A large-capacity ozone water preparation device and an ozone concentration intelligent adjusting method thereof

Abstract

The present application relates to a kind of large capacity ozone water preparation device and its ozone concentration intelligent adjusting method, including ozone water tank, new water tank, water tank three major parts, and is equipped with corresponding ozone concentration sensor, water quality sensor, ozone generator, air supplement device, circulating water pump etc., while PID control, BP neural network, support vector machine SVR three kinds of algorithms are applied to ozone water storage stage, new water injection stage, ozone water use stage three stages.The present application realizes whether new water injection or ozone water discharge, and no matter ozone water inner water temperature, water quality, PH value, set ozone concentration etc.How to change, can be in the case without manual intervention, adaptively maintain the ozone concentration in ozone water tank unchanged, can provide the large capacity ozone water of specified concentration at any time.The present application can be widely used in water treatment, air purification, food processing, medical disease control, biological pharmaceutical, aquaculture and other fields of sterilization and disinfection.

Description

Technical Field

[0001] This invention belongs to the field of disinfection and sterilization technology, specifically relating to a large-capacity ozone water preparation device and its intelligent ozone concentration adjustment method. Background Technology

[0002] Ozone water, as a disinfection medium, has outstanding advantages: it not only has strong and rapid bactericidal ability, but also releases oxygen upon decomposition, making it safe and without side effects. Ozone water is considered widely applicable in water treatment, air purification, food processing, medical disease control, biopharmaceuticals, and aquaculture. However, compared to other media such as 84 disinfectant, the use of ozone water for disinfection is currently only on a small scale and has not been widely adopted. The fundamental reason for this is that it is difficult to maintain a stable concentration of ozone water.

[0003] In many situations, it is necessary to prepare ozone water of a specified concentration in advance for immediate use. For example, in laboratories used in biopharmaceuticals and disease control, the pure water pipeline network needs to be disinfected with ozone water periodically to prevent bacterial growth and contamination of the pure water, which could affect experimental results. Since the preparation and use of pure water in laboratories is on-demand, there is uncertainty regarding when to shut down the water purifier, when to cut off the flow of pure water in the pipeline network, and when to fill the network with ozone water for disinfection. Therefore, it is usually necessary to prepare ozone water of the specified concentration in advance for immediate use.

[0004] Commonly used tools for ozone concentration adjustment include packed towers, bubble column contact reactors, gas-liquid mixing pumps, and ejectors. However, these devices are only suitable for adjusting ozone concentration and are best used immediately, but not for storing large volumes of ozone water with a constant concentration.

[0005] Chinese patent CN 108268064 A discloses an ozone disinfection water preparation machine with an automatic outlet concentration control system, and Chinese patent CN 214596604 U discloses a disinfection device that automatically adjusts ozone concentration. Both patents mention using a PID algorithm combined with multiple auxiliary algorithms to precisely control ozone concentration. However, the PID method has limitations in practical applications, especially when water quality, water temperature, pH value, and ozone concentration change. The P, I, and D parameters need to be readjusted to maintain closed-loop control performance and keep the ozone concentration stable. However, adjusting the P, I, and D parameters every time water quality, water temperature, pH value, and ozone concentration change is impractical. Summary of the Invention

[0006] To overcome the shortcomings of existing technologies, this invention provides a large-capacity ozone water preparation device for disinfection and sterilization, and a method for intelligently adjusting the ozone concentration. Regardless of whether new water is injected or ozone water is discharged, and regardless of changes in the water temperature, quality, pH value, etc., the ozone concentration in the ozone water tank can be maintained constant without human intervention, providing a large-capacity ozone water of the specified concentration at any time.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0008] A large-capacity ozone water preparation device includes an ozone water tank, an ozone concentration sensor, an ozone water circulating pump, an ozone water tank aeration device, an ozone water tank water quality sensor, an ozone water tank level sensor, an ozone water sampling point, a fresh water tank, a fresh water pump, a fresh water tank aeration device, a fresh water tank water quality sensor, a fresh water tank concentration valve, a fresh water tank water injection valve, an ozone generator, an ozone water aeration valve, a fresh water aeration valve, an ozone outlet valve, a water tank for use, a delivery pump, an ozone outlet water quality sensor, an ozone outlet pipe, a PLC controller, an edge gateway, and a cloud platform.

[0009] The ozone water sampling point is connected to the ozone concentration sensor, which is connected to the inlet of the ozone water circulation pump. The outlet of the ozone water circulation pump is connected to the bottom inlet of the ozone water tank aeration device. The outlet pipe of the ozone water tank aeration device extends to the bottom of the ozone water tank, thereby realizing the circulation of ozone water. At the same time, the outlet of the ozone generator is connected to the upper inlet of the ozone water tank aeration device through the ozone water aeration valve, realizing the injection of ozone into the ozone water tank.

