Dosing control system in water treatment process

Abstract

The present disclosure relates to the technical field of water treatment, and provides a dosing control system in a water treatment process. The system comprises: a programmable logic controller configured to transmit current characteristic values of water inlet parameter characteristics and water outlet parameter characteristics collected by sensors to an edge gateway, and drive a camera to capture a floc image; the camera is configured to transmit the captured floc image to the edge gateway; the edge gateway is configured to obtain a dosing amount prediction value based on a preset dosing amount prediction model according to the water inlet parameter characteristics, the water outlet parameter characteristics and the floc image, and feed back the obtained dosing amount prediction value to the programmable logic controller; and the programmable logic controller is further configured to control the dosing pump to perform a dosing operation in the water treatment process according to the dosing amount prediction value. The present scheme can improve the dosing control efficiency of the water treatment process.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of water treatment, and particularly relates to a dosing control system in a water treatment process. BACKGROUND

[0002] The conventional dosing control system in a water treatment process mainly comprises a sensor, a programmable logic controller or an industrial computer, and a dosing pump.

[0003] Taking a PAC (Polyaluminum Chloride) dosing system as an example, the water quality parameters are collected by a sensor, the water quality parameters collected by the sensor are received by a programmable logic controller or an industrial computer, the corresponding dosing amount is predicted according to the received water quality parameters, and then the dosing pump is driven to operate at a corresponding frequency according to the predicted dosing amount.

[0004] However, the conventional water treatment process predicts the dosing amount by using a programmable logic controller or an industrial computer, and due to the limited computing resources, only some simple prediction methods such as linear regression can be used to predict the dosing amount, which results in insufficient accuracy of the dosing amount prediction, and the programmable logic controller or the industrial computer is used for both the dosing amount prediction and the dosing pump control, which increases the burden of the equipment and reduces the overall working efficiency of the equipment, thereby affecting the dosing control efficiency in the water treatment process.

[0005] It should be noted that the information disclosed in the above background section is only used to strengthen the understanding of the background of the present disclosure, and therefore can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0006] The purpose of the present disclosure is to provide a dosing control system in a water treatment process, thereby at least to some extent improving the dosing pump control efficiency and accuracy in the water treatment process.

[0007] Other characteristics and advantages of the present disclosure will become apparent from the following detailed description, or will be learned by practice of the present disclosure.

[0008] According to one aspect of the present disclosure, a dosing control system in a water treatment process is provided, comprising: a programmable logic controller configured to transmit current characteristic values of water inlet parameter characteristics and water outlet parameter characteristics collected by sensors to an edge gateway, and drive a camera to take a floc image; the camera is configured to take the floc image in response to the driving instruction sent by the programmable logic controller, and transmit the taken floc image to the edge gateway; the edge gateway is configured to obtain a dosing amount prediction value based on a preset dosing amount prediction model according to the water inlet parameter characteristics, the water outlet parameter characteristics and the floc image, and feed back the obtained dosing amount prediction value to the programmable logic controller; the programmable logic controller is further configured to control a dosing pump to perform dosing operation in the water treatment process according to the dosing amount prediction value fed back by the edge gateway.

[0009] From the above technical solutions, the dosing control system in the water treatment process in the present disclosure has at least the following advantages and positive effects:

[0010] In the technical solutions provided in some embodiments of the present disclosure, in the water treatment process, the programmable logic controller is responsible for sending the characteristic values of the water inlet parameter characteristics and the water outlet parameter characteristics collected by the sensors to the edge gateway, and is responsible for driving the camera to take the floc image. The camera can also send the taken floc image to the edge gateway. Then, the edge gateway predicts the dosing amount according to the water inlet parameter characteristic values, the water outlet parameter characteristics and the floc image, and feeds back the predicted dosing amount to the programmable logic controller. The programmable logic controller adjusts and controls the operating frequency of the dosing pump according to the predicted dosing amount. Compared with the related art, on the one hand, the present disclosure splits the core computing task in the water treatment process and assigns it to the edge gateway and the programmable logic controller, which can reduce the computing pressure of the programmable logic controller and the edge gateway, and overall improve the computing efficiency, thereby assisting to improve the dosing control efficiency in the water treatment process and reduce the dosing delay in the water treatment process. On the other hand, the present disclosure migrates the complex dosing amount prediction process to the edge gateway, which can improve the accuracy and reliability of the dosing prediction based on the powerful computing capability of the edge gateway.

[0011] It should be understood that the foregoing general description and the following detailed description are only exemplary and explanatory, and cannot limit the present disclosure. BRIEF DESCRIPTION OF DRAWINGS

[0012] The drawings herein are incorporated into the specification and form part of the specification, show embodiments consistent with the present disclosure, and together with the specification serve to explain the principles of the present disclosure. Obviously, the drawings in the following description are only some embodiments of the present disclosure, and other drawings can be obtained by those skilled in the art without creative labor on the basis of these drawings.

[0013] Figure 1 Fig. 1 shows a schematic diagram of an architecture of a dosing control system in a water treatment process according to an example embodiment of the present disclosure;

[0014] Figure 2 Fig. 2 shows a schematic diagram of an architecture of a dosing control system in a water treatment process according to another example embodiment of the present disclosure;

[0015] Figure 3 Fig. 3 shows a schematic diagram of a method for determining a preset dosing amount prediction model according to an example embodiment of the present disclosure;

[0016] Figure 4 Fig. 4 shows a schematic diagram of a method for determining a required operating frequency of a dosing pump according to an example embodiment of the present disclosure;

[0017] Figure 5 Fig. 5 shows a schematic diagram of an architecture of a dosing control system in a water treatment process according to still another example embodiment of the present disclosure. DETAILED DESCRIPTION

[0018] Example implementations will now be described more fully with reference to the accompanying drawings. Example implementations can be implemented in any

[0019] In the following description, numerous specific details are provided, such as examples of programming, software modules, user selections, client

[0020] The terms "one", "a", "an", and "the" as used herein mean "at least one" or "one or more" unless expressly specified otherwise. The term "plurality", as used herein, means "two or more". The term "exemplary" as used herein means "serving as an example, instance, or illustration," and should not necessarily be construed as preferred or advantageous over other examples. The term "step" as used herein, means any act or operation, or a portion thereof, that involves physical manipulation of physical quantities. The term "step" as used herein, can also include a

[0021] Further, the accompanying drawings are included to provide a further understanding of the present disclosure and are incorporated in and constitute a part of this specification. The drawings are not necessarily to scale, the same or similar reference numerals designate the same or similar parts throughout the several views, and embodiments will be described with reference to these drawings in which:

[0022] In the related art, the dosing amount prediction control is performed by a programmable logic controller or an industrial computer. Limited by the computing capability, only some simple prediction methods can be used for the dosing amount prediction, which leads to insufficient accuracy of the dosing amount prediction in some complex environments. For example, the programmable logic controller needs to perform the dosing amount prediction and control the dosing pump, which increases the computing burden of the device and affects the overall working efficiency of the device.

