverse osmosis, ultrafiltration, activated carbon adsorption, etc. However, in the subsequent taste adjustment process, a relatively static and extensive management approach is often used. Typically, the system will add taste adjusters such as minerals and electrolytes at the total water outlet in a fixed proportion at one time, expecting them to naturally mix evenly during subsequent pipe transportation. In addition, existing systems have poor adaptability to water temperature changes and are difficult to dynamically adjust according to external environments or user needs, resulting in unstable drinking experiences. The mixing and diffusion process of water flow in the pipe is difficult to control accurately, and changes in flow rate, pressure and pipe diameter in the pipe can lead to uneven distribution of taste adjusters such as minerals or flavoring ingredients. For example, in long-distance pipes, the taste near the water source may be rich, while the taste at the far end may be weak or even have an odor due to insufficient mixing. This unevenness is due to the complex influence of fluid mechanics on the diffusion of taste adjusters, making it difficult to achieve consistency throughout the system through simple static adjustment.

[0004] This static adjustment approach ignores the complex dynamic characteristics of the pipe system. First, direct drinking water pipe networks are usually long, have many branches, and have complex topologies. The motion state of water flow in the pipe is governed by fluid mechanics, and its flow rate and pressure will change dynamically in real time with changes in user water consumption, such as differences between peak and off-peak periods, pipe bends, diameter changes, etc. These changes directly lead to complex and uneven mixing and diffusion of taste adjusters in the water body. For example, at the end of a long-distance pipe or in a branch pipe network with low water flow velocity, the adjuster may be insufficiently mixed and have a low concentration, resulting in a weak taste; while in areas close to the injection point or areas with intense water flow disturbances, local concentrations may be too high.

[0005] Secondly, water temperature is another key variable affecting taste. On the one hand, water temperature directly affects human perception of taste. At different temperatures, the taste threshold of minerals in water is different. On the other hand, water temperature deeply affects the physical and chemical properties of the regulator in water, such as solubility and diffusion coefficient. In summer, the solubility and diffusion of the regulator are accelerated with the increase of water temperature, while in winter, the solubility and diffusion of the regulator are reduced with the decrease of water temperature. The existing system generally lacks dynamic perception and adaptive capacity for water temperature changes, and the fixed dosing strategy is difficult to cope with the water temperature fluctuations caused by seasonal changes or day and night temperature differences, further exacerbating the instability and inconsistency of taste.

[0006] In summary, the existing pipeline direct drinking water technology has obvious technical bottlenecks in taste control: first, it lacks real-time perception and analysis capability of the dynamic environment of water flow in the pipeline, and cannot grasp the actual distribution state of the regulator in the pipe network; second, the control strategy is fixed, and cannot be dynamically and prospectively adjusted according to the changes of flow rate, pressure, water temperature and other key factors; third, it lacks effective closed-loop feedback and optimization mechanism, and it is difficult to quantitatively evaluate and continuously improve the taste consistency of the whole pipe network. Therefore, it is urgent to develop an intelligent taste control method that can dynamically adapt to the changes of pipe network environment and realize accurate dosing and uniform mixing of the regulator, to ensure that the drinking water experience of all user terminals reaches a high-consistency high-quality standard, which has become a problem to be solved in the field. SUMMARY

[0007] The main purpose of the present application is to provide a taste control method and system for pipeline direct drinking water, aiming to solve the technical problem of inconsistent taste of different water intake points caused by uneven distribution of taste regulator due to dynamic changes of flow rate, pressure, water temperature and other working conditions in the pipeline.

[0008] In order to achieve the above purpose, the present application provides a taste control method for pipeline direct drinking water, comprising the following steps:

[0009] To achieve the above object, the method provided by the application comprises the following steps: firstly, real-time dynamic parameters of water flow in a pipeline are acquired, the parameters comprising variation data of one or more of flow rate, pressure and pipe diameter, and a water flow mixing state parameter representing an initial distribution state of a taste adjusting agent is generated based on the parameters. Then, the distribution of the adjusting agent is grouped and processed by using a preset clustering algorithm according to the water flow mixing state parameter, so as to identify and determine a distribution uneven area in which a concentration deviation of the adjusting agent exceeds a preset range. Then, it is judged whether the concentration deviation of the distribution uneven area exceeds a preset concentration deviation threshold value, and if so, the dosing proportion of the taste adjusting agent is adjusted by using a water treatment injection device, and an optimized adjusting agent distribution model is generated. On this basis, a dissolution efficiency index is acquired based on the optimized adjusting agent distribution model, and fusion processing is performed in combination with real-time water temperature variation data, so as to determine a diffusion efficiency correction value reflecting the influence of temperature. Then, the diffusion efficiency correction value and collected seasonal environmental data are analyzed by using a prediction model, so as to generate a dynamic adjustment scheme. Before the scheme is applied, it is judged whether the scheme satisfies a preset system stability requirement, and if so, the dosing proportion of the adjusting agent is updated by using the scheme, and a whole-system taste consistency index is calculated. Finally, real-time feedback data are acquired according to the whole-system taste consistency index, the index is optimized by using a cyclic iteration mode, so as to determine final mixing and diffusion process control parameters for guiding continuous operation of a water treatment device.

[0010] Correspondingly, the application further provides a taste adjusting control system for pipeline direct drinking water, which is designed to execute the above method. The system comprises:

[0011] A data acquisition module is configured to acquire real-time dynamic parameters of water flow in a pipeline, the real-time dynamic parameters comprising at least variation data of one or more of flow rate, pressure and pipe diameter;

[0012] An analysis module is connected with the data acquisition module and is configured to determine an initial distribution state of a taste adjusting agent in water based on the real-time dynamic parameters, to generate a water flow mixing state parameter, and to identify and determine a distribution uneven area according to the water flow mixing state parameter by using a preset clustering algorithm;

[0013] A first adjusting module is connected with the analysis module and is configured to generate an adjusting instruction to adjust a dosing proportion of a taste adjusting agent when a concentration deviation of the distribution uneven area exceeds a preset concentration deviation threshold value, and to construct an optimized adjusting agent distribution model;

[0014] A correction module is connected with the first adjusting module and is configured to acquire a dissolution efficiency index based on the optimized adjusting agent distribution model, and to calculate a diffusion efficiency correction value in combination with real-time water temperature data;

[0015] A prediction module, connected with the correction module, is configured to analyze the diffusion efficiency correction value and seasonal environmental data through a prediction model to generate a dynamic adjustment scheme;

[0016] An execution and update module, connected with the prediction module, is configured to execute the scheme to update the dosing ratio of the conditioning agent and calculate a full-system taste consistency index when the dynamic adjustment scheme meets the system stability requirement;

[0017] An iterative optimization module, connected with the execution and update module, is configured to determine final mixed diffusion process control parameters through a cyclic iteration mode based on real-time feedback of the full-system taste consistency index and send the parameters to the water treatment equipment.