[0010] The outlet of the fresh water tank is connected to the inlet of the fresh water pump. The outlet of the fresh water pump is divided into two paths: one path goes through the fresh water tank injection valve to the ozone tank, and the other path goes through the fresh water tank concentration valve to the bottom inlet of the fresh water tank aeration device. The outlet pipe of the fresh water tank aeration device extends to the bottom of the fresh water tank to achieve the circulation of fresh water. At the same time, the outlet of the ozone generator is connected through the fresh water aeration valve to the upper inlet of the fresh water tank aeration device to achieve the injection of ozone into the fresh water tank.

[0011] The ozone tank is connected to the water tank via an ozone outlet valve. The bottom outlet of the water tank is connected to the inlet of the delivery pump. The outlet of the delivery pump is connected to the ozone outlet via a long flexible hose. An ozone water quality sensor is installed near the ozone outlet.

[0012] The output signal lines of the ozone tank water quality sensor, the fresh water tank water quality sensor, the ozone outlet water quality sensor, the ozone tank level sensor, and the ozone concentration sensor are all connected to the input interface of the PLC controller; the control signal lines of the fresh water tank concentration valve, the fresh water tank filling valve, the ozone water aeration valve, the fresh water aeration valve, and the ozone outlet valve are all connected to the output interface of the PLC controller; the control signal line of the ozone generator is connected to the analog output interface of the PLC controller.

[0013] Preferably, the ozone tank aeration device and the fresh water tank aeration device are any one of the following: packed tower contact reactor, bubble tower contact reactor, gas-liquid mixing pump, and jet generator.

[0014] Preferably, there are 3 to 5 ozone water sampling points, which are distributed in different areas at the bottom of the ozone water tank to ensure the accuracy of ozone concentration detection.

[0015] Preferably, the ozone tank water quality sensor includes a first water temperature sensor, a first conductivity sensor, and a first pH sensor; the fresh water tank water quality sensor includes a second water temperature sensor, a second conductivity sensor, and a second pH sensor; and the ozone outlet water quality sensor includes a third water temperature sensor, a third conductivity sensor, and a third pH sensor.

[0016] The present invention also provides a method for intelligent adjustment of ozone concentration of the above-mentioned device, the method covering three stages: ozone water storage stage, new water injection stage, and ozone water use stage;

[0017] During the ozone water storage stage, the PLC collects the signal from the ozone concentration sensor, then performs PID calculation according to the following formula, and adjusts the analog voltage value sent by the PLC to the ozone generator based on the calculation result, thereby adjusting the ozone output of the ozone generator to compensate for the dynamic balance of ozone concentration in the ozone water tank.

[0018]

[0019] In the above formula, u(t) is the analog voltage value output by the PLC to the ozone generator at time t, and K p K i K d The three parameters of the PID are the proportional parameter, integral parameter, and derivative parameter. e(t) and e(t-1) are the differences between the set ozone concentration and the actual ozone concentration at time t and time t-1, respectively.

[0020] During the ozone water storage stage, when any parameters such as water temperature, conductivity, and pH value in the ozone water tank change, or when the set ozone concentration value changes, the P, I, and D parameters in the above formula will be adjusted by the BP neural network described in the following formula; the input parameters of the BP neural network are the ozone concentration difference e(t) and the ozone concentration change rate e(t)-e(t-1), and the output parameters are the three parameters P, I, and D;

[0021] Y = f(W, X)

[0022] In the above formula: X is the input term, X = (X1, X2), where X1 represents e(t) and X2 represents e(t) - e(t-1); Y is the output term, Y = (K p K i K d W is the weight matrix between the hidden layer and the output layer, and satisfies the following relationship:

[0023]

[0024] In the above formula: M represents the number of nodes in the output layer, L represents the number of nodes in the hidden layer, and w ij f represents the weights between the input layer and the hidden layer. 1i y represents the non-linear activation function between the input layer and the hidden layer. 1i Indicates the hidden layer output value; w ki f represents the weights between the hidden layer and the output layer. 2k This represents the nonlinear activation function between hidden layers and the output layer;

[0025] During the ozone water storage phase, the Support Vector Machine (SVR) shown in the following formula continuously judges the relevant parameters in the ozone water tank, which serve as training sample data for the SVR's self-learning. The input parameters of the SVR are water temperature, conductivity, pH value, and ozone concentration, and the output parameter is the ozone dissipation rate.