[0023] Figure 1 An example of a dosing control system in a water treatment process is shown to at least partially solve one or more of the above problems. Referring to Figure 1 The system can include:

[0024] A programmable logic controller 11 is configured to transmit current characteristic values of water inlet parameter characteristics and water outlet parameter characteristics collected by a sensor 12 to an edge gateway 13, and drive a camera 14 to capture a floc image.

[0025] The camera 14 is configured to capture a floc image in response to a driving instruction sent by the programmable logic controller 11, and transmit the captured floc image to the edge gateway 13.

[0026] The edge gateway 13 is configured to obtain a dosing amount prediction value based on a preset dosing amount prediction model according to the water inlet parameter characteristics, the water outlet parameter characteristics, and the floc image, and feed back the obtained dosing amount prediction value to the programmable logic controller 11.

[0027] The programmable logic controller 11 is further configured to control a dosing pump 15 to perform a dosing operation in a water treatment process according to the dosing amount prediction value fed back by the edge gateway 13.

[0028] In Figure 1In the technical solution provided by the illustrated embodiment, in the water treatment process, the programmable logic controller is responsible for sending the characteristic values of the water inlet parameter characteristics and water outlet parameter characteristics collected by the sensor to the edge gateway, and is responsible for driving the camera to take flocculation images. The camera can also send the taken flocculation images to the edge gateway. Then, the edge gateway predicts the dosing amount according to the water inlet parameter characteristic values, water outlet parameter characteristics and flocculation images, and feeds back the predicted dosing amount to the programmable logic controller. The programmable logic controller adjusts and controls the operating frequency of the dosing pump according to the predicted dosing amount. Compared with the related art, on the one hand, the core computing task in the water treatment process is split and distributed to the edge gateway and the programmable logic controller, which can reduce the computing pressure of the programmable logic controller and the edge gateway, and can improve the computing efficiency as a whole, thereby assisting in improving the dosing control efficiency in the water treatment process and reducing the dosing delay in the water treatment process. On the other hand, the complex dosing amount prediction process is migrated to the edge gateway, which can improve the accuracy and reliability of dosing prediction based on the powerful computing capability of the edge gateway.

[0029] Next, first, the specific implementation of "the programmable logic controller 11 is used to transmit the current characteristic values of the water inlet parameter characteristics and water outlet parameter characteristics collected by the sensor 12 to the edge gateway 13, and drive the camera 14 to take flocculation images" is described in detail.

[0030] In an exemplary embodiment, the programmable logic controller (PLC) is a digital operation electronic system specially designed for industrial environment. It stores instructions in programmable memory to accurately manage automation tasks such as mechanical action, process control, device state monitoring, etc. in the production process.

[0031] For example, the sensor 12 can be configured in a water tank, such as an inlet tank, a flocculation tank, a sedimentation tank, a filter tank, and an outlet tank. For example, the water inlet parameter characteristics include water inlet pH value, water inlet flow, water inlet turbidity, and water inlet temperature. The water inlet parameter detection sensor, such as temperature sensor, pH detection sensor, water inlet turbidity detection sensor, and water inlet flow detection sensor, can be configured in the water inlet tank. For example, the water outlet parameter characteristics include water outlet turbidity, water outlet pH, and water outlet suspended solids content. The water outlet turbidity detection sensor, water outlet pH detection sensor, and water outlet suspended solids content detection sensor can be configured in the water outlet tank. Of course, other sensors can also be configured in the water tank, which is not specially limited in the present exemplary embodiment.

[0032] In an exemplary embodiment, the sensors 12 of the water inlet tank and the water outlet tank can collect current characteristic values of the water inlet parameter characteristics and the water outlet parameter characteristics at regular time intervals, and then transmit the collected current characteristic values to the programmable logic controller 11, which transmits the received water inlet parameter characteristics and water outlet parameter characteristics to the edge gateway 13. The programmable logic controller 11 can also regularly drive the camera 14 in the flocculation tank to take flocculation images, and the camera 14 can transmit the taken flocculation images to the edge gateway 13. After receiving the current characteristic values of the water inlet parameter characteristics and the water outlet parameter characteristics sent by the programmable logic controller 11 and the flocculation images sent by the camera, the edge gateway 13 can obtain a predicted value of the dosing amount based on a preset dosing amount prediction model according to the current characteristic values of the water inlet parameter characteristics and the water outlet parameter characteristics and the flocculation images, and then feed back the predicted value of the dosing amount to the programmable logic controller 11. After receiving the predicted value of the dosing amount sent by the edge gateway 13, the programmable logic controller 11 can calculate a required operating frequency of the dosing pump 15 according to the built-in logic, and then send a control instruction to the dosing pump 15 to drive the dosing pump 15 to dose at the required operating frequency.

[0033] In an exemplary embodiment, as shown in Figure 2 The above-mentioned programmable logic controller 11 can include a first programmable logic controller 111 and a second programmable logic controller 112. The first programmable logic controller 111 is configured to transmit the current characteristic values of the water inlet parameter characteristics and the water outlet parameter characteristics collected by the sensors 12 to the edge gateway 13, and drive the camera to take flocculation images.

[0034] For example, the sensors 12 transmit the collected current characteristic values of the water inlet parameter characteristics and the water outlet parameter characteristics to the first programmable logic controller 111, which transmits the data collected by the sensors 12 to the edge gateway 13.

