An intelligent water treatment method and system for ship exhaust gas

Abstract

The present application provides an intelligent water treatment method and system for ship exhaust gas, which relates to the technical field of exhaust gas water treatment. The method includes: determining the exhaust gas component set and multiple sewage image sets in the sewage solution; establishing an implicit association between sewage feature descriptors and exhaust gas components; building an exhaust gas component identification channel; collecting real-time sewage images, and transferring them to the exhaust gas component identification channel, and outputting component identification data; calculating the component deviation between the component identification data and the adjacent historical component data, performing simulation optimization of flocculant addition parameters, and determining the optimal addition parameters; correcting the working frequency; sending the sewage solution into the flotation tank, and controlling the flocculant addition of the flotation tank based on the corrected working frequency. Through this application, the technical problem of low accuracy of drug dosage control for waste gas water treatment in the prior art can be solved, the technical goal of improving the degree of intelligence of water treatment can be achieved, and the technical effect of improving the quality of waste gas water treatment can be achieved.

Description

Technical Field

[0001] The present application relates to the technical field of waste gas water treatment, and in particular to an intelligent water treatment method and system for ship exhaust gas. Background Art

[0002] Ship exhaust refers to the exhaust gas produced by the combustion of fuel in the engine of a ship during navigation, which mainly includes sulfur oxides, nitrogen oxides, particulate matter and carbon dioxide. These exhaust gases not only pollute the environment, but also pose a threat to human health.

[0003] At present, the existing ship exhaust water treatment method mainly controls the dosage of the reagent manually, and the accuracy of monitoring the actual state of the sewage solution is low, resulting in low accuracy of dosage control, waste of reagent resources and low treatment accuracy of sewage treatment. Therefore, a method is needed to solve the above problems.

[0004] To sum up, the prior art has technical problems such as low intelligence level in the dosage control process of the reagent, low accuracy in monitoring the actual state of the sewage solution, low accuracy in dosage control, waste of reagent resources and low treatment accuracy of sewage treatment. Summary of the invention

[0005] The purpose of the present application is to provide an intelligent water treatment method and system for ship exhaust gas, so as to solve the technical problems in the prior art that the degree of intelligence in the control process of the dosage addition of the reagent is low, resulting in low accuracy in monitoring the actual state of the sewage solution, low accuracy in dosage control, waste of reagent resources and low treatment accuracy of sewage treatment.

[0006] In view of the above problems, the present application provides an intelligent water treatment method and system for ship exhaust gas.

[0007] In the first aspect, the present application provides an intelligent water treatment method for ship exhaust gas, which is implemented by an intelligent water treatment system for ship exhaust gas, wherein the method comprises: reading the exhaust gas treatment log of the target ship, and determining the exhaust gas component set and multiple sewage image sets in the sewage solution based on the exhaust gas treatment log, wherein the exhaust gas components include the exhaust gas type and the corresponding concentration; performing feature extraction on the multiple sewage image sets in turn, determining multiple sewage feature descriptors, and establishing implicit associations between the sewage feature descriptors and the exhaust gas components; building an exhaust gas component identification channel based on the implicit association; and at a preset node, through a CCD image The sensor collects real-time sewage images in the neutralization treatment cabinet, transfers the real-time sewage images to the exhaust gas component identification channel, and outputs component identification data; calculates the component deviation between the component identification data and the adjacent historical component data, and with the purpose of satisfying the component deviation and the flocculant type as a constraint, performs simulation optimization of flocculant addition parameters to determine the optimal addition parameters; reads the working frequency of the dosing pump, corrects the working frequency according to the optimal addition parameters, and determines the corrected working frequency; sends the sewage solution in the neutralization treatment cabinet into the flotation cabinet through the water supply pump, and controls the flocculant addition of the flotation cabinet based on the corrected working frequency.