[0026] f(x) = w T x+b

[0027] In the above formula: f(x) represents the ozone dissipation rate; x is the input variable of the support vector machine (SVR), namely the four parameters after normalization: current water temperature T, conductivity D, pH value P, and set ozone concentration Q, i.e., x = [TDPQ]′, where the symbol ' represents matrix transpose; w represents the classification hyperplane normal vector of the SVR, and b represents the bias value;

[0028] During the new water injection phase, the ozone generator rapidly injects ozone into the new water to equalize the ozone concentration in the ozone tank before delivering the ozone water to the tank. The ozone gas injected during this phase consists of two parts: the first part is a certain volume of ozone gas used to rapidly increase the ozone concentration, calculated based on the ozone concentration in the ozone tank and the total volume of new water in the tank; the second part is ozone gas continuously input at a certain rate to compensate for dissipated ozone. The input parameters for the SVR are the new water temperature, conductivity, and pH value detected by the water quality sensor in the new water tank, along with the current ozone concentration in the ozone tank. The SVR then calculates the ozone dissipation rate and the injection rate of the second part of the ozone gas.

[0029] During the use of ozone water, the areas disinfected by ozone water will rapidly change the water temperature, conductivity, and pH value parameters. The water temperature, conductivity, and pH parameters detected by the ozone outlet water quality sensor, along with the current ozone concentration in the ozone water tank, are used as input parameters for the SVR. The SVR then calculates the ozone dissipation rate, and based on the ozone dissipation rate, it estimates the disinfection time that the current ozone can support, and at the same time estimates the total amount of ozone water required.

[0030] Preferably, the intelligent ozone concentration adjustment method includes the following steps:

[0031] (1) Power on, the user sets the required ozone water capacity and ozone water concentration, and the PLC detects the water temperature, conductivity and pH value of the new water injected;

[0032] (2) The edge gateway quickly calculates the ozone dissipation rate corresponding to the new water based on the trained SVR model, and then executes the algorithm of the new water injection stage to quickly turn the new water into ozone water of a specified concentration and enter the ozone water storage stage.

[0033] (3) When the user needs to use ozone water, the PLC controls the ozone outlet valve to inject the required amount of ozone water into the water tank. At the same time, by executing the algorithm of the ozone water usage stage, the PLC displays the estimated disinfection time that the ozone in the water tank can maintain on the display screen, and also displays the total amount of ozone water required for the user to use, as well as how much more ozone water needs to be injected into the water tank.

[0034] (4) When the total amount of ozone water is too low, remind the user to add water to the new water tank and then repeat steps (1) to (4).

[0035] As a preferred method, during the ozone water storage stage, the criteria for determining whether relevant parameters in the ozone water tank can be used as training sample data for SVR are as follows: when the ozone concentration is consistently stable at ±0.01 mg / L of the set concentration for one minute, the average value of the five parameters—water temperature, conductivity, pH value, ozone concentration, and ozone dissipation rate—within that one minute will be used as a set of training samples.

[0036] Preferably, the PID algorithm is executed within the PLC, while the Support Vector Machine (SVR) algorithm and the neural network algorithm are executed within the edge gateway.

[0037] The beneficial effects of this invention are as follows:

[0038] 1. This invention is low in cost and has a very high cost-performance ratio. Compared with traditional ozone water preparation devices, this invention, based on hardware, combined with the application of BP neural network and PID algorithm, realizes the preparation of large-capacity, stable ozone water at a specified concentration that can be used at any time.

[0039] 2. By separating the ozone tank, the fresh water tank, and the used water tank, this invention ensures, from a hardware perspective, that neither the water replenishment process nor the water usage process will affect the stability of the ozone concentration in the ozone tank.

[0040] 3. This invention uses the Support Vector Machine (SVR) algorithm to quickly calculate the ozone required for injecting new water and the total amount of water needed for disinfection, thus ensuring, from a software perspective, that the impact of the water replenishment and water usage processes on ozone concentration can be rapidly eliminated.

[0041] 4. Existing ozone concentration sensors are expensive. In the ozone water usage stage, this invention can estimate the disinfection time that ozone can support without an ozone concentration sensor by using only ordinary sensors and related algorithms. At the same time, in the new water injection stage, ozone in the new water can be accurately adjusted to the specified concentration by using only ordinary sensors and related algorithms without an ozone concentration sensor. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of the structure of the large-capacity ozone water preparation device of the present invention;

[0043] Figure 2 This is a flowchart of the intelligent ozone concentration adjustment method of the present invention;