[0035] Next, the specific embodiment of the camera 14, which is configured to take flocculation images in response to the driving instruction sent by the programmable logic controller 11 and transmit the taken flocculation images to the edge gateway 13, will be described in detail.

[0036] For example, the above-mentioned camera 14 can be a smart camera, which can directly extract flocculation features from the taken flocculation images according to the built-in algorithm logic, and transmit the extracted flocculation features to the edge gateway 13. In this way, the prediction efficiency of the dosing amount can be further improved.

[0037] In an exemplary embodiment, the flocculation features can include flocculation area distribution characteristics, flocculation quantity per unit area, flocculation average particle size, etc.

[0038] For example, the floe recognition model can be pre-trained, and the floe features can be extracted by the floe recognition model. For example, a large number of historical floe images can be pre-collected, and then the historical floe images are manually labeled to label the contour or boundary of each floe in the floe image, the floe area and floe particle size corresponding to the contour or boundary, and the number of floes in the floe image as training data.

[0039] A machine learning model is trained by the training data, so that the machine learning model can identify the contour or boundary of the floe in the image, the area of each floe, and the number of floes in the image according to the identification. The existing target detection model can also be fine-tuned by the training data to train the target detection model to detect and identify the floe, thereby obtaining the floe recognition model. The floe recognition model can accurately identify the floe in the image containing the floe, and determine the area and floe particle size of the identified floe, and the number of floes in the image. Then the identification result of the floe recognition model is analyzed and processed to obtain the floe features. For example, the area distribution of the identified floe (such as the number ratio of different areas, etc.), the average floe particle size obtained by averaging all floe particle sizes, and the number of floes per unit area determined according to the number of floes in the image and the real scene area indicated by the image. Of course, the floe features can include other features such as floe morphology and edge sharpness, which are not specially limited in the present exemplary embodiment.

[0040] Of course, the camera 14 described above can also be a general camera with shooting and data transmission functions, and the camera 14 directly transmits the floe images shot to the edge gateway 13 for floe feature extraction, which is not specially limited in the present exemplary embodiment.

[0041] Next, the specific embodiment of the "edge gateway 13, configured to obtain a coagulant dosage prediction value based on a preset coagulant dosage prediction model according to the water inlet parameter feature, the water outlet parameter feature, and the floe image, and feed back the obtained coagulant dosage prediction value to the programmable logic controller 11" will be described in detail.

[0042] For example, the edge gateway is a key device in the industrial Internet of Things and edge computing architecture, and its essence is a "bridge and local computing node" connecting "edge terminal devices" and "cloud". It not only undertakes the tasks of terminal data collection and transmission, but also can realize part of data processing, calculation and control functions on the "edge side" close to the data source, avoiding the delay, bandwidth occupation and security risks caused by uploading all data to the cloud.

[0043] Exemplarily, the edge gateway can obtain the dosing amount prediction value in the following manner: the edge gateway processes the flocculation image to obtain flocculation features; according to the water inlet parameter features, the water outlet parameter features and the flocculation features, a dosing amount classification prediction model in a preset dosing amount prediction model is used to obtain a first dosing amount prediction level; according to the water inlet parameter features, the water outlet parameter features and the flocculation features, a dosing amount regression prediction model in the preset dosing amount prediction model is used to obtain a first dosing amount prediction value; and based on the overlapping relationship between the first dosing amount prediction level and the first dosing amount prediction value, the dosing amount prediction value is obtained.

[0044] In an exemplary embodiment, the water inlet parameter features, the water outlet parameter features and the flocculation features can be collectively referred to as dosing amount prediction features. That is, the dosing amount prediction features include the water inlet parameter features, the water outlet parameter features and the flocculation features.

[0045] Exemplarily, the current feature value of each dosing amount prediction feature can be respectively input into the dosing amount regression prediction model in the preset dosing amount prediction model corresponding to the dosing amount prediction feature and taking the dosing amount prediction feature as a single input, and the candidate dosing amount prediction value corresponding to each dosing amount prediction feature can be obtained according to the output of each dosing amount regression prediction model, and the first dosing amount prediction value can be obtained according to each candidate dosing amount prediction value. At the same time, the feature value of each dosing amount prediction feature can be respectively input into the dosing amount classification prediction model in the preset dosing amount prediction model corresponding to the dosing amount prediction feature and taking the dosing amount prediction feature as a single input, and the candidate dosing amount prediction level corresponding to each dosing amount prediction feature can be obtained according to the output of each dosing amount classification prediction, and the first dosing amount prediction level can be obtained according to each candidate dosing amount prediction level.

[0046] Exemplarily, Figure 3 A flowchart for illustrating a method for determining a preset dosing amount prediction model in an exemplary embodiment of the present disclosure is shown. Referring to FIG. 6, the method can include steps S310 to S370. Wherein: Figure 3

[0047] In step S310, according to the historical data corresponding to each dosing amount prediction feature and the historical dosing amount value corresponding to the historical data, a first label data set corresponding to each dosing amount prediction feature is generated.

[0048] ​For example, the characteristic value of the historical inflow parameter feature and the correct coagulant dosage value under the characteristic value can be collected to obtain a first label data set corresponding to the inflow parameter feature. The historical flocculation image is collected, and the flocculation feature is extracted from the historical flocculation image. The first label data set corresponding to the flocculation feature is obtained according to the extracted flocculation feature and the correct coagulant dosage value corresponding to the historical flocculation image. The characteristic value of the historical outflow parameter feature and the correct coagulant dosage value under the characteristic value are collected to generate a first label data set corresponding to the outflow parameter feature. In other words, one coagulant dosage prediction feature corresponds to one label data set, and the number of coagulant dosage prediction features corresponds to the number of first label data sets.

[0049] The correct coagulant dosage value can be understood as the coagulant dosage that can make the outflow water quality reach the preset standard under the characteristic value.

[0050] In an exemplary embodiment, the historical data of the coagulant dosage prediction feature can be collected, and if the coagulant dosage corresponding to the historical data is not the correct coagulant dosage value, it can be rejected, that is, the collected historical data is filtered according to whether it corresponds to a normal coagulant dosage value, thereby obtaining a first label data set corresponding to each coagulant dosage prediction feature.