[0018] To sum up, the application discloses a taste adjustment control method and system for pipeline direct drinking water, and a full-closed-loop dynamic control system from real-time sensing, intelligent diagnosis to predictive adjustment and continuous optimization is constructed.

[0019] Compared with the prior art, the application has the following beneficial effects: first, the application can respond to changes in the flow rate, pressure and other hydraulic conditions in the pipeline in real time, dynamically adjust the dosing ratio of the conditioning agent, effectively overcome the limitations of the static dosing strategy in the prior art that cannot adapt to fluctuations in working conditions, and improve the uniformity level of the distribution of the conditioning agent in the complex pipe network, thereby ensuring the consistency of the taste at different water intake points. Second, by introducing the correction of the water temperature on the diffusion efficiency and the prediction model based on seasonal environmental data, the application not only can adapt to the current environmental changes, but also can predict and adjust in advance the changes that may occur subsequently, so that the system shows better robustness when facing seasonal changes or climate mutations, and maintains the long-term stability of the taste. Finally, the application defines the full-system taste consistency index to provide an objective and quantitative evaluation standard for the adjustment effect, and the cyclic iteration optimization process based on the index can make the running parameters of the system converge to the optimal solution continuously, which helps to optimize the running energy consumption of the equipment and the dosing cost of the conditioning agent under the premise of ensuring the taste. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 A whole flowchart of a taste adjustment control method for pipeline direct drinking water provided by the embodiment of the application is shown.

[0021] Figure 2 The internal detailed flowchart of the step (S105) of generating a dynamic adjustment scheme in the embodiment of the present application is shown.

[0022] Figure 3 The internal detailed flowchart of the step (S106) of judging system stability and updating the dosing ratio in the embodiment of the present application is shown.

[0023] Figure 4 The internal detailed flowchart of the step (S107) of iterative optimization in the embodiment of the present application is shown.

[0024] Figure 5 The functional module structure block diagram of the taste adjusting control system for the pipeline direct drinking water provided by the embodiment of the present application is shown.

[0025] Figure 6 The technical effect diagram of the embodiment of the present application in a specific application scenario is shown.

[0026] Figure 7 The technical effect diagram of the embodiment of the present application relative to the prior art is shown. DETAILED DESCRIPTION

[0027] The technical solutions in the embodiments of the present application will be clearly and completely described in the following with reference to the drawings. Figures 1 to 7 It should be noted that the technical solutions in the embodiments of the present application are described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0028] It should be noted that the “connection”, “connection” and the like mentioned in the embodiments of the present application can refer to the physical direct connection or the logical indirect connection. The terms “first”, “second” are only for the purpose of description, and cannot be understood as indicating or implying the relative importance or implicitly indicating the number of the indicated technical features.

[0029] Embodiment one

[0030] The embodiment details the specific steps of the taste adjusting control method for the pipeline direct drinking water. Please refer to Figure 1 The figure shows the overall flow of the method of the present application.

[0031] Step S101: Obtain the change data of the flow rate, pressure and pipe diameter in the pipeline, and monitor the changes in real time through the water treatment system sensor to obtain the water flow mixing state parameters, which represent the initial distribution characteristics of the adjusting agent in the water body.

[0032] In this step, a comprehensive and real-time perception of the physical state of the pipe system is required. In one specific embodiment, to achieve this goal, a sensor network composed of various high-performance sensors can be pre-deployed on key nodes of the direct drinking water pipe network. These nodes usually include the outlet of the main pipe, the branches of the main pipe, the relay points of long-distance conveying pipes, and representative user ends.

[0033] The selection of sensor types is targeted. For example, flow rate data can be obtained by non-contact ultrasonic flow meters or electromagnetic flow meters, which can accurately measure the average flow rate of the pipe cross section without invading the water flow. Pressure data is collected by piezoresistive or capacitive pressure sensors installed on the pipe wall, which reflects the real-time fluctuations of the water pressure in the pipe. For changes in pipe diameter, although the pipe material is fixed in most cases, thermal expansion and contraction caused by temperature changes or small deformations under long-term water pressure can be monitored by high-precision laser displacement sensors or strain gauges. In addition, platinum resistance temperature sensors, such as PT100, are used to accurately measure water temperature. All sensors have high sampling frequency, such as 1 to 10 data collected per second, and high accuracy to ensure that transient operating condition fluctuations caused by changes in user water consumption behavior can be captured.

[0034] The collected raw data is transmitted in real time to the central processing unit through wired (such as Modbus bus) or wireless (such as LoRa, NB-IoT) communication. The central processing unit first preprocesses the data, such as filtering and denoising, data alignment, etc.

[0035] The obtained water flow mixing state parameter is not a single value, but a set of characteristic vectors that can comprehensively describe the mixing state of the adjusting agent. Preferably, the parameter is a multi-dimensional vector. Its generation process is as follows:

[0036] Correlation analysis and trend extraction, first analyze the correlation between flow rate change data and pressure change data. For example, by calculating the covariance of the two, the trend of small changes in pipe diameter caused by hydraulic action can be indirectly inferred. Because at a certain flow rate, a sudden increase in pressure may indicate a contraction or blockage of the downstream pipe diameter.

[0037] Preliminary modeling of initial concentration distribution, according to the initial dosing concentration of the adjusting agent injection point and the main pipe flow rate, a basic one-dimensional convection-diffusion model is established to preliminarily estimate the concentration distribution of the adjusting agent along the main flow direction. At this time, temperature data is introduced for preliminary correction, because temperature will affect the viscosity of water, and then affect the flow state. For example, using the weighted average method, the pipe diameter change trend and the temperature correction factor are fused to obtain a more accurate initial concentration distribution curve.

[0038] Hydrodynamic refinement, in order to get more accurate local mixing state, more complex hydrodynamic model is used, such as Navier-Stokes equation model. The model is a partial differential equation set describing the motion of viscous incompressible fluid. The system takes the initial concentration distribution obtained by the above preliminary modeling and the real-time collected three-dimensional flow field data as the input boundary conditions of the model. By numerical calculation method, such as finite element method or finite volume method, the velocity field and concentration field of any point in the pipeline can be obtained. From the concentration field, the key diffusion rate value, such as the modulus of concentration gradient, can be calculated.

[0039] Finally, the water flow mixing state parameter is constituted by multiple dimensions of data such as initial concentration distribution and diffusion rate value. The parameter can comprehensively and dynamically represent the initial mixing and distribution characteristics of the adjusting agent under the influence of the current hydraulic working condition after being injected into the water body.