[0044] Diagram Description: 1. Ozone Water Tank; 2. Ozone Concentration Sensor; 3. Ozone Water Circulation Pump; 4. Ozone Water Tank Aeration Device; 5a. First Water Temperature Sensor; 5b. First Conductivity Sensor; 5c. First pH Sensor; 6. Ozone Water Tank Level Sensor; 7. Ozone Water Sampling Point; 8. Fresh Water Tank; 9. Fresh Water Pump; 10. Fresh Water Tank Aeration Device; 11. Fresh Water Tank Inlet; 12a. Second Water Temperature Sensor; 12b. Second pH Sensor; 12c. Fresh Water Tank Concentration Valve; 13. Fresh Water Tank Injection Valve; 14. Ozone Generator; 15a. Ozone Water Aeration Valve; 15b. Fresh Water Aeration Valve; 15c. Ozone Outlet Valve; 16. Water Tank; 17. Water Delivery Pump; 18. Third Water Temperature Sensor; 19a. Third Conductivity Sensor; 19b. Third pH Sensor; 19c. Ozone Outlet Pipe; 20. PLC Controller; 21. Edge Gateway; 22. Cloud Platform; 23. Detailed Implementation

[0045] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be further described below with reference to the accompanying drawings and embodiments, but the scope of protection of the present invention is not limited thereto.

[0046] Reference Figure 1 , Figure 2 A large-capacity ozone water preparation device, comprising an ozone water tank 1, an ozone concentration sensor 2, an ozone water circulation pump 3, an ozone water tank aeration device 4, an ozone water tank water quality sensor, an ozone water tank liquid level sensor 6, an ozone water sampling point 7, a fresh water tank 8, a fresh water pump 9, a fresh water tank aeration device 10, a fresh water tank water quality sensor, a fresh water tank concentration valve 13, a fresh water tank water injection valve 14, an ozone generator 15a, an ozone water aeration valve 15b, a fresh water aeration valve 15c, an ozone outlet valve 16, a water tank for use 17, a delivery pump 18, an ozone outlet water quality sensor, an ozone outlet pipe 20, a PLC controller 21, an edge gateway 22, and a cloud platform 23.

[0047] The number of ozone water sampling points is 3 to 5, and they are distributed in different areas at the bottom of the ozone water tank to ensure that ozone concentration can be collected regardless of the amount of ozone water in the tank. By collecting ozone water from multiple areas, an average method can be used to ensure the accuracy of ozone concentration detection.

[0048] Ozone water sampling point 7 is connected to ozone concentration sensor 2, ozone concentration sensor 2 is connected to the inlet of ozone water circulation pump 3, ozone water circulation pump 3 is connected to the bottom inlet of ozone water tank aeration device 4, and the outlet pipe of ozone water tank aeration device 4 extends to the bottom of ozone water tank 1, thereby realizing the circulation of ozone water. The circulation of ozone water can ensure that even if the ozone water volume in the tank is large, the ozone concentration in different areas is balanced.

[0049] Meanwhile, the outlet of ozone generator 15a is connected to the upper inlet of ozone water tank gas supply device 4 via ozone water gas supply valve 15b, realizing the injection of ozone into the ozone water tank; the specific forms of ozone water tank gas supply device and fresh water tank gas supply device are any one of packed tower contact reactor, bubble tower contact reactor, gas-liquid mixing pump, and jet injector; in the gas supply device, the water flows from bottom to top, while the ozone is transported from top to bottom, thereby realizing the deep and rapid fusion of ozone gas and ozone water.

[0050] The outlet of the fresh water tank 8 is connected to the inlet of the fresh water pump 9. The outlet of the fresh water pump 9 is divided into two paths: one path connects to the ozone tank 1 via the fresh water tank injection valve 14, and the other path connects to the bottom inlet of the fresh water tank aeration device 10 via the fresh water tank concentration valve 13. The specific path of the water flow is determined by the PLC control: when the fresh water is circulating and ozone gas is being injected, the fresh water tank concentration valve 13 is open and the fresh water tank injection valve 14 is closed; when the ozone concentration is adjusted and ozone water is injected into the ozone tank, the fresh water tank concentration valve 13 is closed and the fresh water tank injection valve 14 is open. It should be noted that in this invention, fresh water refers to water that does not contain ozone and is injected into the fresh water tank through the inlet 11 (external inlet). It can be tap water, pure water, or ultrapure water, etc.

[0051] The outlet pipe of the fresh water tank aeration device 10 extends to the bottom of the fresh water tank 8, thereby realizing the circulation of fresh water between the fresh water tank 8, the fresh water pump 9, and the fresh water tank aeration device 10. At the same time, the outlet of the ozone generator 15a is connected to the upper inlet of the fresh water tank aeration device 10 through the fresh water aeration valve 15c, realizing the injection of ozone into the fresh water tank. Similar to the ozone tank, in the fresh water tank aeration device, the water flows from bottom to top, while the ozone is transported from top to bottom.