[0051] In step S320, a second label data set corresponding to each coagulant dosage prediction feature is generated according to the historical data corresponding to each coagulant dosage prediction feature and the coagulant dosage level to which the historical coagulant dosage value corresponding to the historical data belongs in the different coagulant dosage levels corresponding to the coagulant dosage prediction feature.

[0052] In an exemplary embodiment, the different coagulant dosage levels corresponding to each coagulant dosage prediction feature can be determined in advance. Different coagulant dosage levels correspond to different coagulant dosage ranges.

[0053] For example, the number of levels of the coagulant dosage levels corresponding to different coagulant dosage prediction features is the same, and the levels representing the same coagulation degree have the same level identifier, such as the coagulant dosage levels corresponding to each coagulant dosage prediction feature are divided into low, medium and high levels. However, the coagulant dosage range indicated by the coagulant dosage levels of the same level corresponding to different coagulant dosage prediction features may be the same or may be different, such as the coagulant dosage range indicated by the low level coagulant dosage corresponding to the inflow parameter feature may be [a1, b1], and the coagulant dosage range indicated by the low level coagulant dosage corresponding to the flocculation feature may be [a2, b2]. Wherein, a1, a2, b1, b2 are different values.

[0054] For example, the determination of the different dosing amount levels corresponding to any dosing amount prediction feature can include: collecting historical data of the dosing amount prediction feature, and clustering the collected historical data; determining different dosing amount prediction levels according to the dosing amount indicated by the historical data in each cluster category in the clustering result.

[0055] For example, for each dosing amount prediction feature, the historical data of the dosing amount prediction feature can be collected, and the historical data can also be filtered according to whether there is a correct dosing amount value, so as to obtain the historical data of the dosing amount prediction feature. For example, the first historical data corresponds to the water inflow parameter feature, the second historical data corresponds to the flocculation feature, and the third historical data corresponds to the water outflow parameter feature. Then, the first clustering is performed on the first historical data to obtain the first clustering result, the second clustering is performed on the second historical data to obtain the second clustering result, and the third clustering is performed on the third historical data to obtain the third clustering result.

[0056] In the embodiment, the number of cluster categories of each dosing amount prediction feature is the same, that is, the K value in the clustering algorithm is the same, that is, the K values of the first clustering, the second clustering, and the third clustering are the same, for example, all are 3. The specific K value can be determined according to experience or test results, and the embodiment does not specially limit the K value. For example, different K values are taken, the clustering results of different K values are tested, and the K value with the best test effect is selected as the final K value.

[0057] For example, the size relationship of the dosing amount degree between the different dosing amount levels corresponding to different cluster categories can be determined according to the size relationship of the average value of the dosing amount indicated by the historical data in each cluster category in the clustering result; and the dosing amount range indicated by the dosing amount level corresponding to the cluster category can be determined according to the minimum value and the maximum value of the dosing amount indicated by the historical data in the cluster category.

[0058] For example, when the K value is 3, each cluster category corresponds to one level, and there are three levels in total. The average value of the correct dosing amount value corresponding to the historical data in the cluster category 1 is c1, the average value of the correct dosing amount value corresponding to the historical data in the cluster category 2 is c2, and the average value of the correct dosing amount value corresponding to the historical data in the cluster category 3 is c3. The size relationship of c1, c2, and c3 is c1 < c2 < c3. The cluster category 1 corresponds to the first dosing amount level, the cluster category 2 corresponds to the second dosing amount level, and the cluster category 3 corresponds to the third dosing amount level. The dosing amount degree of the first dosing amount level is less than that of the second dosing amount level, and the dosing amount degree of the second dosing amount level is less than that of the third dosing amount level. For example, the first dosing amount level is a low dosing amount level, the second dosing amount level is a medium dosing amount level, and the third dosing amount level is a high dosing amount level.

[0059] For example, the minimum value and the maximum value of the dosing amount in the cluster category corresponding to each dosing amount level can be used to determine the dosing amount range indicated by the dosing amount level. For example, there are 100 historical data in a certain dosing amount level, and the 100 historical data correspond to 100 dosing amount values. The interval range composed of the minimum value and the maximum value of the 100 dosing amount values is the dosing amount range indicated by the dosing amount level.

[0060] Taking the first dosing amount level, the second dosing amount level, and the third dosing amount level as examples, the dosing amount ranges corresponding to the dosing amount levels determined according to the clustering results can have overlapping intervals between the dosing ranges indicated by different dosing amount levels. In this case, manual verification and adjustment can be performed to ensure that the dosing amount ranges of adjacent dosing amount levels are continuous and do not overlap. Alternatively, the dosing amount ranges corresponding to the dosing amount levels obtained according to the clustering results can be used as initial dosing amount ranges, and the initial dosing amount ranges can be adjusted according to a first preset rule to obtain the final dosing amount ranges corresponding to the dosing amount levels.

[0061] For example, the first preset rule can include: in the case where there is an overlapping interval between the dosing amount ranges indicated by adjacent dosing amount levels, the overlapping interval is used as a buffer interval, i.e., the dosing amount of the buffer interval belongs to both adjacent dosing amount levels, and the dosing amount of the non-overlapping part belongs to the respective dosing amount level. That is, the dosing amount range indicated by each dosing amount level is composed of a non-overlapping interval and a buffer interval. In this way, when data labeling is performed, the dosing amount level of the historical data whose dosing amount value falls within the buffer interval has two labels. Subsequently, when prediction is performed, the final output result is also selected by class confidence, e.g., the class with a confidence greater than 0.9 is the final prediction class, thereby the problem of low prediction accuracy caused by unreasonable interval range division of the dosing amount level can be avoided to the greatest extent.

[0062] For example, the first preset rule can also include: the overlapping interval is evenly divided and allocated to the dosing amount ranges indicated by adjacent dosing amount levels, i.e., the overlapping interval is divided into two equal parts by the median value of the overlapping interval, and the final dosing amount range corresponding to the lower level in the adjacent dosing amount levels is from the minimum value in the initial dosing amount range corresponding to the lower level to the median value of the overlapping interval, and the final dosing amount range corresponding to the higher level in the adjacent dosing amount levels is from the median value of the overlapping interval to the maximum value in the initial dosing amount range corresponding to the higher level.