[0040] Step S102: According to the water flow mixing state parameter, a preset clustering algorithm is used to analyze the distribution of the adjusting agent, and a distribution uneven area is determined, which represents a pipeline part with large adjusting agent concentration deviation.

[0041] After obtaining the quantified water flow mixing state parameter, the next step is to intelligently evaluate the mixing uniformity of the entire pipe network and accurately locate the problem area. This step uses the clustering algorithm in unsupervised learning method in machine learning to achieve.

[0042] Specifically, the system will collect and calculate the water flow mixing state parameter from each monitoring point in the pipe network to form a large data set. Each data point in the data set represents the mixing state of a specific position on the pipeline at a certain time. Then, the system calls the preset clustering algorithm to process the data set.

[0043] Optionally, the K-means clustering algorithm is used to analyze the data set. The workflow is as follows:

[0044] 1. Determine the number of clusters (K): K value can be preset according to experience, for example, the mixing state of the pipe network is divided into "excellent, good, medium, and poor" four levels, then K=4. The elbow rule and other methods can also be used to adaptively determine the best K value.

[0045] 2. Initialize the centroid: randomly select K data points as the initial clustering center (centroid).

[0046] 3. Assign data points: calculate the distance of each data point in the data set to the K centroids, preferably, use Euclidean distance, and assign each data point to the cluster where the nearest centroid is located.

[0047] 4. Update the centroids: Recalculate the average of all data points in each cluster and use this average as the new centroid.

[0048] 5. Iteration: Repeat steps 3 and 4 until the positions of the centroids no longer change significantly, or a preset number of iterations is reached, at which point the clustering is complete.

[0049] After clustering is complete, all monitoring points on the pipe network are divided into K clusters. Each cluster represents a specific mixing state. The system then calculates the average concentration deviation for each cluster, which is the difference between the average concentration of all data points in the cluster and the overall average concentration of the entire pipe network. If the absolute value of the average concentration deviation for a cluster exceeds a preset threshold (for example, more than 10% of the overall average concentration), then the pipe section where all monitoring points in that cluster are located is identified and marked as a non-uniform distribution area by the system.

[0050] Preferably, to enhance the robustness of the method, especially in complex water flow environments containing noise or abnormal data points, other clustering algorithms such as DBSCAN (Density-Based Spatial Clustering Application) can also be used. DBSCAN does not require the number of clusters to be specified in advance, and can well identify clusters of arbitrary shape and distinguish outliers (noise) within them, thus more accurately locating areas where there are real systemic mixing problems.

[0051] Through this step, the originally complex and continuous concentration distribution problem is transformed into the identification and location of a few discrete problem areas, laying the foundation for subsequent precise adjustment.

[0052] Step S103: Determine whether the non-uniform distribution area exceeds a preset concentration deviation threshold. If it does, adjust the dosing ratio of the conditioning agent through the water treatment injection device to obtain an optimized conditioning agent distribution model, which represents a uniform conditioning agent distribution state.

[0053] After identifying the non-uniform distribution area, the decision and execution phase is entered. First, the deviation degree is judged to check whether the maximum concentration deviation value calculated in S102 exceeds a strict, preset concentration deviation threshold. This threshold is set according to the drinking water taste standards and system operation requirements, for example, requiring that the concentration deviation of any point not exceed 5%.

[0054] If the judgment result is yes, i.e., the deviation is out of specification, the system will immediately start the adjustment program. The core executive mechanism of the adjustment is the water treatment injection device, which is usually composed of high-precision metering pumps, electric control valves, and possibly static or dynamic mixers. The adjustment process follows a closed-loop feedback control logic, for example, using a classic proportional-integral-derivative (PID) controller.

[0055] The specific adjustment process is as follows:

[0056] 1. Calculate the adjustment amount: The PID controller takes the current concentration deviation value as input. The proportional term (P) directly calculates a basic adjustment amount according to the size of the current deviation; the integral term (I) is used to eliminate static error, and if the deviation persists, it will accumulate error and gradually enhance the adjustment; the derivative term (D) anticipates the trend of deviation change, plays a role in advance inhibition and prevents overshoot. The controller synthesizes the outputs of the three terms to calculate the precise dosage change amount (increase or decrease) required for adjustment.

[0057] 2. Execute the adjustment: The central processing unit sends the calculated adjustment instructions to the controller of the injection device (such as PLC). The PLC converts the instructions into specific control signals for the frequency of the metering pump or the opening degree of the valve, thereby accurately changing the injection rate of the adjusting agent.

[0058] 3. Build an optimized model: While executing the adjustment, the system will perform a simulation calculation again based on the new dosage ratio and the current fluid mechanics model. This calculation aims to simulate the distribution of the adjusted adjusting agent in the pipe network, and the result constitutes the optimized adjusting agent distribution model. This model can be a grid-based concentration matrix in form, where each grid point represents the predicted concentration value of a small section of the pipeline. This model is a theoretical prediction of the system's adjustment effect, representing the ideal uniform distribution state.

[0059] This step uses mature control theory for quantitative adjustment, and through modeling simulation, it pre-evaluates the adjustment effect, ensuring the effectiveness and accuracy of the adjustment.

[0060] Step S104: Obtain the dissolution efficiency index from the optimized adjusting agent distribution model, combine the water temperature change data for fusion processing, and determine the diffusion efficiency correction value, which reflects the influence of temperature on the diffusion of the adjusting agent.

[0061] After completing the preliminary adjustment based on the hydraulic working condition, this method introduces the consideration of water temperature, a key influencing factor. This step aims to quantify the influence of temperature on the dissolution and diffusion process of the adjusting agent in water and generate a correction parameter for subsequent more detailed adjustment.

[0062] First, the system extracts a solubility efficiency indicator from the optimized conditioner distribution model, i.e., the concentration matrix generated in S103. This indicator quantifies how quickly the conditioner transitions from a concentrated state upon injection to a uniformly dissolved state. A simple yet effective calculation method is to analyze the rate of change of the concentration gradient in the model. Downstream from the injection point, if the concentration rapidly decreases from an extremely high value to a stable state close to the average value, it indicates high solubility diffusion efficiency; conversely, if there is severe concentration tailing, the efficiency is low. The system can calculate the norm of the concentration gradient along the main flow direction and integrate or average it to obtain a quantitative solubility efficiency indicator.

[0063] Next, this indicator is fused with real-time monitored water temperature data. The physical basis of this process is that temperature rise generally accelerates molecular Brownian motion, thereby increasing the diffusion coefficient, which can be referred to as the Arrhenius equation. The fusion process can be done in the following ways:

[0064] 1. Establish a baseline model: The system pre-stores a model of the relationship between the solubility efficiency indicator and the diffusion efficiency at a standard temperature, such as 25°C, which can be a lookup table or a functional expression.