[0052] Ozone water tank 1 is connected to water tank 17 via ozone outlet valve 16. When the user needs ozone water, the PLC controls the ozone outlet valve 16 to open, injecting the user-defined volume of ozone water into water tank 17. The bottom outlet of water tank 17 is connected to the inlet of water pump 18, and the outlet of water pump 18 is connected to the ozone outlet via a long hose. An ozone water quality sensor is installed near the ozone outlet. This sensor is closest to the components being disinfected by the ozone water and can detect changes in the ozone water temperature, conductivity, and pH value caused by the disinfected components in real time, thus facilitating the SVR to assess the ozone dissipation rate caused by the disinfected components immediately.

[0053] The output signal lines of the ozone tank water quality sensors (5a, 5b, 5c), the fresh water tank water quality sensors (12a, 12b, 12c), the ozone outlet water quality sensors (19a, 19b, 19c), the ozone tank level sensor 6, and the ozone concentration sensor 2 are all connected to the input interface of the PLC controller 21. The corresponding parameters are collected by the PLC, and the PLC will upload the collected data to the edge gateway 22 via Ethernet. The control signal lines of the fresh water tank concentration valve 13, the fresh water tank water injection valve 14, the ozone water aeration valve 15b, the fresh water aeration valve 15c, and the ozone outlet valve 16 are all connected to the output interface of the PLC controller 21. The control signal line of the ozone generator 15a is connected to the analog output interface of the PLC controller 21.

[0054] The ozone tank water quality sensor, the fresh water tank water quality sensor, and the ozone outlet water quality sensor are all sensor combinations. The ozone tank water quality sensor includes a first water temperature sensor 5a, a first conductivity sensor 5b, and a first pH sensor 5c; the fresh water tank water quality sensor includes a second water temperature sensor 12a, a second conductivity sensor 12b, and a second pH sensor 12c; and the ozone outlet water quality sensor includes a third water temperature sensor 19a, a third conductivity sensor 19b, and a third pH sensor 19c. Water temperature, water quality, and pH value all affect the solubility and dissipation rate of ozone in water. Taking water quality parameters as an example, the purer the water, the slower the ozone dissipates, meaning the longer the ozone concentration can be maintained. For example, ozone dissipates quickly in tap water, while it dissipates slowly in pure water. There are many parameters that can characterize water quality, such as resistivity, conductivity, TOC value, etc. This paper selects resistivity as a parameter to characterize water purity. When water contains more impurities (i.e., the water is not pure enough), the water is more likely to conduct electricity. At this time, the resistivity is small, and the ozone in the water is easily dissipated.

[0055] The ozone concentration intelligent regulation method of the above-mentioned device includes an ozone water storage stage, a new water injection stage, and an ozone water usage stage; it mainly uses PID algorithm and BP neural network, and also uses support vector machine SVR algorithm.

[0056] During the ozone water storage stage, the PLC collects the signal from the ozone concentration sensor, then performs PID calculation according to the following formula, and adjusts the analog voltage value sent by the PLC to the ozone generator based on the calculation result, thereby adjusting the ozone output of the ozone generator (the ozone output is directly proportional to the analog voltage value), so as to make up for the dynamic balance of ozone concentration in the ozone water tank.

[0057]

[0058] In the above formula, u(t) is the analog voltage value output by the PLC to the ozone generator at time t, and K p Ki K d The three parameters of the PID are proportional parameter, integral parameter, and derivative parameter. e(t) and e(t-1) are the differences between the set ozone concentration and the actual ozone concentration at time t and time t-1, respectively.

[0059] During the ozone water storage phase, when any parameters such as water temperature, conductivity, and pH value in the ozone water tank change, or when the set ozone concentration value changes, the ozone dissolution rate and dissipation rate will change. The PID parameters that previously achieved optimal control performance may no longer be optimal. At this time, the P, I, and D parameters in the above formula will be adjusted by the BP neural network described in the following formula; the input parameters of the BP neural network are the ozone concentration difference e(t) (set ozone concentration minus actual ozone concentration) and the ozone concentration change rate e(t) - e(t-1) (partial derivative of the ozone concentration difference with respect to time), and the output parameters are the three parameters P, I, and D.

[0060] Y = f(W, X)

[0061] In the above formula: X is the input term, X = (X1, X2), where X1 represents e(t) and X2 represents e(t) - e(t-1); Y is the output term, Y = (K p K i K d W is the weight matrix between the hidden layer and the output layer, and satisfies the following relationship:

[0062]

[0063] In the above formula: M represents the number of nodes in the output layer. In this invention, the output terms are the three parameters of a PID controller, therefore the value of M is 3; L represents the number of nodes in the hidden layer. By repeatedly training and comparing the prediction performance of the neural network, a suitable value of L is finally determined; w ij f represents the weights between the input layer and the hidden layer. 1i y represents the non-linear activation function between the input layer and the hidden layer. 1i Indicates the hidden layer output value; w ki f represents the weights between the hidden layer and the output layer. 2k This represents the nonlinear activation function between hidden layers and the output layer. In the device described in this invention, the BP neural network is typically pre-trained by the manufacturer before leaving the factory, with training data originating partly from the manufacturer's experimental data and partly from data uploaded by the device.