[0063] Similarly, for the case where the initial dosing amount ranges corresponding to adjacent dosing amount levels are discontinuous, the adjustment can also be based on a second preset rule.

[0064] Exemplarily, the second preset rule can comprise: taking a range interval between the initial dosing amount ranges corresponding to adjacent dosing amount grades as a buffer interval, and the buffer interval belongs to the adjacent dosing amount grades.

[0065] Exemplarily, the second preset rule can also comprise: taking a range interval between the initial dosing amount ranges corresponding to adjacent dosing amount grades as a buffer interval, and dividing the buffer interval into two equal parts by the median value of the buffer interval, and assigning the two equal parts to the adjacent dosing amount grades. That is, the final dosing amount range corresponding to the lower grade of the adjacent dosing amount grades is from the minimum value in the initial dosing amount range corresponding to the lower grade to the median value of the buffer interval, and the final dosing amount range corresponding to the higher grade of the adjacent dosing amount grades is from the median value of the buffer interval to the maximum value in the initial dosing amount range corresponding to the higher grade.

[0066] In an exemplary factual manner, if there is an overlapping interval between the dosing amount ranges indicated by the dosing amount grades that are not adjacent, such as the overlapping interval between the dosing amount ranges corresponding to the lower dosing amount grade and the higher dosing amount grade, the K value is adjusted and the clustering is performed again, and the determination of the dosing amount grades is performed again according to the clustering result. The dosing amount ranges indicated by the dosing amount grades that are not adjacent can also be adjusted based on the dosing amount range indicated by the intermediate dosing amount grade between the dosing amount grades that are not adjacent, so as to ensure that there is no overlapping interval between the dosing amount ranges indicated by the dosing amount grades that are not adjacent, such as taking the minimum value of the intermediate dosing amount grade as the upper limit value of the lower dosing amount grade among the dosing amount grades that are not adjacent, and taking the maximum value of the intermediate dosing amount grade as the lower limit value of the higher dosing amount grade among the dosing amount grades that are not adjacent.

[0067] Adjusting the initial dosing amount range based on the above-mentioned manner can ensure the rationality of the determined dosing amount range, and ensure that the dosing amount range corresponding to the dosing amount grade is continuous and there is no fault, so that the dosing amount prediction value can be more accurately determined according to the overlapping relationship between the first dosing amount prediction value and the first dosing amount prediction grade in the subsequent step.

[0068] Through the above-mentioned manner, the dosing amount range indicated by the different dosing amount grades corresponding to different dosing amount prediction features can be automatically determined according to the historical data. Compared with the manual determination of the dosing amount grade, this clustering automatic determination of the dosing amount grade not only improves the determination efficiency of the dosing amount grade, but also provides the basis for the determination of the dosing amount grade, so that the determined dosing amount grade has high explainability and reliability.

[0069] For example, after obtaining the dosing amount range indicated by the different dosing amount grades corresponding to each dosing amount prediction feature, the first label data set corresponding to each dosing amount prediction feature can be copied to generate a first label copy data set. For any first label copy data set corresponding to a dosing amount prediction feature, the correct dosing amount value label in the first label copy data set is modified to the specific dosing amount grade to which the correct dosing amount value belongs in the different dosing amount grades corresponding to the dosing amount prediction feature, thereby generating a second label data set.

[0070] For example, the first label data set A corresponding to the water inflow parameter feature is copied to generate the first label copy data set A1 corresponding to the water inflow parameter feature. According to the dosing amount range indicated by the different dosing amount grades corresponding to the water inflow parameter feature, it is determined which range in the dosing amount range indicated by the different dosing amount grades corresponding to the water inflow parameter feature the label of the training data in A1, i.e. the specific correct dosing amount value, belongs to, thereby obtaining which grade in the different dosing amount grades corresponding to the water inflow parameter feature the training data in A1 belongs to. The grade is taken as a new training label, thereby generating a second label data set.

[0071] In other words, the training data in the first label data set and the second label data set can be the same, but the labels of the two are different. The label of each training data in the first label data set is a specific dosing amount value, and the label of each training data in the second label data set is a dosing amount grade.

[0072] In step S330, a first number of candidate regression prediction models are trained based on the first label data set corresponding to each dosing amount prediction feature, respectively, to obtain a second number of dosing amount regression prediction models.

[0073] In an exemplary embodiment, the first number is an integer greater than 1. The candidate regression prediction model can include any machine learning model capable of regression prediction, such as a neural network model, a support vector machine-based regression model, etc., which is not specially limited in the present exemplary embodiment.

[0074] For example, the first number is 3. Based on the first label data set corresponding to the water inflow parameter feature, three candidate regression prediction models are trained, and according to the training results, three dosing amount regression prediction models can be obtained. Based on the first label data set corresponding to the flocculation feature, three candidate regression prediction models are trained, and according to the training results, three dosing amount regression prediction models can be obtained. Based on the first label data set corresponding to the water outflow parameter feature, three candidate regression prediction models are trained, and according to the training results, three dosing amount regression prediction models can be obtained. In total, nine dosing amount regression prediction models can be obtained. That is, the second number and the first number are in a multiple relationship, and the second number is three times the first number.

[0075] In step S340, a third number of candidate classification prediction models are trained based on the second label data set corresponding to each dosing amount prediction feature, respectively, to obtain a fourth number of dosing classification prediction models.

[0076] In an exemplary embodiment, the third number is an integer greater than 1. The candidate classification prediction model can include any machine learning model capable of classification prediction, such as a neural network classification model, a decision tree classification model, etc., which is not particularly limited in the present exemplary embodiment.

[0077] Taking the third number as 4 as an example, similarly, four dosing classification prediction models can be obtained based on each dosing amount prediction feature, and a total of twelve dosing classification prediction models can be obtained. That is, the fourth number and the third number are also in a multiple relationship, and the fourth number is also three times the third number.