[0065] 2. Calculate the temperature influence factor: Calculate a temperature influence factor based on the difference between the current real-time water temperature and the standard temperature. For example, a linear or exponential model can be established to describe the percentage increase in diffusion efficiency for every 1°C increase in temperature.

[0066] 3. Weighted fusion: Use weighted averaging to fuse the solubility efficiency indicator calculated based on the model and the influence factor calculated based on the temperature. For example, the final diffusion efficiency = α * solubility efficiency indicator + (1-α) * temperature influence factor, where the weight α can be optimized based on historical data.

[0067] 4. Generate correction value: Compare the current actual diffusion efficiency obtained after fusion processing with the baseline diffusion efficiency at standard temperature. The difference or ratio between the two is the diffusion efficiency correction value. If the current water temperature is high, the correction value is usually positive, indicating that diffusion is promoted; otherwise, it is negative.

[0068] Through this step, this method not only considers the macroscopic mixing effect of the water flow, but also incorporates the influence of temperature in the form of a quantitative correction value into the control model, greatly improving the accuracy of the model.

[0069] Step S105: Analyze the diffusion efficiency correction value and seasonal environmental data through a prediction model to obtain a dynamic adjustment scheme that integrates environmental factors to optimize the water treatment process.

[0070] Please refer to Figure 2This figure details the internal flow of this step. In order to make the control strategy not only cope with the current situation, but also adapt to future environmental changes, this step introduces a machine learning prediction model to achieve dynamic adjustment.

[0071] 1. Data collection and preparation Figure 2 In the "acquire seasonal environmental data set", "acquire diffusion efficiency correction value set"), seasonal environmental data such as daily average temperature, humidity, air pressure, etc. will be continuously collected from internal sensors (water temperature) and external meteorological service interfaces. At the same time, the system will store the diffusion efficiency correction values calculated in S104 during the historical operation process. These data together constitute the data set required for training the prediction model.

[0072] 2. Model training: A prediction model is used, such as linear regression, ARIMA time series model, or even more complex neural network model. Taking linear regression as an example, the model will use historical seasonal environmental data such as temperature and humidity as independent variables (input features), and the corresponding historical diffusion efficiency correction values as dependent variables (output labels). By training the historical data set, the model can learn the potential relationship between external environmental factors and water body internal diffusion efficiency, and solidify it in the form of a set of regression coefficients.

[0073] 3. Prediction and scheme generation Figure 2 In the "generate dynamic adjustment scheme"), the latest collected external environmental data, such as 24-hour weather forecast and the currently calculated diffusion efficiency correction value, are input into the trained prediction model. The model will predict the possible trend of diffusion efficiency correction value in the future.

[0074] Based on this prediction result, the system will generate a dynamic adjustment scheme, which is no longer a simple increase / decrease of the dosage ratio, but a more complex instruction set, such as: "In the next 6 hours, due to the expected temperature drop of 5℃, it is recommended to pre-adjust the dosage ratio by 2% and increase the mixing intensity of the downstream pipe mixer by one level to compensate for the diffusion slowdown caused by low temperature".

[0075] This step introduces a prediction model, which can layout in advance and smoothly cope with environmental changes, avoiding frequent and drastic lagging adjustment, thereby greatly improving the stability and economic efficiency of the system.

[0076] Step S106: Determine whether the dynamic adjustment scheme meets the preset system stability requirements. If it meets the requirements, update the dosage ratio of the adjustment agent according to the scheme to obtain the taste consistency index of the whole system, which represents the uniformity of the overall quality of the water body.

[0077] Please refer to Figure 3This diagram illustrates the internal process of this step in detail. After obtaining a dynamic adjustment plan, it cannot be implemented immediately; a rigorous safety and stability assessment must be conducted.

[0078] 1. Stability requirements preset ( Figure 3 The "Preset System Stability Requirements" state that the system pre-sets a series of stable operating ranges for key performance indicators (KPIs), which are the system stability requirements. These requirements are the red lines to ensure the safe operation of the water supply system. For example, the fluctuation range of pipeline pressure must not exceed ±0.05MPa; the change rate of water flow rate must not exceed 10%; and the fluctuation range of pH value must not exceed ±0.2, etc.

[0079] 2. Comparison and Judgment of Solutions ( Figure 3 (In the "Judging whether the system stability requirements are met"): The system will perform a rapid simulation analysis on the dynamic adjustment scheme generated by S105 to predict the possible changes in the above key performance indicators after the scheme is implemented. Then, the predicted changes will be compared with the preset stability requirements.

[0080] 3. Decision-making and execution ( Figure 3 (In the "update ratio if satisfied"): Only when the predicted changes of all KPIs are within the allowable range, and the result is "meets requirements," is the dynamic adjustment plan considered safe and feasible. At this point, the system will formally adopt the plan and issue its instructions (such as the new addition ratio) to the implementing agency. The new addition ratio is usually obtained by multiplying the original ratio by the dynamic adjustment value in the plan (e.g., ...). Figure 3 (The new ratio = the original ratio × the dynamic value).

[0081] 4. Quantitative evaluation of results ( Figure 3 The "Calculation of System-wide Taste Consistency Index" is as follows: After implementing a new dosage ratio, the system needs to conduct a global quantitative evaluation of the final adjustment effect. To this end, the system introduces the concept of a system-wide taste consistency index. This index is calculated by collecting real-time data of key sensory parameters from all monitoring points in the pipeline network, such as turbidity, conductivity, and pH value, and then calculating the standard deviation or coefficient of variation of these data points. The smaller this statistical value, the closer the water quality parameters are to each other across the entire pipeline network, the higher the uniformity, and the better the taste consistency.

[0082] This step ensures that all optimization and adjustment actions are carried out without sacrificing the safety and stability of the system. At the same time, it provides a clear and quantifiable evaluation target for subsequent closed-loop optimization.

[0083] Step S107: According to the full-system taste consistency index, real-time feedback data is obtained, and the index is optimized through a cyclic iteration method to determine the final mixing and diffusion process control parameters, which are used to guide the continuous adjustment of the water treatment equipment.

[0084] Figure 4 The internal detailed flowchart of the cyclic iteration optimization step (S107) in the embodiment of the application is shown. The full-system taste consistency index calculated in S106 is taken as the objective function, and the optimization goal is to minimize the index. The flowchart is started with the input of the obtained full-system taste consistency index. First, a decision is made in the judgment block whether the index converges or reaches the preset iteration number. If the answer is yes, the flowchart ends, and the final control parameters are determined and output. If the answer is no, the optimization cycle is entered, the gradient of the objective function is first calculated, and then the mixing and diffusion process control parameters are updated (for example, through the formula "new parameter = old parameter - learning rate x gradient"). The updated new parameters are applied to the system to produce a new running state and a new taste consistency index, which is taken as feedback to enter the starting point of the flowchart again, forming an iterative process of closed-loop optimization.