[0008] In a second aspect, the present application also provides an intelligent water treatment system for ship exhaust gas, which is used to execute an intelligent water treatment method for ship exhaust gas as described in the first aspect, wherein the system includes: a sewage image set determination module, the sewage image set determination module is used to read the exhaust gas treatment log of the target ship, and determine the exhaust gas component set and multiple sewage image sets in the sewage solution based on the exhaust gas treatment log, wherein the exhaust gas components include exhaust gas types and corresponding concentrations; an implicit association establishment module, the implicit association establishment module is used to extract features from the multiple sewage image sets in turn, determine multiple sewage feature descriptors, and establish implicit associations between sewage feature descriptors and exhaust gas components; an exhaust gas component identification channel construction module, the exhaust gas component identification channel construction module is used to build an exhaust gas component identification channel based on the implicit association; a component identification data output module, the component identification data output module is used to Under the preset node, the real-time sewage image in the neutralization treatment cabinet is collected by the CCD image sensor, and the real-time sewage image is transferred to the exhaust gas component identification channel to output component identification data; the optimal addition parameter determination module, the optimal addition parameter determination module is used to calculate the component deviation between the component identification data and the adjacent historical component data, and to meet the component deviation as the purpose, with the flocculant type as the constraint, to simulate and optimize the flocculant addition parameters to determine the optimal addition parameters; the correction working frequency determination module, the correction working frequency determination module is used to read the working frequency of the dosing pump, correct the working frequency according to the optimal addition parameter, and determine the correction working frequency; the flocculant addition control module, the flocculant addition control module is used to send the sewage solution in the neutralization treatment cabinet into the flotation cabinet through the water supply pump, and control the flocculant addition of the flotation cabinet based on the correction working frequency.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] By reading the exhaust gas treatment log of the target ship, the exhaust gas component set and multiple sewage image sets in the sewage solution are determined based on the exhaust gas treatment log, wherein the exhaust gas components include the exhaust gas type and the corresponding concentration; the multiple sewage image sets are sequentially feature extracted to determine multiple sewage feature descriptors, and an implicit association between the sewage feature descriptors and the exhaust gas components is established; an exhaust gas component identification channel is built based on the implicit association; at a preset node, the real-time sewage image in the neutralization treatment cabinet is collected by a CCD image sensor, and the real-time sewage image is transferred to the exhaust gas component identification channel, and component identification data is output; the component deviation between the component identification data and the adjacent historical component data is calculated, and with the purpose of satisfying the component deviation and the flocculant type as a constraint, the exhaust gas component identification channel is identified. The flocculant addition parameters are simulated and optimized to determine the optimal addition parameters; the working frequency of the dosing pump is read, and the working frequency is corrected according to the optimal addition parameters to determine the corrected working frequency; the sewage solution in the neutralization treatment tank is sent to the flotation tank through the water supply pump, and the flocculant addition of the flotation tank is controlled based on the corrected working frequency. That is to say, the actual state of the sewage solution is intelligently analyzed for the drug concentration, and then the drug addition is intelligently controlled to improve the accuracy of the drug addition. At the same time, the drug addition flow of the dosing pump can be automatically controlled according to the ideal drug concentration, thereby improving the rationality and accuracy of resource use, avoiding resource waste, achieving the technical goal of improving the automation and intelligence of water treatment equipment, and achieving the technical effect of improving the quality and efficiency of ship exhaust water treatment.

[0011] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented according to the contents of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0013] Figure 1 A schematic diagram of a process for an intelligent water treatment method for ship exhaust gas in this application;

[0014] Figure 2This application discloses a structural schematic diagram of an intelligent water treatment system for ship exhaust gas.

[0015] Description of reference numerals:

[0016] Sewage image set determination module 11, implicit association establishment module 12, waste gas component identification channel construction module 13, component identification data output module 14, optimal addition parameter determination module 15, correction working frequency determination module 16, flocculant addition control module 17. DETAILED DESCRIPTION

[0017] This application provides an intelligent water treatment method and system for ship exhaust gas, which solves the technical problems in the prior art that the low degree of intelligence in the process of controlling the dosage of the agent leads to low accuracy in monitoring the actual state of the sewage solution, low accuracy in controlling the dosage, waste of agent resources and low accuracy in sewage treatment. The technical goal of improving the automation and intelligence of water treatment equipment is achieved, and the technical effect of improving the quality and efficiency of ship exhaust gas water treatment is achieved.

[0018] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.

[0019] Embodiment 1

[0020] Please see attached Figure 1 The present application provides an intelligent water treatment method for ship exhaust gas, wherein the method is applied to an intelligent water treatment system for ship exhaust gas, and the method specifically comprises the following steps:

[0021] Step 1: reading the exhaust gas treatment log of the target ship, and determining an exhaust gas component set and a plurality of sewage image sets in the sewage solution based on the exhaust gas treatment log, wherein the exhaust gas component includes an exhaust gas type and a corresponding concentration;

[0022] Specifically, the target ship is a ship to be subjected to exhaust gas treatment. The exhaust gas treatment log of the target ship at a historical time is read. Further, based on the exhaust gas treatment log, an exhaust gas component set in the sewage solution and a corresponding plurality of sewage image sets containing the exhaust gas component set are determined. Further, the exhaust gas component includes an exhaust gas type and a corresponding concentration. Among them, the exhaust gas type is an exhaust gas composition type.

[0023] Step 2: extracting features from the plurality of sewage image sets in sequence, determining a plurality of sewage feature descriptors, and establishing implicit associations between the sewage feature descriptors and waste gas components;

[0024] Specifically, pixel features are extracted from multiple sewage image sets in sequence, for example, pixel features such as gradient intensity and direction are extracted, and then multiple sewage feature descriptors corresponding to the multiple sewage image sets are determined by fitting. According to the correspondence between the waste gas component set and the multiple sewage image sets, an implicit association between the sewage feature descriptors corresponding to the multiple sewage image sets and the corresponding waste gas component sets is established, that is, an implicit association between each sewage feature descriptor and the corresponding waste gas component.

[0025] Step 3: Building an exhaust gas component identification channel based on the implicit association;

[0026] Specifically, a waste gas component identification channel is built, and multiple sewage feature descriptors in implicit associations are embedded in the waste gas component identification channel for identifying waste gas components and waste gas content, thereby completing the construction of the waste gas component identification channel.