[0078] In step S350, the second number of dosing regression prediction models are tested, and a target dosing regression prediction model is determined from the second number of dosing regression prediction models according to the test results.

[0079] For example, the second number of dosing regression prediction models can be tested by a test data set, such as testing the prediction accuracy of the model, and the dosing regression prediction model with a prediction accuracy greater than a preset threshold is determined as the target dosing regression prediction model. The preset threshold can be determined according to requirements or experience, which is not particularly limited in the present exemplary embodiment.

[0080] The dosing regression prediction model corresponding to each dosing amount prediction feature can also be selected, and the dosing regression prediction model with a test accuracy ranking in the top N is selected as a candidate target dosing regression prediction model, and the target dosing regression prediction model is obtained according to the union of the candidate target dosing regression prediction models corresponding to each dosing amount prediction feature. Wherein, N is greater than or equal to 1 and less than the first number.

[0081] In step S360, the fourth number of dosing classification prediction models are tested, and a target dosing classification prediction model is determined from the fourth number of dosing classification prediction models according to the test results.

[0082] Exemplarily, the specific implementation of step S360 can refer to step S350, which will not be described here.

[0083] In step S370, the preset dosing amount prediction model is determined according to the target dosing regression prediction model and the target dosing classification prediction model.

[0084] Exemplarily, the preset dosing amount prediction model can be obtained according to the combination of the target dosing amount regression prediction model and the target dosing amount classification prediction model. In other words, the preset dosing amount prediction model can be composed of a plurality of target dosing amount regression prediction models and a plurality of target dosing amount classification prediction models.

[0085] Through the steps S310 to S370 described above, in the process of generating the preset dosing amount prediction model, data labeling and training are respectively performed based on a plurality of dosing amount prediction features, so as to obtain a plurality of target dosing amount regression prediction models and a plurality of target dosing amount classification prediction models. The plurality of models can adapt to a plurality of different water quality environments, so as to ensure the accuracy and reliability of the dosing amount prediction under different water quality environments.

[0086] Exemplarily, as described above, each model in the preset dosing amount prediction model takes a single dosing amount prediction feature as input. Therefore, for each dosing amount prediction feature, the dosing amount prediction feature can be input into the dosing amount regression prediction model taking the dosing amount prediction feature as the single input, and the candidate dosing amount prediction value corresponding to the dosing amount prediction feature can be obtained according to the output of the dosing amount regression prediction model.

[0087] Taking the dosing amount regression prediction as an example, the number of dosing amount prediction values that can be predicted by each dosing amount prediction feature is determined according to the number of dosing amount regression prediction models in the preset dosing amount prediction model taking the dosing amount prediction feature as the single input. For example, if the water inlet parameter feature corresponds to two dosing amount regression prediction models, then two candidate dosing amount prediction values can be predicted by the water inlet parameter feature. In other words, the number of candidate dosing amount prediction values that can be predicted by each dosing amount prediction feature is the same as the number of target dosing amount regression prediction models corresponding to the dosing amount prediction feature in the preset dosing amount regression prediction model.

[0088] In an exemplary embodiment, a first dosing amount prediction value can be obtained according to the mean value of the candidate dosing amount prediction values corresponding to each dosing amount prediction feature, and a first dosing amount prediction level can be obtained according to the candidate dosing amount prediction level with the largest number in the candidate dosing amount prediction levels corresponding to each dosing amount prediction feature.

[0089] Exemplarily, when the first dosing amount prediction value belongs to the first dosing amount range interval indicated by the first dosing amount prediction level, the first dosing amount prediction value is determined as the dosing amount prediction value; when the first dosing amount prediction value does not belong to the first dosing amount range interval indicated by the first dosing amount prediction level, the current feature value of the dosing amount prediction feature is matched with the feature values of the dosing amount prediction features in the experience library, and the dosing amount prediction value is determined according to the dosing amount indicated by the feature value of the dosing amount prediction feature matched successfully.

[0090] For example, when the first dosing amount prediction value does not belong to the dosing amount range interval corresponding to the first dosing amount prediction level, the current characteristic value of the water inlet parameter characteristic, the current characteristic value of the water outlet parameter characteristic, and the current characteristic value of the flocculation body characteristic are combined into a first data record, the similarity of the first data record and the second data record in the experience library is calculated, and the dosing amount prediction value is determined according to the dosing amount corresponding to the second data record with the largest similarity.

[0091] Taking the first dosing amount prediction level as the medium dosing amount level as an example, the dosing amount range indicated by each dosing amount prediction characteristic corresponding to the medium dosing amount level may be different. In an exemplary embodiment, as long as the first dosing amount prediction value falls within the dosing amount range indicated by any target dosing amount prediction characteristic corresponding to the medium dosing amount level, it is considered that the first dosing amount prediction value belongs to the dosing amount range indicated by the medium dosing amount level, and if the first dosing amount prediction value does not fall within the dosing amount range indicated by any target dosing amount prediction characteristic corresponding to the medium dosing amount level, it is considered that the first dosing amount prediction value does not belong to the dosing amount range indicated by the medium dosing amount level. The target dosing amount prediction characteristic includes a dosing amount prediction characteristic corresponding to the same candidate dosing amount prediction level and the finally determined first dosing amount prediction level. For example, there are 6 models in the target dosing classification prediction model, and the prediction results of 4 models in the 6 models are medium dosing amount level, and the prediction results of 2 models are low dosing amount level, that is, the second prediction dosing amount level is medium dosing amount level, and then the dosing amount prediction characteristics input by the target dosing classification prediction model with the prediction result of medium dosing amount level are target dosing amount prediction characteristics.

[0092] In the present disclosure, dosing prediction can be performed by different kinds of models, and the prediction results of different kinds of models can be verified and constrained with each other, thereby improving the accuracy of the predicted dosing amount prediction value. At the same time, by taking each dosing amount prediction characteristic as an input, the dosing amount can be predicted from multiple angles through multiple dosing amount prediction characteristics, so that the model can adapt to multiple different water quality environments, and further improve the accuracy and reliability of the predicted dosing amount.