[0085] Specifically, after the current full-system taste consistency index is calculated in S106, the value is taken as real-time feedback of system performance and input into the iterative optimizer of this step. The goal of the optimizer is to fine-tune a series of underlying mixing and diffusion process control parameters to minimize the taste consistency index, for example, the TDS standard deviation.

[0086] Preferably, an efficient iterative optimization algorithm, the gradient descent method, is used. The working principle is as follows:

[0087] 1. Define the objective function: define the taste consistency index, for example, the standard deviation σ of the conductivity at each point, as the objective function J. The optimization goal is to minimize J.

[0088] 2. Determine the adjustable parameters: determine a group of device operating parameters that can directly affect the mixing and diffusion process, for example, the operating frequency (P_freq) of the injection device metering pump and the stirring speed (M_speed) of the pipeline mixer. These are the variables that need to be optimized by the gradient descent method.

[0089] 3. Calculate the gradient: at the current operating parameter point, a small perturbation is applied, for example, P_freq is increased by 0.1%, and the change in the objective function J is observed. Thus, the partial derivative of the objective function J with respect to each adjustable parameter, i.e., the gradient, can be calculated. The gradient direction is the direction in which the objective function grows fastest.

[0090] 4. Update parameters: update the parameters by a small step in the opposite direction of the gradient. The update formula is: new parameter = old parameter - learning rate x gradient. The learning rate is a hyperparameter that controls the step size of each update.

[0091] 5. Loop iteration: repeat steps 3 and 4. In each iteration, the parameters move towards the direction that can reduce the objective function J. After multiple iterations, when the value of the objective function J converges (no longer significantly reduced) or reaches the preset number of iterations, the optimization process ends.

[0092] At this time, the set of (P_freq, M_speed,...) values output by the optimizer is the final mixing and diffusion process control parameter that can achieve the best taste consistency under the current macroscopic working condition. These parameters will be locked and used to guide the continuous and stable operation of the related water treatment equipment in the next control period.

[0093] Through this loop iteration and continuous optimization process, the method of the present application finally forms an intelligent closed-loop control system that can adaptively, self-learn, and continuously approach the optimal state.

[0094] As can be seen from the embodiment, the method first acquires various dynamic parameters of the water flow in the pipeline in real time through the sensor network installed at each key node of the pipeline. These parameters at least cover the flow rate, pressure, and possible pipe diameter change data. Based on these first-hand data, the system can construct a mathematical model that can represent the initial distribution state of the taste regulator after being injected into the water body, and generate a set of quantitative water flow mixing state parameters.

[0095] Then, the method uses advanced data analysis techniques, specifically a preset clustering algorithm, to intelligently group the water flow mixing state parameters generated above. In this way, pipeline sections with similar regulator concentration distributions can be effectively classified into one category, thereby quickly and accurately identifying those distribution uneven areas with a regulator concentration deviating too much from the overall average level.

[0096] After identifying the uneven distribution areas, the method enters the judgment and adjustment stage, and compares the actual concentration deviation of these areas with a pre-set, acceptable concentration deviation threshold. If the deviation exceeds the threshold, it means that the taste unevenness problem has reached the level that requires intervention, and the system will immediately activate the water treatment injection device to accurately adjust the dosage proportion of the taste regulator, which can be to increase or decrease the dosage. After adjustment, the system will rebuild an optimized regulator distribution model that theoretically represents a more uniform distribution state.

[0097] Considering the important influence of water temperature on the adjustment agent diffusion process, the method further introduces a temperature correction mechanism. The system extracts a key indicator, the dissolution efficiency indicator, from the optimized adjustment agent distribution model, and combines it with real-time monitored water temperature change data to process the indicator. The result of the processing is to generate a diffusion efficiency correction value that can scientifically quantify and reflect the promotion or inhibition of the current water temperature on the diffusion speed of the adjustment agent in the water.

[0098] To achieve more forward-looking and adaptive control, the method also introduces a prediction model. The model analyzes the diffusion efficiency correction value just calculated and the seasonal environmental data (such as air temperature, humidity, etc.) collected from the external environmental monitoring system. Through the operation of the machine learning algorithm, the prediction model can generate a dynamic adjustment scheme that not only considers the current working condition but also incorporates the prediction of future environmental changes, making the optimization of the water treatment process more forward-looking.

[0099] Before applying any new adjustment scheme, it is determined whether the execution of the dynamic adjustment scheme will cause the key operating parameters of the pipe network system (such as water pressure, flow rate) to fluctuate sharply, thereby affecting water supply safety. Only when it is confirmed that the scheme meets the preset system stability requirements, the system will adopt the scheme and update the actual dosage proportion of the adjustment agent accordingly. After updating, the system also calculates a global index that can quantitatively evaluate the uniformity of the overall water quality, i.e. the overall taste consistency index.

[0100] Finally, the method designs a cyclic iteration link that will continuously obtain real-time operation effect feedback data according to the latest calculated overall taste consistency index. Through a cyclic iteration algorithm, the index is continuously optimized and calculated to drive the fine tuning of the control parameters of the mixing and diffusion process (such as the frequency of the injection pump, the speed of the mixer, etc.), until the taste consistency index converges to the ideal target range. In this way, a set of optimal final control parameters for guiding the long-term, stable and efficient operation of the water treatment equipment is determined.

[0101] Example Two

[0102] This embodiment provides a taste adjustment control system for direct drinking water in pipes. Please refer to Figure 4 , which is designed to implement the method described in Example One. The system includes the following modules:

[0103] Data acquisition module 10, which provides real-time and accurate pipe network physical state data for the entire system, is responsible for obtaining original physical quantity signals from various sensors deployed in the pipe system and converting them into digital signals for processing by upper-level modules.

[0104] Specifically, the implementation of the data acquisition module 10 is to integrate a sensor network and a data acquisition terminal. The sensor network includes but is not limited to: a non-contact ultrasonic flowmeter 11 installed on the outside of the pipe wall for measuring the fluid velocity without disturbance; a diffusion silicon pressure sensor 12 embedded in the pipe wall for accurately measuring the water pressure; a high-precision laser displacement sensor 13 for monitoring the slight change of the pipe radius caused by temperature or pressure change; and an armored PT100 platinum resistance temperature sensor 14 for measuring the water temperature. The data acquisition terminal can be a programmable logic controller (PLC) or a dedicated remote terminal unit (RTU) with multiple input / output (I / O) interfaces. It digitizes the analog signals (such as 4-20mA) output by the sensors through an A / D converter, and transmits the data to the analysis module 20 through an industrial Ethernet or RS485 bus.