[0027] Step 4: Under the preset node, the real-time sewage image in the neutralization treatment cabinet is collected by the CCD image sensor, and the real-time sewage image is transferred to the exhaust gas component identification channel to output the component identification data;

[0028] Specifically, the CCD image sensor can convert optical images into digital signals. The neutralization treatment cabinet is used to treat acid and alkali waste liquid, and can neutralize harmful substances in the acid and alkali waste liquid to meet the emission standards. Further, at a preset time node, the sewage in the neutralization treatment cabinet is collected in real time by the CCD image sensor to obtain a real-time sewage image. Among them, the preset time node, that is, the preset node is a time interval, which is obtained by a technician in this field according to the actual situation. For example, when the preset time interval is 10 minutes, the real-time image is collected according to the 10-minute time interval as the preset node. Further, the real-time sewage image is transferred to the exhaust gas component identification channel, and the real-time sewage image is identified by the similarity comparison of multiple convolutional layers. When the real-time sewage image meets the exhaust gas component identification channel, the real-time sewage image includes the sewage feature descriptor embedded in the exhaust gas component identification channel, that is, the real-time sewage image includes the exhaust gas component, and then the concentration of the exhaust gas component is obtained by the number of times the exhaust gas component identification channel is met, and then the component identification data is output.

[0029] Step 5: Calculate the component deviation between the component identification data and the adjacent historical component data, and simulate and optimize the flocculant addition parameters with the purpose of satisfying the component deviation and the flocculant type as a constraint to determine the optimal addition parameters;

[0030] Specifically, the deviation values ​​of the exhaust gas components and the concentrations of the exhaust gas components in the component identification data and the exhaust gas components and the concentrations of the exhaust gas components in the historical component data of the previous adjacent time node are calculated to determine the component deviation.

[0031] Step 6: Read the working frequency of the dosing pump, calibrate the working frequency according to the optimal addition parameters, and determine the corrected working frequency;

[0032] Specifically, the operating frequency of the dosing pump is read. For example, the operating frequency of the dosing pump DFD-06-05-L is 140 times per minute. Further, the operating frequency is corrected according to the optimal addition parameters. For example, the recommended concentration of the PAC solution is 20%, and the recommended dosing amount for raw water treatment is 50ppm. At this time, the operating frequency of the dosing pump is corrected to 44 times per minute.

[0033] Step 7: Send the sewage solution in the neutralization treatment tank into the flotation tank through a water supply pump, and control the flocculant addition of the flotation tank based on the correction working frequency.

[0034] Specifically, the sewage solution in the neutralization treatment tank is sent into the flotation tank through a water supply pump, and the flocculant is added to the flotation tank to treat the sewage solution based on the correction working frequency control.

[0035] The intelligent water treatment method for ship exhaust gas is applied to an intelligent water treatment system for ship exhaust gas, which can achieve the technical goal of improving the automation and intelligence level of water treatment equipment and achieve the technical effect of improving the quality and efficiency of ship exhaust gas water treatment.

[0036] Furthermore, the present application also includes the following steps:

[0037] selecting a first exhaust gas component and a first sewage image set including the first exhaust gas component based on the exhaust gas treatment log;

[0038] Performing image denoising on the first sewage image set by using a Gaussian filtering algorithm, and performing image local enhancement on the image denoising results to obtain a first standard image set;

[0039] A mapping association between the first exhaust gas component and the first standard image set is established, and the exhaust gas component set and the plurality of sewage image sets are obtained based on the mapping association.

[0040] Specifically, the exhaust gas components are randomly selected from the exhaust gas treatment log to obtain the first exhaust gas component, and the sewage image having the first exhaust gas component is collected as the first sewage image set.

[0041] Then, the first sewage image set is subjected to image denoising by using a Gaussian filtering algorithm, wherein a Gaussian kernel of the first sewage image set is determined, and then the first sewage image set is scanned to obtain pixels of the first sewage image set, and pixel values ​​of the first sewage image set are obtained, and weighted average calculation is performed based on the Gaussian kernel and the pixel values, and then the pixel values ​​of the original images in the first sewage image set are replaced to obtain image denoising results of the first sewage image set. Furthermore, the texture contained in the image denoising results is enhanced to obtain image local enhancement results, and then the first standard image set is obtained.

[0042] Next, any waste gas component and the corresponding sewage image set containing the any waste gas component are selected from the waste gas treatment log, and image denoising and enhancement are performed on the corresponding sewage image set to obtain the corresponding standard image set. Further, a mapping association between the first waste gas component and the first standard image set is established, and based on the mapping association relationship between the first waste gas component and the first standard image set, any waste gas component is mapped and associated with the corresponding standard image set, and then the waste gas component set and multiple sewage image sets are integrated.

[0043] The first standard image set is obtained by image denoising and image local enhancement methods, thereby improving the accuracy of obtaining the standard image set.

[0044] Furthermore, the present application also includes the following steps:

[0045] resizing the plurality of sewage image sets according to a preset aspect ratio to obtain a plurality of adjusted sewage image sets;

[0046] Selecting a first adjusted sewage image set from the plurality of adjusted sewage image sets, and dividing the first adjusted sewage image set according to a predetermined cell size to determine a first pre-processed image set;

[0047] Calculating the gradient features of the pixel points in the first preprocessed image set by using a histogram of oriented gradients, and fitting to determine a first sewage feature descriptor set;

[0048] The sewage feature descriptor with the highest frequency in the first sewage feature descriptor set is extracted and set as the first sewage feature descriptor, and is added to the multiple sewage feature descriptors.

[0049] Specifically, the preset aspect ratio is obtained by a person skilled in the art through custom setting according to actual conditions, for example, the preset aspect ratio may be 3: 2. Further, the multiple sewage image sets are resized according to the preset aspect ratio to obtain multiple unified adjusted sewage image sets.