[0093] Next, the specific implementation of "the programmable logic controller 11 is further configured to control the dosing pump 15 to perform the dosing operation in the water treatment process according to the dosing amount prediction value fed back by the edge gateway 13" will be described in detail.

[0094] In an exemplary embodiment, the first programmable logic controller 111 is further configured to receive the predicted dosing amount from the edge gateway 13, determine a required operating frequency of the dosing pump based on the predicted dosing amount, and transmit the required operating frequency of the dosing pump to the second programmable logic controller 112; and the second programmable logic controller 112 is configured to adjust the current operating frequency of the dosing pump 15 according to the required operating frequency of the dosing pump 15 to control the dosing pump 15 to perform the dosing operation in the water treatment process.

[0095] For example, as shown in FIG. 1, the first programmable logic controller is responsible for sensor data transmission and calculation of the required operating frequency of the dosing pump, and the second programmable logic controller is responsible for adjusting and controlling the current operating frequency of the dosing pump according to the required operating frequency calculated by the first programmable logic controller. When both programmable logic controllers are normal, the two programmable logic controllers work in cooperation, which reduces the burden of the two programmable logic controllers and improves the processing efficiency of the two programmable logic controllers, thereby improving the efficiency of the dosing control and speeding up the system response speed as a whole. At the same time, the two programmable logic controllers are mutually primary and backup, and when any one of the programmable logic controllers fails, the other programmable logic controller can temporarily assume the task of the failed programmable logic controller, so that the water treatment system can continue to operate normally. Figure 2

[0096] In an exemplary embodiment, the dosing pump in the present disclosure includes a first dosing pump and a second dosing pump, and an exemplary embodiment of the dosing pump is shown in FIG. 1. Figure 4 FIG. 2 shows a flowchart of a method for determining a required operating frequency of a dosing pump according to an exemplary embodiment of the present disclosure. Referring to FIG. 2, the method can include steps S410-S430. In which: Figure 4

[0097] In step S410, a target operating frequency corresponding to the predicted dosing amount is determined according to a first mapping relationship between the dosing amount and the operating frequency of the dosing pump.

[0098] In step S420, a total equivalent operating frequency of the first dosing pump and the second dosing pump is determined as the target operating frequency, the adjustment priority of a single pump is higher than the priority of simultaneous adjustment of the first dosing pump and the second dosing pump, and the required operating frequency of the first dosing pump and the second dosing pump is within a respective safe operating frequency range, and a multi-objective optimization function pre-constructed with the dosing pump life and the dosing pump energy consumption as optimization objectives is solved, and the first operating frequency change value of the first dosing pump and the second operating frequency change value of the second dosing pump are determined according to the solving result.

[0099] ​​Exemplarily, a multi-objective optimization function can be pre-constructed with the life span of the dosing pump and the energy consumption of the dosing pump as the optimization objectives, and the demand operating frequency of the first dosing pump and the second dosing pump can be updated according to the solving result of the multi-objective optimization function.

[0100] In an exemplary embodiment, the multi-objective optimization function is determined according to a dosing pump life span optimization function and a dosing pump energy consumption optimization function.

[0101] The dosing pump life span optimization function is determined according to a first operating frequency change value of the first dosing pump, a second operating frequency change value of the second dosing pump, a first cumulative operating time length of the first dosing pump, and a second cumulative operating time length of the second dosing pump. The dosing pump energy consumption optimization function is determined according to a first operating frequency of the first dosing pump, a first relationship index between the first operating frequency and the power of the first dosing pump, a second operating frequency of the second dosing pump, and a second relationship index between the second operating frequency and the power of the second dosing pump.

[0102] For example, the dosing pump life span optimization function can be expressed as formula (1) as follows, the dosing pump energy consumption optimization function can be expressed as formula (2) as follows, and the multi-objective optimization function can be expressed as formula (3) as follows:

[0103] (1)

[0104] In formula (1), is the operating frequency adjustment weight of the first dosing pump, is the operating frequency adjustment weight of the second dosing pump, is the weight of the cumulative operating time length, is the first operating frequency change value of the first dosing pump, is the second operating frequency change value of the second dosing pump, is the cumulative operating time length of the first dosing pump, is the cumulative operating time length of the second dosing pump, is the rated life span length of the first dosing pump, is the rated life span length of the second dosing pump.

[0105] (2)

[0106] In formula (2), is the power coefficient of the first dosing pump, is the power coefficient of the second dosing pump, m is the first relationship index between the power and the frequency of the first dosing pump, n is the second relationship index between the power and the frequency of the second dosing pump, is the first operating frequency at which the first dosing pump is currently actually operating, a second running frequency currently actually running for the second chemical feeding pump, a first running frequency change value for the first chemical feeding pump, a second running frequency change value for the second chemical feeding pump.

[0107] (3)

[0108] In formula (3), a weight of the life optimization function, which can be customized according to the user's needs, such as the user needs to prioritize the protection of the equipment, that is, the life optimization is prioritized, the value is greater than 0.5 is less than 1, a weight of the energy consumption optimization function.

[0109] When solving the above multi-objective optimization function, the total equivalent running frequency of the first chemical feeding pump and the second chemical feeding pump is taken as the target running frequency, the adjustment priority of a single pump is higher than that of double pumps, and the demand running frequency of the first chemical feeding pump and the second chemical feeding pump is in the respective safe running frequency interval as the constraint condition, to obtain the first running frequency change value of the first chemical feeding pump and the second running frequency change value of the second chemical feeding pump.

[0110] In step S430, the first demand running frequency of the first chemical feeding pump is determined according to the first running frequency change value, and the second demand running frequency of the second chemical feeding pump is determined according to the second running frequency change value.

[0111] For example, after obtaining the first running frequency change value and the second running frequency change value according to the solving result of the multi-objective optimization function, the first demand frequency of the first chemical feeding pump can be determined as the sum of the current actual running frequency of the first chemical feeding pump and the first running frequency change value, for example, the current actual running frequency of the first chemical feeding pump is , the first running frequency change value is , and the first demand running frequency of the first chemical feeding pump is updated to .