[0064] During the ozone water storage phase, the Support Vector Machine (SVR), as shown in the following formula, continuously evaluates relevant parameters within the ozone water tank, using these parameters as training data for the SVR's self-learning. The SVR's input parameters are water temperature, conductivity, pH value, and ozone concentration. The output parameter is the ozone dissipation rate (the amount of ozone gas injected divided by the total amount of ozone water when the ozone concentration is stable). When the ozone concentration remains stable within ±0.01 mg / L of the set concentration for one minute, the average value of the five parameters—water temperature, conductivity, pH value, ozone concentration, and ozone dissipation rate—within that minute will be used as a set of training samples.

[0065] f(x) = w T x+b

[0066] In the above formula: f(x) represents the ozone dissipation rate; x is the input variable of the Support Vector Machine (SVR), namely, the four parameters after normalization: current water temperature (T), conductivity (D), pH value (P), and set ozone concentration (Q), i.e., x = [TDPQ]′, where the symbol ' denotes matrix transpose; w represents the classification hyperplane normal vector of the SVR, and b represents the bias value. Training the SVR with sample data essentially involves iteratively optimizing the normal vector w and the bias value b. After determining the optimal normal vector w and bias value b, each input x = [TDPQ] is used. T The ozone dissipation rate can be calculated from the value.

[0067] During the new water injection phase, the ozone generator rapidly injects ozone into the new water to equalize the ozone concentration in the ozone tank before delivering the ozone water to the tank. The ozone gas injected during this phase consists of two parts. The first part is a fixed volume of ozone gas used to rapidly increase the ozone concentration; this volume is calculated based on the ozone concentration in the ozone tank and the total volume of new water in the tank. The second part is a continuous inflow of ozone gas at a specific rate to compensate for dissipated ozone. The new water temperature, conductivity, and pH value detected by the water quality sensors (12a, 12b, 12c) in the new water tank, along with the current ozone concentration in the ozone tank, are used as input parameters for the SVR. The SVR then calculates the ozone dissipation rate, thus determining the injection rate of the second part of the ozone gas.

[0068] During the ozone water usage phase, the ozone water disinfection process rapidly alters the water temperature, conductivity, and pH parameters. The water temperature, conductivity, and pH parameters detected by the ozone outlet water quality sensors (19a, 19b, 19c), along with the current ozone concentration in the ozone water tank, are used as input parameters for the SVR. The SVR then calculates the ozone dissipation rate, and based on the ozone dissipation rate, it estimates the disinfection time that the current ozone can support, as well as the total amount of ozone water required.

[0069] The aforementioned PID algorithm is executed within the PLC, while the Support Vector Machine (SVR) algorithm and neural network algorithm are executed within the edge gateway, thereby distributing the computational load.

[0070] In this invention, the PLC can be a Siemens S7-200 or similar product, and the edge gateway can be a small industrial control host priced between 1,000 and 3,000 yuan, with a Linux or Windows operating system installed inside. The main control task is undertaken by the PLC to leverage its stability and reliability; the PID algorithm is also implemented by a built-in module in the PLC. Algorithms such as BP neural networks and Support Vector Machines (SVR) are executed within the edge gateway to fully utilize its computing resources. Furthermore, the execution process and results of the BP neural network within the edge gateway are fed back to the cloud platform via 4G or 5G. After collecting the BP neural network execution parameters from multiple ozone water preparation devices, the manufacturer can better train the BP neural network in ozone water preparation devices that have not yet left the factory. Alternatively, the BP neural network parameters in ozone water preparation devices already in production can be updated via 4G or 5G communication to ensure better ozone concentration control.

[0071] Upon startup, the user sets the desired ozone water volume and concentration. The PLC detects the temperature, conductivity, and pH of the incoming water. The edge gateway then quickly calculates the ozone dissipation rate of the new water using a pre-trained SVR model and executes the algorithm described in the new water injection phase to rapidly convert the new water into ozone water of the specified concentration. When the user needs to use the ozone water, the PLC controls the ozone outlet valve to inject the required volume of ozone water into the water tank. Simultaneously, by executing the algorithm described in the ozone water usage phase, the PLC displays the estimated disinfection time that the ozone in the water tank can sustain, the recommended total volume of ozone water needed, and how much more ozone water needs to be added to the water tank. When the total ozone water volume is too low, the user is prompted to add water to the new water tank, and the above steps are repeated.