[0050] Then, a first adjusted sewage image set is randomly selected from the plurality of adjusted sewage image sets to obtain a first adjusted sewage image set. Furthermore, the predetermined cell size is the size of the cell grid for cell grid division of the image. The predetermined cell size is obtained by a technician in this field according to the actual situation. For example, the first adjusted sewage image set is divided into an image set containing 8×8 cells according to the predetermined cell size. Furthermore, the first adjusted sewage image set is divided according to the predetermined cell size to determine a first pre-processed image set for performing image processing cell by cell.

[0051] Next, the oriented gradient histogram is used to detect the shape and texture of the image. The gradient strength and direction of each pixel in the first preprocessed image set are calculated by the oriented gradient histogram, and a histogram of statistical gradient features is constructed in each cell. Among them, the horizontal axis of the histogram represents the gradient direction, and the vertical axis represents the cumulative value of the gradient strength in the corresponding direction, so that each cell is represented as a histogram of a gradient feature set. The gradient feature sets of multiple cells are fitted to determine the first sewage feature descriptor set. Further, the frequency of occurrence of sewage feature descriptors in the first sewage feature descriptor set is counted, that is, the number of cells with the same gradient feature is counted. The sewage feature descriptor with the highest frequency in the first sewage feature descriptor set is extracted, that is, the gradient feature corresponding to the number of cells with the same gradient feature is extracted, and set as the first sewage feature descriptor. Further, the first sewage feature descriptor is obtained by processing the first adjusted sewage image set, and multiple adjusted sewage image sets are processed according to the method for obtaining the first sewage feature descriptor to obtain multiple sewage feature descriptors.

[0052] By performing feature extraction on multiple sewage image sets, multiple sewage feature descriptors are obtained, thereby improving the accuracy of the multiple sewage feature descriptors.

[0053] Furthermore, the present application also includes the following steps:

[0054] Building a network structure of an exhaust gas component identification channel based on a convolutional neural network, and embedding the sewage feature descriptor in the implicit association into a feature convolution layer of the exhaust gas component identification channel, wherein each sewage feature descriptor corresponds to a feature convolution layer;

[0055] Determine the similarity comparison operator and the preset similarity threshold of the feature convolution layer, and complete the construction of the exhaust gas component identification channel based on the similarity comparison operator and the preset similarity threshold.

[0056] Specifically, the network structure of the exhaust gas component identification channel is constructed using convolutional neural network technology to identify the exhaust gas components. The exhaust gas component identification channel has multiple feature convolution layers for feature identification. Furthermore, each sewage feature descriptor in the implicit association is embedded in the feature convolution layer of the exhaust gas component identification channel, wherein each type of sewage feature descriptor corresponds to a feature convolution layer.

[0057] Then, the similarity comparison operator and the preset similarity threshold of the feature convolution layer are determined, wherein the similarity comparison operator is an image feature similarity comparison algorithm. The preset similarity threshold is a threshold for image feature similarity comparison, and the preset similarity threshold is obtained by a technician in this field according to the actual situation. For example, the preset similarity threshold can be 80%. Further, the exhaust gas component identification channel is built based on the similarity comparison operator and the preset similarity threshold.

[0058] By building an exhaust gas component identification channel, a basis is provided for improving the identification accuracy for subsequent exhaust gas identification through the exhaust gas component identification channel.

[0059] Furthermore, the present application also includes the following steps:

[0060] Performing grid processing on the real-time sewage image according to a preset grid size to obtain a sewage grid image set;

[0061] Inputting the sewage grid image set into the exhaust gas component identification channel for similarity comparison, extracting similar image features that meet the preset similarity threshold, and counting the number of similar image features;

[0062] Transferring the number of similar image features to a component concentration predictor for prediction to determine the concentration of similar features;

[0063] The component identification data is obtained based on the similar image features and the similar feature concentrations.

[0064] Specifically, the preset grid size is the size of the grid used to grid the real-time sewage image. The preset grid size is obtained by a person skilled in the art through custom settings according to actual conditions. Further, the real-time sewage image is grid-processed according to the preset grid size, for example, the real-time sewage image is divided into images containing 16×16 grids according to the preset grid size, thereby obtaining a sewage grid image set.

[0065] Then, each sewage grid image in the sewage grid image set is input into each feature convolution layer in turn for similarity comparison. When a sewage grid image meets the preset similarity threshold of a feature convolution layer, the first similar image feature of the convolution layer is obtained. The number of times a sewage grid image meets the preset similarity threshold of each feature convolution layer is obtained as the number of similar features corresponding to each feature convolution layer, and the content of the exhaust gas component corresponding to each feature convolution layer is obtained, and then the concentration of the exhaust gas component corresponding to each feature convolution layer is obtained by subsequent calculation. The number of times a sewage grid image meets the preset similarity threshold of different feature convolution layers is counted, and the concentration of the exhaust gas components corresponding to different feature convolution layers is subsequently obtained, that is, the concentrations of multiple exhaust gas components are obtained through each sewage grid image.

[0066] Next, the component concentration predictor is a device or software that can predict the concentration of a certain component in a substance. For example, the concentration is determined by spectral analysis. Further, the number of similar image features, that is, the number of times a sewage grid image meets the preset similarity threshold of different feature convolution layers, is transferred to the component concentration predictor for prediction, and the concentration of the exhaust gas component corresponding to each similar image feature number is determined as the similar feature concentration.