[0105] The analysis module 20 is connected with the data acquisition module. The function of this module is to determine the initial distribution state of the mouthfeel regulator in the water body based on the real-time dynamic parameters, generate water flow mixing state parameters, and identify and determine the uneven distribution area according to the water flow mixing state parameters using a preset clustering algorithm.

[0106] Specifically, the implementation of the analysis module 20 is a software package running on a main controller (such as an industrial PC). The software package contains: a state parameter calculation engine that receives real-time data streams and calls an embedded, simplified computational fluid dynamics (CFD) solver or an interface linked with an external professional CFD software (such as ANSYS Fluent) to calculate and generate water flow mixing state parameters; a clustering analysis engine that can integrate open-source machine learning libraries (such as Scikit-learn or TensorFlow) and implement K-means or DBSCAN algorithms. It takes the state parameters as input, performs clustering calculation, and sends the output cluster labels (representing different mixing state areas) and corresponding deviation data to the first adjustment module 30.

[0107] The first adjustment module 30 is connected with the analysis module. The function of this module is to generate adjustment instructions to adjust the dosage proportion of the mouthfeel regulator when the concentration deviation in the uneven distribution area exceeds the preset concentration deviation threshold, and to construct an optimized regulator distribution model.

[0108] Specifically, the implementation of the first adjustment module 30 is a software logic unit, which internally contains: a deviation judge that compares the input deviation data with the concentration deviation threshold stored in the configuration database; a PID controller algorithm implementation that calculates the precise dosage adjustment amount according to the P, I, and D parameters (which can be self-tuned according to historical data) when the deviation exceeds the threshold; and a model updater that generates an optimized adjustment agent distribution model by calling the fluid dynamics model again based on the adjusted dosage ratio, and passes the model data structure to the correction module 40.

[0109] The correction module 40 is connected with the first adjustment module, and the function of this module is to specifically handle the influence of water temperature on the adjustment agent diffusion process, and to provide more accurate input for subsequent prediction and control by calculating a quantitative correction value.

[0110] Specifically, the implementation of the correction module 40 is a function library or an independent calculation service. It contains: an index extraction function that calculates the spatial variation rate of concentration gradient from the input distribution model matrix to obtain the dissolution efficiency index; a fusion calculation function that internally contains the Arrhenius equation or empirical fitting curve describing the relationship between temperature and diffusion coefficient, which takes the dissolution efficiency index and real-time water temperature data as input, and outputs the final diffusion efficiency correction value through table lookup or formula calculation.

[0111] The prediction module 50 is connected with the correction module, and the function of this module is to generate a dynamic adjustment scheme by analyzing the diffusion efficiency correction value and seasonal environmental data through a prediction model.

[0112] Specifically, the implementation of the prediction module 50 is a software service that integrates model loading and inference functions. It regularly obtains weather forecast data from external meteorological data services through an API interface. It internally loads a pre-offline trained prediction model 51 (for example, a model trained using the gradient boosting tree algorithm such as XGBoost), which can receive diffusion efficiency correction values and multi-dimensional environmental features, and output a prediction sequence of future diffusion efficiency. The module also contains a scheme generator that translates the prediction sequence into a specific dynamic adjustment scheme containing multiple parameter adjustment suggestions based on a set of "if-then" rules.

[0113] The execution and update module 60 is connected with the prediction module, and the function of this module is to serve as a bridge between the decision layer and the physical execution layer, for executing the dynamic adjustment scheme to update the dosage ratio of the adjustment agent when the scheme meets the system stability requirements, and calculating the overall system taste consistency index.

[0114] Specifically, the implementation of the execution and update module 60 is a core logic scheduling unit in the control system. It contains a stability checker that performs a quick simulation before executing any scheme and compares the simulation results with the stored system stability requirements 61 (a set of safety thresholds for key parameters). After the check, the instruction executor converts the adjustment instructions (such as new dosing ratios) in the scheme into control signals that the underlying hardware can recognize, such as PWM signals for frequency converters or on-off signals for electromagnetic valves, and sends them to the water treatment injection device 70 through the I / O interface. Then, the index calculator collects network data, calculates and publishes the latest full-system taste consistency index.

[0115] The iterative optimization module 80 is connected to the execution and update module. The function of this module is to determine the final mixing and diffusion process control parameters through iterative optimization based on real-time feedback of the full-system taste consistency index, and send them to the water treatment equipment.

[0116] Specifically, the implementation of the iterative optimization module 80 is a background running optimization service. This service monitors the taste consistency index published by the execution and update module 60 and uses it as the objective function of the optimization algorithm 81 (e.g., the gradient-based L-BFGS algorithm or the gradient-free particle swarm optimization algorithm). The optimization algorithm runs in a separate computing thread, fine-tuning a set of underlying control parameters through iterative calculations. When the algorithm converges, i.e., the optimal solution is found, it updates the optimal parameters to the running configuration of the execution and update module 60 through the parameter server, thus completing a complete closed-loop optimization.

[0117] Example Three

[0118] This embodiment illustrates the application process of the method in a complete event cycle in conjunction with a specific and more detailed application scenario. This embodiment is applied to a high-end residential area with a centralized water treatment station and an independent direct drinking water pipe network covering the entire community. The total length of the pipe network is about 5 kilometers, with multiple branches, climbs, and bends. The flow rate difference between the peak period (7-9 am in the morning and 6-8 pm in the evening) and the flat period is huge. The community is located in the north, with a large temperature difference between summer and winter. The water temperature in the pipe network can reach 30°C in summer and approach 5°C in winter. The system needs to add composite mineral adjusters containing strontium and metasilicic acid to improve the taste and health value of drinking water. The goal is to stabilize the TDS (total dissolved solids) value of all users in the network within the range of 120±10 mg / L.

[0119] Scenario setting: Saturday morning at 9 am during a week of continuous high summer temperatures. Due to the hot weather and the weekend water peak, the community's water consumption reaches its peak. At this time, the method of the present invention is fully executed according to the following steps:

[0120] Step 1 (corresponding to S101): At 9:05 am, the data acquisition module 10 reports that the average flow rate across the network has surged from the usual 0.8 m / s to 2.2 m / s; the pressure sensor readings at the end of the pipe network (in the 25-30 high-rise area) have dropped from 0.35 MPa at the end to 0.25 MPa; the water temperature sensor at the outlet of the water treatment station shows a water temperature of 29°C. Based on these sudden changes in real-time dynamic parameters, the system inputs into the Navier-Stokes equation model for calculation to generate the current water flow mixing state parameters. Analysis of these parameters shows that under the high flow rate scour, the main body of the regulating agent is rapidly carried to the far end, but vortexes occur at multiple U-shaped bends and reducers, and in the climbing pipes at the very end of the pipe network, the diffusion rate is significantly lower than the average level of the pipe network due to the pressure drop and flow pattern changes.