[0072] This invention consists of three main parts: an ozone water tank, a fresh water tank, and a used water tank. It is equipped with corresponding ozone concentration sensors, water quality sensors, an ozone generator, an aeration device, and a circulating water pump. Furthermore, it applies three types of algorithms—PID control, BP neural network, and Support Vector Machine (SVR)—to the ozone water storage stage, the fresh water injection stage, and the ozone water usage stage. This enables the invention to adaptively maintain a constant ozone concentration within the ozone water tank without human intervention, regardless of changes in water temperature, water quality, pH value, or the set ozone concentration, whether it's fresh water injection or ozone water discharge. It can provide a large volume of ozone water at the specified concentration at any time. This invention can be widely used for sterilization and disinfection in water treatment, air purification, food processing, medical disease control, biopharmaceuticals, and aquaculture.

[0073] The present invention has been described in detail above with reference to the embodiments. However, the content described is only a specific implementation of the present invention and should not be construed as limiting the scope of the present invention. It should be noted that for those skilled in the art, any modifications and improvements made in accordance with the scope of the present invention without departing from the concept of the present invention should still fall within the patent coverage of the present invention.

Claims

1. A method for intelligent adjustment of ozone concentration, characterized in that: The method covers three stages: ozone water storage, fresh water injection, and ozone water usage. During the ozone water storage stage, the PLC collects the signal from the ozone concentration sensor, then performs PID calculation according to the following formula, and adjusts the analog voltage value sent by the PLC to the ozone generator based on the calculation result, thereby adjusting the ozone output of the ozone generator to compensate for the dynamic balance of ozone concentration in the ozone water tank. ； In the formula, u(t) is the analog voltage value output by the PLC to the ozone generator at time t, K p , K i , K d are PID three parameters, i.e. proportional parameter, integral parameter and differential parameter, e(t) and e(t-1) are respectively the difference between the set ozone concentration and the actual ozone concentration at time t and time t-1. During the ozone water storage stage, when any parameter among the water temperature, conductivity, and pH value in the ozone water tank changes, or when the set ozone concentration value changes, the P, I, and D parameters in the above formula will be adjusted by the BP neural network described in the following formula; the input parameters of the BP neural network are the ozone concentration difference e(t) and the ozone concentration change rate e(t) - e(t-1), and the output parameters are the three parameters P, I, and D; ； In the above formula: X is an input term, X=(X1, X2), wherein X1 represents e(t), and X2 represents e(t)-e(t-1); Y is an output term, Y=(K p , K i , K d ); W is a weight matrix between the hidden layer and the output layer, and satisfies the following relationship: ； In the above formula: M represents the number of nodes in the input layer, L represents the number of nodes in the hidden layer, and w ij f represents the weights between the input layer and the hidden layer. 1i y represents the non-linear activation function between the input layer and the hidden layer. 1i This represents the output value of the hidden layer; w ki denotes the weights between the hidden and output layers, f 2k denotes the non-linear activation function between the hidden layers and the output layer; During the ozone water storage phase, the Support Vector Machine (SVR) shown in the following formula continuously judges the relevant parameters in the ozone water tank, which serve as training sample data for the SVR's self-learning. The input parameters of the SVR are water temperature, conductivity, pH value, and ozone concentration, and the output parameter is the ozone dissipation rate. ； In the above formula: f(x) represents the ozone dissipation rate; x is the input variable of the support vector machine (SVR), namely, the four parameters after normalization: current water temperature T, conductivity D, pH value P, and set ozone concentration Q. The symbol ' denotes matrix transpose; w represents the classification hyperplane normal vector of SVR, and b represents the bias value; During the new water injection phase, the ozone generator rapidly injects ozone into the new water to equalize the ozone concentration in the ozone tank before delivering the ozone water to the tank. The ozone gas injected during this phase consists of two parts: the first part is a certain volume of ozone gas used to rapidly increase the ozone concentration, calculated based on the ozone concentration in the ozone tank and the total volume of new water in the tank; the second part is ozone gas continuously input at a certain rate to compensate for dissipated ozone. The input parameters for the SVR are the new water temperature, conductivity, and pH value detected by the water quality sensor in the new water tank, along with the current ozone concentration in the ozone tank. The SVR then calculates the ozone dissipation rate and the injection rate of the second part of the ozone gas. During the use of ozone water, the areas disinfected by ozone water will rapidly change the water temperature, conductivity, and pH value parameters. The water temperature, conductivity, and pH parameters detected by the ozone outlet water quality sensor, along with the current ozone concentration in the ozone water tank, are used as input parameters for the SVR. The SVR then calculates the ozone dissipation rate, and based on the ozone dissipation rate, it estimates the disinfection time that the current ozone can support, and at the same time estimates the total amount of ozone water required.