[0067] Next, based on the similar image features and the corresponding similar feature concentrations, that is, the exhaust gas concentration corresponding to each exhaust gas component is obtained according to the exhaust gas components and the corresponding concentrations, and the similar image features and the corresponding similar feature concentrations are integrated as component identification data.

[0068] The real-time sewage image is identified through the exhaust gas component identification channel to obtain component identification data, thereby improving the accuracy of the obtained component identification data.

[0069] Furthermore, the present application also includes the following steps:

[0070] reading a real-time sewage volume in a neutralization treatment tank, and determining a component treatment volume based on the component deviation and the real-time sewage volume;

[0071] For the purpose of satisfying the component processing amount, a simulation process is performed in a simulation platform based on the flocculant type to determine a plurality of flocculant addition parameters, wherein the flocculant addition parameters include the flocculant type and the corresponding addition amount;

[0072] The plurality of flocculant addition parameters are comprehensively evaluated according to a preset evaluation function, and the optimal addition parameter is determined according to the evaluation result.

[0073] Specifically, the real-time sewage volume in the neutralization treatment cabinet is the capacity of the real-time sewage to be neutralized, that is, the volume of the real-time sewage. The real-time sewage volume in the neutralization treatment cabinet is read, and the quality of the sewage at the last adjacent time node deviation is determined based on the component deviation and the real-time sewage volume, that is, the deviation concentration represented by the component deviation and the volume of the sewage, that is, the component treatment volume at the last adjacent time node deviation.

[0074] Then, during the simulation process of sewage neutralization by adding flocculant types in the simulation platform, due to the deviation of the amount of flocculant type added, that is, the amount of flocculant type added is too much or too little, the deviation of the component treatment amount is generated, that is, the quality of the deviated sewage. Furthermore, in order to meet the quality of the deviated sewage at the previous time node, the flocculant type is added or reduced to the deviated quality, and then the deviation addition parameters of different flocculant types corresponding to different sewage corresponding to the multiple exhaust gas components are determined. Among them, the flocculant addition parameters are obtained by integrating the multiple flocculant types and the addition amount corresponding to each flocculant type.

[0075] Then, a comprehensive evaluation is performed on multiple flocculant addition parameters according to a preset evaluation function, and a comprehensive evaluation coefficient obtained by the comprehensive evaluation is used as an evaluation result to further determine the optimal addition parameters.

[0076] The component processing amount is obtained through the component deviation, and then the optimal addition parameters are determined according to the component processing amount, thereby further improving the accuracy of obtaining the optimal addition parameters.

[0077] Furthermore, the present application also includes the following steps:

[0078] The preset evaluation function expression is:

[0079] ;

[0080] in, is the comprehensive evaluation coefficient, , , are the first weight, the second weight and the third weight respectively, where , , The sum of is 1, and Greater than , The sum of, N is the number of flocculant types in the flocculant addition parameter, n is any flocculant type among the N flocculant types, is the amount of flocculant type n added, is the unit cost of the nth flocculant type, is the complete reaction time of the nth flocculant type, It is the composition deviation value after the complete reaction of n flocculant types.

[0081] Specifically, the expression of the preset evaluation function for comprehensive evaluation of flocculant addition parameters is:

[0082] ;

[0083] in, is the comprehensive evaluation coefficient, where The higher it is, the better the flocculant addition parameters are, and vice versa.

[0084] Further, N is the number of flocculant types in the flocculant addition parameter, and n is any flocculant type among the N flocculant types. Further, is the amount of flocculant type n added, is the unit cost of the nth flocculant type. When the amount of the nth flocculant type added is The less and the unit cost of the nth flocculant type The less, the The higher the value of The higher it is, the lower it is.

[0085] Further, is the complete reaction time of the nth flocculant type. When the complete reaction time of the nth flocculant type is shorter, The higher the value of The higher it is, the lower it is.

[0086] Further, is the component deviation value after the complete reaction of n flocculant types, and the solution component deviation before and after the treatment of n flocculant types is obtained. When the solution component deviation is larger, it means that on the basis of meeting the component treatment capacity, the quality of the sewage that can be additionally neutralized can generate a deviation, that is, when fewer flocculant types or flocculant masses are used for sewage treatment, the better the sewage treatment effect is obtained, and the component deviation value after the complete reaction of n flocculant types is larger. The larger the value, the smaller the value. Therefore, the component deviation value after the complete reaction of n flocculant types is The larger the comprehensive evaluation coefficient The higher it is, the lower it is.

[0087] Further, , , are the first weight, the second weight and the third weight respectively, where , , The sum of is 1, and Greater than , sum.

[0088] A comprehensive evaluation of multiple flocculant addition parameters is performed through a preset evaluation function, thereby improving the accuracy of obtaining the optimal addition parameters.

[0089] Furthermore, the present application also includes the following steps:

[0090] As the flocculant is added, the floating oil on the top of the flotation tank is sensed and monitored based on an oil content meter to determine real-time optical sensing data, wherein the oil content meter is constructed based on light refraction technology and includes M light sources at different angles, where M is an integer greater than or equal to 3;

[0091] When the real-time light sensing data does not meet the preset light threshold, an overflow warning signal is generated, and the booster pump and the liquid level control valve are closed based on the overflow warning signal.