[0121] Step 2 (corresponding to S102): The analysis module 20 receives the above state parameters and immediately executes the K-means clustering algorithm. The algorithm divides the data points of the hundreds of monitoring nodes across the network into 5 clusters. The calculation results show that the average TDS concentration prediction value of one cluster (Cluster4) is only 98 mg / L, which has a huge deviation of -18.3% from the target value of 120 mg / L. The map mapping shows that this cluster covers all high-rise users of buildings 28, 29, and 30, and the system thus determines the specific uneven distribution area.

[0122] Step 3 (corresponding to S103): The first adjustment module 30 determines that the absolute value of the -18.3% deviation has far exceeded the preset concentration deviation threshold of 8%. The PID controller immediately intervenes and calculates that the frequency of the metering pump of the injection device 70 needs to be increased by 25% according to the deviation size and the integral term (which indicates that the deviation has lasted for several minutes). After the instruction is issued, the system simultaneously re-performs simulation based on this new injection ratio to generate an optimized regulating agent distribution model, which shows that the end concentration is expected to return to the target range within 15 minutes.

[0123] Step 4 (corresponding to S104): The correction module 40 analyzes the concentration gradient changes from the above-optimized distribution model and extracts the dissolution efficiency index as 0.92 (with a benchmark value of 1.0). At the same time, the current water temperature of 29°C is obtained. According to the built-in Arrhenius temperature effect model, 29°C relative to the standard 25°C will increase the diffusion coefficient by about 9%. Combining these two factors, the system calculates the final diffusion efficiency correction value as +1.20, which is a positive value indicating that the current high-temperature environment has promoted the diffusion of the regulating agent, which is a favorable factor.

[0124] Step 5 (corresponding to S105): The prediction module 50 is activated. It not only receives the diffusion efficiency correction value of +1.20, but also obtains the 24-hour weather forecast through the API, showing that the high temperature will continue, and there may be thunder showers in the afternoon (which may cause slight fluctuations in the quality of municipal raw water). After comprehensive analysis, the prediction model generates a more refined dynamic adjustment scheme: maintain the 25% increase in the dosage rate for the next 2 hours to make up for the current deficit; starting from the third hour, considering the continuous promoting effect of high temperature and the possible decline in water consumption from the peak, the dosage rate will be reduced to a level 15% higher than the base value; and it is recommended to double the frequency of subsequent iterative optimization during the thunder shower warning period to cope with possible changes in raw water.

[0125] Step 6 (corresponding to S106): The execution and update module 60 performs stability checking on the scheme. The simulation shows that the scheme execution will not cause the pipe network pressure to be lower than the safety lower limit of 0.2 MPa, and it is determined to meet the requirements. Subsequently, the module formally executes the scheme and updates the operating curve of the metering pump. After 30 minutes of scheme execution, the system collects the TDS data of the whole network, and calculates that the taste consistency index (TDS standard deviation) of the whole system has been reduced from 9.5 mg / L at the peak to 3.2 mg / L, successfully controlling the water quality uniformity at an excellent level.

[0126] Step 7 (corresponding to S107): After passing through the water consumption peak, the system enters normal operation, and the iterative optimization module 80 is activated. In the next 48 hours, it performs multiple iterations of calculation with the joint optimization target of minimizing the TDS standard deviation and the energy consumption of the injection device. Finally, it is found that by changing the injection method from smooth and continuous to "pulsed injection" with the same frequency as the water flow pulsation of the main pipe network caused by the periodic start and stop of large water pumps, the total operating power consumption of the metering pump can be reduced by about 7% without reducing or even slightly improving the mixing uniformity. This final mixing diffusion process control parameter is fixed and used to guide subsequent normal operation.

[0127] Figure 6The technical effect comparison of the straight drinking water taste adjustment technology before and after the application of the technical solution of the present application is shown. The abscissa is time, and the ordinate is the taste consistency index, the conductivity standard deviation, and the smaller the value, the better the consistency. Before T0 moment, the system adopts the static adding mode of the prior art, and it can be seen that in the water peak period (region A in the figure), the index value increases sharply, indicating that the taste consistency is poor. At T0 moment, the control method of the present application is started. It can be seen that after starting, even in the similar water peak period (region B in the figure), the index value only has a small fluctuation, and can be quickly adjusted by the system to fall to a lower stable level (within the target range shown in the figure). This directly shows that the technical solution of the present application can effectively inhibit the taste unevenness caused by the change of working conditions, and maintain the taste consistency of the whole system at a high level.

[0128] Figure 7 The technical effect comparison of the present application embodiment relative to the prior art in multiple dimensions is shown in the schematic diagram, which includes four subgraphs (a), (b), (c) and (d).

[0129] Figure (a) shows the comparison of taste consistency (measured by TDS standard deviation) in the water peak period. The curve shows that, using the prior art scheme, the TDS standard deviation greatly exceeds the target stable range during the peak period; while using the scheme of the present application, the TDS standard deviation can be effectively controlled within the target range even during the peak period, and the control effect on the taste consistency is better under high dynamic working conditions.

[0130] Figure (b) shows the adaptability comparison at different seasonal temperatures. The curve shows that the prior art has a significant decrease in the taste consistency index at low winter temperature (5℃) and high summer temperature (30℃); while the scheme of the present application can maintain a high level of taste consistency through temperature correction and prediction, and has stronger adaptability to temperature changes.

[0131] Figure (c) shows the comparison of pipe network end performance, with the abscissa being the distance from the water source and the ordinate being the concentration deviation of the adjusting agent. The curve shows that, with the increase of the distance, the concentration deviation of the prior art increases exponentially; while the scheme of the present application can always maintain the end concentration deviation at a very low level, effectively solving the mixing unevenness problem under long-distance transportation.

[0132] Figure (d) shows the comparison of system energy consumption. The curve shows that the prior art uses continuous operation mode to achieve mixing, and the power consumption is relatively constant (average 100.1 W); while the iterative optimization module of the present application finds a pulse injection energy-saving strategy with the same frequency as the water flow pulsation of the main pipe network, which reduces the average power consumption of the metering pump to 93.0 W while ensuring or even improving the mixing effect, achieving an energy-saving effect of about 7.1%.

[0133] Through this complete process, the application not only solves the problem of uneven taste caused by sudden water peak in this specific application scenario, but also prospectively considers temperature and weather changes, and through continuous self-optimization, improves the economic efficiency of the system under the premise of ensuring consistent taste.

[0134] Although the present application is disclosed in the preferred embodiments as above, it is not intended to limit the present application, and any person skilled in the art can make many possible changes and modifications without departing from the spirit and scope of the present application, therefore the protection scope of the present application shall be defined by the appended claims.