2. The method of claim 1, wherein Includes the following steps: (1) Power on, the user sets the required ozone water capacity and ozone water concentration, and the PLC detects the water temperature, conductivity and pH value of the new water injected; (2) The edge gateway quickly calculates the ozone dissipation rate corresponding to the new water based on the trained SVR model, and then executes the algorithm of the new water injection stage to quickly turn the new water into ozone water of a specified concentration and enter the ozone water storage stage. (3) When the user needs to use ozone water, the PLC controls the ozone outlet valve to inject the required amount of ozone water into the water tank. At the same time, by executing the algorithm of the ozone water usage stage, the PLC displays the estimated disinfection time that the ozone in the water tank can maintain on the display screen, and also displays the total amount of ozone water required for the user to use, as well as how much more ozone water needs to be injected into the water tank. (4) When the total amount of ozone water is too low, remind the user to add water to the new water tank and then repeat steps (1) to (4).

3. The method of claim 1, wherein: During the ozone water storage phase, the criteria for determining whether relevant parameters in the ozone water tank can be used as training sample data for SVR are as follows: when the ozone concentration is consistently stable at ±0.01 mg / L of the set concentration for one minute, the average value of the five parameters—water temperature, conductivity, pH value, ozone concentration, and ozone dissipation rate—within that one minute will be used as a set of training samples.

4. The method of claim 1, wherein: The PID algorithm is executed within the PLC, while the Support Vector Machine (SVR) algorithm and the neural network algorithm are executed within the edge gateway.

5. A large capacity ozone water preparation device for the method of claim 1, characterized by: The device includes an ozone water tank, an ozone concentration sensor, an ozone water circulation pump, an ozone water tank aeration device, an ozone water tank water quality sensor, an ozone water tank level sensor, an ozone water sampling point, a fresh water tank, a fresh water pump, a fresh water tank aeration device, a fresh water tank water quality sensor, a fresh water tank concentration valve, a fresh water tank water injection valve, an ozone generator, an ozone water aeration valve, a fresh water aeration valve, an ozone outlet valve, a water tank for use, a delivery pump, an ozone outlet water quality sensor, an ozone outlet pipe, a PLC controller, an edge gateway, and a cloud platform. The ozone water sampling point is connected to the ozone concentration sensor, the ozone concentration sensor is connected to the inlet of the ozone water circulation pump, the outlet of the ozone water circulation pump is connected to the bottom inlet of the ozone water tank aeration device, and the outlet pipe of the ozone water tank aeration device extends to the bottom of the ozone water tank; at the same time, the outlet of the ozone generator is connected to the upper inlet of the ozone water tank aeration device through the ozone water aeration valve. The outlet of the fresh water tank is connected to the inlet of the fresh water pump. The outlet of the fresh water pump is divided into two paths: one path goes through the fresh water tank injection valve to the ozone tank, and the other path goes through the fresh water tank concentration valve to the bottom inlet of the fresh water tank aeration device. The outlet pipe of the fresh water tank aeration device extends to the bottom of the fresh water tank. At the same time, the outlet of the ozone generator is connected to the upper inlet of the fresh water tank aeration device through the fresh water aeration valve. The ozone tank is connected to the water tank via an ozone outlet valve. The bottom outlet of the water tank is connected to the inlet of the delivery pump. The outlet of the delivery pump is connected to the ozone outlet via a long flexible hose. An ozone water quality sensor is installed near the ozone outlet. The output signal lines of the ozone tank water quality sensor, the fresh water tank water quality sensor, the ozone outlet water quality sensor, the ozone tank level sensor, and the ozone concentration sensor are all connected to the input interface of the PLC controller; the control signal lines of the fresh water tank concentration valve, the fresh water tank filling valve, the ozone water aeration valve, the fresh water aeration valve, and the ozone outlet valve are all connected to the output interface of the PLC controller; the control signal line of the ozone generator is connected to the analog output interface of the PLC controller.

6. The large capacity ozone water producing apparatus according to claim 5, wherein: The specific forms of ozone tank aeration devices and fresh water tank aeration devices are any one of the following: packed tower contact reactor, bubble tower contact reactor, gas-liquid mixing pump, and jet generator.

7. The large capacity ozone water producing apparatus according to claim 5, wherein: There are 3 to 5 ozone water sampling points, which are distributed in different areas at the bottom of the ozone water tank.

8. The large capacity ozone water producing apparatus according to claim 5, wherein: The ozone water tank water quality sensor includes a first water temperature sensor, a first conductivity sensor, and a first pH sensor; the fresh water tank water quality sensor includes a second water temperature sensor, a second conductivity sensor, and a second pH sensor; the ozone outlet water quality sensor includes a third water temperature sensor, a third conductivity sensor, and a third pH sensor.

Images