[0092] Specifically, the oil content meter is used to detect whether the oil content in the water meets the discharge standards. As the flocculant is added, the oil content meter senses and monitors the top floating oil of the flotation tank. When the monitoring obtains the oil content, it is determined as real-time optical sensing data. Furthermore, the oil content meter is constructed based on the light refraction technology. The oil content meter includes M light sources at different angles, where M is an integer greater than or equal to 3. The monitoring result obtained by the oil content meter is the light refracted after being reflected from the top floating oil, and the content of the top floating oil is measured by the amount of refracted light.

[0093] Then, when the real-time optical sensing data does not meet the preset light threshold, it means that there are too many turbid substances, resulting in a low degree of satisfaction for the required three light source refraction angles, and the accuracy of oil content monitoring through refracted light is low, or even impossible to monitor oil content, and then an overflow warning signal is generated to warn of excessive turbid substances. Further, when the overflow warning signal is generated, the booster pump is turned off to stop the sewage from flowing in, and the liquid level control valve is closed to allow the turbid substances to settle.

[0094] The flocculant addition of the flotation tank is controlled by correcting the working frequency control, thereby improving the accuracy of sewage treatment.

[0095] In summary, the intelligent water treatment method for ship exhaust gas provided in this application has the following technical effects:

[0096] By reading the exhaust gas treatment log of the target ship, the exhaust gas component set and multiple sewage image sets in the sewage solution are determined based on the exhaust gas treatment log, wherein the exhaust gas components include the exhaust gas type and the corresponding concentration; the multiple sewage image sets are sequentially feature extracted to determine multiple sewage feature descriptors, and an implicit association between the sewage feature descriptors and the exhaust gas components is established; an exhaust gas component identification channel is built based on the implicit association; at a preset node, the real-time sewage image in the neutralization treatment cabinet is collected by a CCD image sensor, and the real-time sewage image is transferred to the exhaust gas component identification channel, and component identification data is output; the component deviation between the component identification data and the adjacent historical component data is calculated, and with the purpose of satisfying the component deviation and the flocculant type as a constraint, the exhaust gas component identification channel is identified. The flocculant addition parameters are simulated and optimized to determine the optimal addition parameters; the working frequency of the dosing pump is read, and the working frequency is corrected according to the optimal addition parameters to determine the corrected working frequency; the sewage solution in the neutralization treatment tank is sent to the flotation tank through the water supply pump, and the flocculant addition of the flotation tank is controlled based on the corrected working frequency. That is to say, the actual state of the sewage solution is intelligently analyzed for the drug concentration, and then the drug addition is intelligently controlled to improve the accuracy of the drug addition. At the same time, the drug addition flow of the dosing pump can be automatically controlled according to the ideal drug concentration, thereby improving the rationality and accuracy of resource use, avoiding resource waste, achieving the technical goal of improving the automation and intelligence of water treatment equipment, and achieving the technical effect of improving the quality and efficiency of ship exhaust water treatment.

[0097] Embodiment 2

[0098] Based on the same inventive concept as the intelligent water treatment method for ship exhaust gas in the aforementioned embodiment, the present application also provides an intelligent water treatment system for ship exhaust gas, please refer to the attached Figure 2 , the system comprising:

[0099] A sewage image set determination module 11, the sewage image set determination module 11 is used to read the exhaust gas treatment log of the target ship, and determine the exhaust gas component set and multiple sewage image sets in the sewage solution based on the exhaust gas treatment log, wherein the exhaust gas component includes the exhaust gas type and the corresponding concentration;

[0100] An implicit association establishing module 12, the implicit association establishing module 12 is used to extract features from the plurality of sewage image sets in sequence, determine a plurality of sewage feature descriptors, and establish implicit associations between the sewage feature descriptors and waste gas components;

[0101] An exhaust gas component identification channel building module 13, wherein the exhaust gas component identification channel building module 13 is used to build an exhaust gas component identification channel based on the implicit association;

[0102] A component identification data output module 14, which is used to collect real-time sewage images in the neutralization and treatment cabinet through a CCD image sensor at a preset node, transfer the real-time sewage images to the exhaust gas component identification channel, and output component identification data;

[0103] The optimal addition parameter determination module 15 is used to calculate the component deviation between the component identification data and the adjacent historical component data, and to simulate and optimize the flocculant addition parameters with the flocculant type as a constraint in order to meet the component deviation, so as to determine the optimal addition parameters;

[0104] A correction working frequency determination module 16, wherein the correction working frequency determination module 16 is used to read the working frequency of the dosing pump, correct the working frequency according to the optimal addition parameter, and determine the correction working frequency;

[0105] The flocculant addition control module 17 is used to send the sewage solution in the neutralization treatment tank into the flotation tank through the water supply pump, and control the flocculant addition of the flotation tank based on the correction working frequency.

[0106] Furthermore, the sewage image set determination module 11 in the system is also used for:

[0107] selecting a first exhaust gas component and a first sewage image set including the first exhaust gas component based on the exhaust gas treatment log;

[0108] Performing image denoising on the first sewage image set by using a Gaussian filtering algorithm, and performing image local enhancement on the image denoising results to obtain a first standard image set;

[0109] A mapping association between the first exhaust gas component and the first standard image set is established, and the exhaust gas component set and the plurality of sewage image sets are obtained based on the mapping association.