[0135] Although the present application is disclosed in the preferred embodiments as above, it is not intended to limit the present application, and any person skilled in the art can make many possible changes and modifications without departing from the spirit and scope of the present application, therefore the protection scope of the present application shall be defined by the appended claims.

[0136] The above describes one embodiment of the present application in detail, but the content described is only the preferred embodiment of the present application, and cannot be considered to limit the implementation scope of the present application. Any equivalent changes and improvements made within the scope of the present application shall still belong to the patent coverage scope of the present application.

Claims

1. A method for adjusting the mouthfeel control of direct pipe drinking water, characterized by, The method comprises the following steps: acquiring real-time dynamic parameters of water flow in a pipeline, the real-time dynamic parameters comprising at least variation data of one or more of flow rate, pressure and pipe diameter, determining an initial distribution state of a taste regulator in the water body based on the real-time dynamic parameters, and generating a water flow mixing state parameter; grouping the distribution of the taste regulator using a preset clustering algorithm according to the water flow mixing state parameter, to identify and determine a distribution uneven area where the concentration deviation of the regulator exceeds a preset range; judging whether the concentration deviation of the distribution uneven area exceeds a preset concentration deviation threshold value; if yes, adjusting the addition proportion of the taste regulator by a water treatment injection device, and generating an optimized regulator distribution model; based on the optimized regulator distribution model, acquiring a dissolution efficiency index, and fusing the dissolution efficiency index with real-time monitored water temperature variation data to determine a diffusion efficiency correction value reflecting the influence of temperature on regulator diffusion; analyzing the diffusion efficiency correction value and collected seasonal environmental data by a prediction model, and generating a dynamic adjustment scheme; under the premise that the dynamic adjustment scheme meets a preset system stability requirement, updating the addition proportion of the regulator using the dynamic adjustment scheme, and calculating a whole-system taste consistency index representing the uniformity of the overall quality of the water body; based on the whole-system taste consistency index, acquiring real-time feedback data, and optimizing the index by a cyclic iteration method until it converges to a preset target range, thereby determining final mixing and diffusion process control parameters for guiding the continuous operation of the water treatment equipment.

2. The method of claim 1, wherein, The step of acquiring real-time dynamic parameters of water flow in a pipeline and generating a water flow mixing state parameter comprises: real-time collecting variation data of at least one or more of flow rate, pressure and pipe diameter by a sensor deployed in the pipeline; calculating the covariance of the flow rate variation data and the pressure variation data to represent the change trend of the pipe diameter; integrating the change trend of the pipe diameter and real-time monitored temperature data to determine the initial concentration distribution of the regulator; inputting the initial concentration distribution and the flow rate variation data into a preset fluid dynamics model to calculate a diffusion rate value, which together with the initial concentration distribution constitutes the water flow mixing state parameter.

3. The method of claim 2, wherein, The fluid dynamics model is a Navier-Stokes equation model; the diffusion rate value is obtained by solving the partial differential equations of the velocity field and the concentration field in the Navier-Stokes equation model.

4. The method of claim 1, wherein, The step of grouping the distribution of the taste regulator using a preset clustering algorithm comprises: grouping by a K-means clustering algorithm with the water flow mixing state parameter as input to obtain preliminary distribution groups; calculating the difference between the average concentration of the regulator in each group and the overall average concentration of the water body to obtain a concentration deviation value; if the concentration deviation value exceeds a preset first threshold value, the pipeline part corresponding to the group is determined as the distribution uneven area.

5. The method of claim 1, wherein, The step of analyzing the diffusion efficiency correction value and the collected seasonal environmental data by the prediction model specifically comprises: Collecting historical seasonal environmental data and corresponding historical diffusion efficiency correction values to form a training data set; Training the prediction model using the training data set, wherein the prediction model is a linear regression model, a neural network model, or a time series analysis model; Inputting real-time collected seasonal environmental data and the currently calculated diffusion efficiency correction value into the trained prediction model to output the dynamic adjustment scheme including the adjusting agent dosage and the mixing device operating parameter adjustment.

6. The method of claim 1, wherein, The judgment of whether the dynamic adjustment scheme meets the preset system stability requirement is specifically achieved by monitoring the fluctuation amplitude of key operating parameters in the water treatment system, wherein the key operating parameters include water flow rate fluctuation, pressure fluctuation, or pH value fluctuation; when the fluctuation amplitude of the key operating parameters is less than the preset stability threshold, it is judged that the system stability requirement is met.

7. A mouth feel control system for straight pipe drinking water, characterized by, Comprise: A data acquisition module for acquiring real-time dynamic parameters of water flow in a pipeline, wherein the real-time dynamic parameters at least include change data of one or more of flow rate, pressure, and pipe diameter; An analysis module connected with the data acquisition module, for determining the initial distribution state of the taste adjusting agent in the water body based on the real-time dynamic parameters, generating water flow mixing state parameters, and identifying and determining the uneven distribution area by using a preset clustering algorithm according to the water flow mixing state parameters; A first adjustment module connected with the analysis module, for generating an adjustment instruction to adjust the dosage proportion of the taste adjusting agent when the concentration deviation in the uneven distribution area exceeds a preset concentration deviation threshold, and constructing an optimized adjusting agent distribution model; A correction module connected with the first adjustment module, for acquiring a dissolution efficiency index based on the optimized adjusting agent distribution model, and calculating a diffusion efficiency correction value in combination with real-time water temperature data; A prediction module connected with the correction module, for analyzing the diffusion efficiency correction value and seasonal environmental data by a prediction model to generate a dynamic adjustment scheme; An execution and update module connected with the prediction module, for executing the scheme to update the dosage proportion of the adjusting agent and calculating a full-system taste consistency index when the dynamic adjustment scheme meets the system stability requirement; An iterative optimization module connected with the execution and update module, for optimizing through a cyclic iteration mode according to real-time feedback of the full-system taste consistency index to determine final mixing diffusion process control parameters and send them to a water treatment device.

8. The system of claim 7, wherein, The data acquisition module comprises at least one ultrasonic flowmeter, pressure sensor, displacement sensor, and temperature sensor arranged in the pipeline.

9. The system of claim 7, wherein, The analysis module is specifically configured to: Receive the real-time dynamic parameters and input them into a built-in fluid dynamics model to calculate water flow mixing state parameters representing the initial distribution characteristics of the taste adjusting agent in the water body; The K-means clustering algorithm or the DBSCAN density clustering algorithm is executed to perform clustering analysis on the data set composed of the water flow mixing state parameters, so as to identify a data cluster with large concentration deviation, and determine the uneven distribution area.

10. The system of claim 7, wherein, The iterative optimization module adopts a gradient descent algorithm, constructs a target function based on an error between the full-system taste consistency index and a preset target value, and updates the mixing diffusion process control parameter through iterative calculation until a value of the target function is less than a preset convergence threshold.

Images