[0110] Furthermore, the implicit association establishing module 12 in the system is also used for:

[0111] resizing the plurality of sewage image sets according to a preset aspect ratio to obtain a plurality of adjusted sewage image sets;

[0112] Selecting a first adjusted sewage image set from the plurality of adjusted sewage image sets, and dividing the first adjusted sewage image set according to a predetermined cell size to determine a first pre-processed image set;

[0113] Calculating the gradient features of the pixel points in the first preprocessed image set by using a histogram of oriented gradients, and fitting to determine a first sewage feature descriptor set;

[0114] The sewage feature descriptor with the highest frequency in the first sewage feature descriptor set is extracted and set as the first sewage feature descriptor, and is added to the multiple sewage feature descriptors.

[0115] Furthermore, the exhaust gas component identification channel building module 13 in the system is also used for:

[0116] Building a network structure of an exhaust gas component identification channel based on a convolutional neural network, and embedding the sewage feature descriptor in the implicit association into a feature convolution layer of the exhaust gas component identification channel, wherein each sewage feature descriptor corresponds to a feature convolution layer;

[0117] Determine the similarity comparison operator and the preset similarity threshold of the feature convolution layer, and complete the construction of the exhaust gas component identification channel based on the similarity comparison operator and the preset similarity threshold.

[0118] Furthermore, the component identification data output module 14 in the system is also used for:

[0119] Performing grid processing on the real-time sewage image according to a preset grid size to obtain a sewage grid image set;

[0120] Inputting the sewage grid image set into the exhaust gas component identification channel for similarity comparison, extracting similar image features that meet the preset similarity threshold, and counting the number of similar image features;

[0121] Transferring the number of similar image features to a component concentration predictor for prediction to determine the concentration of similar features;

[0122] The component identification data is obtained based on the similar image features and the similar feature concentrations.

[0123] Further, the optimal adding parameter determination module 15 in the system is also used for:

[0124] reading a real-time sewage volume in a neutralization treatment tank, and determining a component treatment volume based on the component deviation and the real-time sewage volume;

[0125] For the purpose of satisfying the component processing amount, a simulation process is performed in a simulation platform based on the flocculant type to determine a plurality of flocculant addition parameters, wherein the flocculant addition parameters include the flocculant type and the corresponding addition amount;

[0126] The plurality of flocculant addition parameters are comprehensively evaluated according to a preset evaluation function, and the optimal addition parameter is determined according to the evaluation result.

[0127] Further, the optimal adding parameter determination module 15 in the system is also used for:

[0128] The preset evaluation function expression is:

[0129] ;

[0130] in, is the comprehensive evaluation coefficient, , , are the first weight, the second weight and the third weight respectively, where , , The sum of is 1, and Greater than , The sum of, N is the number of flocculant types in the flocculant addition parameter, n is any flocculant type among the N flocculant types, is the amount of flocculant type n added, is the unit cost of the nth flocculant type, is the complete reaction time of the nth flocculant type, It is the composition deviation value after the complete reaction of n flocculant types.

[0131] Furthermore, the flocculant addition control module 17 in the system is also used for:

[0132] As the flocculant is added, the floating oil on the top of the flotation tank is sensed and monitored based on an oil content meter to determine real-time optical sensing data, wherein the oil content meter is constructed based on light refraction technology and includes M light sources at different angles, where M is an integer greater than or equal to 3;

[0133] When the real-time light sensing data does not meet the preset light threshold, an overflow warning signal is generated, and the booster pump and the liquid level control valve are closed based on the overflow warning signal.

[0134] Each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The intelligent water treatment method for ship exhaust gas and the specific examples in the aforementioned embodiment 1 are also applicable to the intelligent water treatment system for ship exhaust gas in this embodiment. Through the aforementioned detailed description of the intelligent water treatment method for ship exhaust gas, those skilled in the art can clearly know the intelligent water treatment system for ship exhaust gas in this embodiment, so for the sake of brevity of the specification, it will not be described in detail here. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0135] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0136] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.

Claims

1. An intelligent water treatment method for ship exhaust gas, characterized in that: The method comprises: reading an exhaust gas treatment log of a target ship, and determining an exhaust gas component set and a plurality of sewage image sets in a sewage solution based on the exhaust gas treatment log, wherein the exhaust gas component includes an exhaust gas type and a corresponding concentration; Extracting features from the plurality of sewage image sets in sequence, determining a plurality of sewage feature descriptors, and establishing implicit associations between the sewage feature descriptors and waste gas components; Building an exhaust gas component identification channel based on the implicit association; At a preset node, a real-time sewage image in the neutralization treatment cabinet is collected by a CCD image sensor, and the real-time sewage image is transferred to the exhaust gas component identification channel to output component identification data; Calculating the component deviation between the component identification data and the adjacent historical component data, and, with the purpose of satisfying the component deviation and the flocculant type as a constraint, performing simulation optimization of flocculant addition parameters to determine the optimal addition parameters; Reading the operating frequency of the dosing pump, correcting the operating frequency according to the optimal addition parameters, and determining the corrected operating frequency; The sewage solution in the neutralization treatment tank is sent into the flotation tank through a water supply pump, and the flocculant addition of the flotation tank is controlled based on the correction working frequency.

2. The method according to claim 1, characterized in that The step of determining a waste gas component set and a plurality of waste gas image sets in a waste gas solution based on the waste gas treatment log comprises: selecting a first exhaust gas component and a first sewage image set including the first exhaust gas component based on the exhaust gas treatment log; Performing image denoising on the first sewage image set by using a Gaussian filtering algorithm, and performing image local enhancement on the image denoising results to obtain a first standard image set; A mapping association between the first exhaust gas component and the first standard image set is established, and the exhaust gas component set and the plurality of sewage image sets are obtained based on the mapping association.

3. The method according to claim 1, characterized in that The plurality of sewage image sets are sequentially subjected to feature extraction to determine a plurality of sewage feature descriptors, including: resizing the plurality of sewage image sets according to a preset aspect ratio to obtain a plurality of adjusted sewage image sets; Selecting a first adjusted sewage image set from the plurality of adjusted sewage image sets, and dividing the first adjusted sewage image set according to a predetermined cell size to determine a first pre-processed image set; Calculating the gradient features of the pixel points in the first preprocessed image set by using a histogram of oriented gradients, and fitting to determine a first sewage feature descriptor set; The sewage feature descriptor with the highest frequency in the first sewage feature descriptor set is extracted and set as the first sewage feature descriptor, and is added to the multiple sewage feature descriptors.

4. The method according to claim 1, characterized in that: The exhaust gas component identification channel is established based on the implicit association, including: Building a network structure of an exhaust gas component identification channel based on a convolutional neural network, and embedding the sewage feature descriptor in the implicit association into a feature convolution layer of the exhaust gas component identification channel, wherein each sewage feature descriptor corresponds to a feature convolution layer; Determine the similarity comparison operator and the preset similarity threshold of the feature convolution layer, and complete the construction of the exhaust gas component identification channel based on the similarity comparison operator and the preset similarity threshold.

5. The method according to claim 4, characterized in that The step of transferring the real-time sewage image to the exhaust gas component identification channel and outputting component identification data includes: Performing grid processing on the real-time sewage image according to a preset grid size to obtain a sewage grid image set; Inputting the sewage grid image set into the exhaust gas component identification channel for similarity comparison, extracting similar image features that meet the preset similarity threshold, and counting the number of similar image features; Transferring the number of similar image features to a component concentration predictor for prediction to determine the concentration of similar features; The component identification data is obtained based on the similar image features and the similar feature concentrations.

6. The method according to claim 1, characterized in that In order to meet the component deviation, the flocculant type is used as a constraint to simulate and optimize the flocculant addition parameters to determine the optimal addition parameters, including: reading a real-time sewage volume in a neutralization treatment tank, and determining a component treatment volume based on the component deviation and the real-time sewage volume; For the purpose of satisfying the component processing amount, a simulation process is performed in a simulation platform based on the flocculant type to determine a plurality of flocculant addition parameters, wherein the flocculant addition parameters include the flocculant type and the corresponding addition amount; The plurality of flocculant addition parameters are comprehensively evaluated according to a preset evaluation function, and the optimal addition parameter is determined according to the evaluation result.

7. The method according to claim 6, characterized in that The method further comprises: The preset evaluation function expression is: ； in, is the comprehensive evaluation coefficient, , , are the first weight, the second weight and the third weight respectively, where , , The sum of is 1, and Greater than , The sum of, N is the number of flocculant types in the flocculant addition parameter, n is any flocculant type among the N flocculant types, is the amount of flocculant type n added, is the unit cost of the nth flocculant type, is the complete reaction time of the nth flocculant type, It is the composition deviation value after the complete reaction of n flocculant types.

8. The method according to claim 1, characterized in that The step of controlling the flocculant addition of the air flotation tank based on the corrected working frequency further comprises: As the flocculant is added, the floating oil on the top of the flotation tank is sensed and monitored based on an oil content meter to determine real-time optical sensing data, wherein the oil content meter is constructed based on light refraction technology and includes M light sources at different angles, where M is an integer greater than or equal to 3; When the real-time light sensing data does not meet the preset light threshold, an overflow warning signal is generated, and the booster pump and the liquid level control valve are closed based on the overflow warning signal.

9. An intelligent water treatment system for ship exhaust gas, characterized in that: For implementing the steps of the method according to any one of claims 1 to 8, the system comprises: a sewage image set determination module, the sewage image set determination module being used to read an exhaust gas treatment log of a target ship, and determine an exhaust gas component set and a plurality of sewage image sets in a sewage solution based on the exhaust gas treatment log, wherein the exhaust gas component includes an exhaust gas type and a corresponding concentration; An implicit association establishment module, the implicit association establishment module is used to extract features from the plurality of sewage image sets in sequence, determine a plurality of sewage feature descriptors, and establish implicit associations between the sewage feature descriptors and waste gas components; An exhaust gas component identification channel building module, the exhaust gas component identification channel building module is used to build an exhaust gas component identification channel based on the implicit association; A component identification data output module, which is used to collect real-time sewage images in the neutralization and treatment cabinet through a CCD image sensor at a preset node, transfer the real-time sewage images to the exhaust gas component identification channel, and output component identification data; An optimal addition parameter determination module, the optimal addition parameter determination module is used to calculate the component deviation between the component identification data and the adjacent historical component data, and to simulate and optimize the flocculant addition parameters with the purpose of satisfying the component deviation and the flocculant type as a constraint to determine the optimal addition parameters; A correction working frequency determination module, the correction working frequency determination module is used to read the working frequency of the dosing pump, correct the working frequency according to the optimal addition parameter, and determine the correction working frequency; The flocculant addition control module is used to send the sewage solution in the neutralization treatment tank into the flotation tank through the water supply pump, and control the flocculant addition of the flotation tank based on the correction working frequency.

Images