[0112] In an exemplary embodiment, the solving process of the multi-objective optimization function can be accelerated by a pre-trained machine learning model. For example, a machine learning model can be trained to learn to predict the life loss and energy consumption of the chemical feeding pump according to different running frequency combinations of the first chemical feeding pump and the second chemical feeding pump based on historical running data. In the case where the running frequencies of both chemical feeding pumps must be adjusted to meet the target running frequency, the trained machine learning model can quickly determine the final running frequency combination in the double-pump running frequency combination that meets the target running frequency, thereby obtaining the solving result of the multi-objective optimization function.

[0113] By the steps S410 to S430, the optimization of the life and energy consumption of the first and second dosing pumps can be realized by the multi-objective optimization function, so as to prolong the life of the dosing pumps and reduce the energy consumption of the dosing pumps as much as possible.

[0114] In an exemplary embodiment, the method for determining the required running frequency of the dosing pump can further comprise: determining a first feature value of the corresponding remaining life prediction feature of the first dosing pump according to the first historical running data of the first dosing pump, and obtaining the first predicted remaining life of the first dosing pump based on the first feature value and the dosing pump remaining life prediction model; determining a second feature value of the corresponding remaining life prediction feature of the second dosing pump according to the second historical running data of the second dosing pump, and obtaining the second predicted remaining life of the second dosing pump based on the second feature value and the dosing pump remaining life prediction model; in the case that the difference between the first predicted remaining life and the second predicted remaining life is greater than a second preset value, taking the minimum difference between the first predicted remaining life and the second predicted remaining life as a new added optimization target, so as to update the multi-objective optimization function, and determining the required running frequency of the first and second dosing pumps according to the solving result of the updated multi-objective optimization function.

[0115] In an exemplary embodiment, the remaining life prediction feature of the dosing pump can include the total cumulative running time of the dosing pump, the distribution of the cumulative running time at each running frequency, the number of start-stop times, the load time, the number of failures, etc. Of course, it can also include other features that affect the life of the dosing pump, which is not specially limited in the present exemplary embodiment.

[0116] For example, a dosing pump remaining life prediction model can be trained in advance, and the remaining life of the first dosing pump and the remaining life of the second dosing pump are predicted by the dosing pump life prediction model every fixed time. In the case that the difference between the first predicted remaining life of the first dosing pump and the second predicted remaining life of the second dosing pump is greater than a second preset value, the minimum difference between the first predicted remaining life and the second predicted remaining life or less than the second preset value can be taken as a new added optimization target, and added to the above multi-objective optimization function, so as to update the multi-objective optimization function. The updated multi-objective optimization function can be expressed as the following formula (4):

[0117] (4)

[0118] In formula (4), pshouming 1 is the first predicted remaining life, pshouming 2 is the second predicted remaining life.

[0119] By taking the predicted residual life of the first and second dosing pumps as a newly added optimization target, the long-term cumulative error of the original multi-objective optimization function can be corrected, the accuracy of the optimization solution result can be improved, and the situation that the cumulative running time of the two dosing pumps is greatly different due to the inaccurate solution result of the original multi-objective optimization function can be avoided, thereby the situation that a single pump is excessively used can be avoided to the greatest extent, the service life of the pump is prolonged, and the hardware cost in the water treatment process is reduced.

[0120] In an exemplary embodiment, the first and second dosing pumps can also be primary and backup to each other, and when any one of the dosing pumps fails, the other dosing pump is started to perform the current dosing control. In other words, in the present disclosure, the two dosing pumps that are primary and backup to each other can be reused, and when both of the dosing pumps are normal, the two dosing pumps can be used at the same time to perform more accurate and wider range of dosing adjustment to cope with various possible emergency situations, and when any one of the dosing pumps fails, the other dosing pump can continue to perform dosing control as a backup pump to ensure the continuity of the production process.

[0121] For example, after the first programmable logic controller obtains the required running frequency of the first and second dosing pumps through the steps S410 to S430, the first programmable logic controller can send the required running frequency of the first and second dosing pumps to the second programmable logic controller, and the second programmable logic controller can adjust the current running frequency of the first and second dosing pumps to the corresponding required running frequency through PID (Proportional-Integral-Derivative) control.

[0122] For example, the second programmable logic controller is also configured to collect the actual dosing amount of the dosing pump, and send the actual dosing amount to the first programmable logic controller. The first programmable logic controller determines a correction coefficient of the required running frequency of the dosing pump according to the deviation between the actual dosing amount and the dosing amount prediction value, and corrects the required running frequency of the dosing pump calculated subsequently according to the correction coefficient.

[0123] For example, due to the aging of the dosing pump or the fluctuation of the voltage, there may be a deviation between the actual dosing amount of the dosing pump and the dosing amount indicated by the required running frequency, so the second programmable logic controller can collect the actual dosing amount of the dosing pump, such as determining the actual dosing amount of the dosing pump by a flow meter, and then feed back the actual dosing amount to the first programmable logic controller. The first programmable logic controller can compare the actual dosing amount with the predicted dosing amount, and obtain a correction coefficient of the required running frequency of the dosing pump according to the deviation between the two. For example, the actual dosing amount is 180 liters per hour, and the predicted dosing amount is 200 liters per hour, so the correction coefficient can be the quotient of 200 divided by 180, which is 1.11. In subsequent calculations, if the calculated required running frequency of the dosing pump is 30 Hz, then the corrected required running frequency of the dosing pump is 30 Hz multiplied by 1.11.

[0124] In an exemplary embodiment, the water treatment system in the present disclosure can further include a remote monitoring platform, based on which, the exemplary, Figure 5 An architecture schematic diagram of still another water treatment system in an exemplary embodiment of the present disclosure is shown.

[0125] Reference Figure 5 The dosing control system in the water treatment process in the present disclosure further includes a remote monitoring platform 16. Exemplarily, the above-mentioned edge gateway is further used to transmit the water inlet parameter features, the water outlet parameter features, the flocculation image corresponding flocculation features, and the dosing amount prediction value to the remote monitoring platform; the remote monitoring platform is used to display the water inlet parameter features, the water outlet parameter features, the flocculation image, and the dosing amount prediction value in a graphical user interface; the remote monitoring platform is further used to send the adjusted dosing amount value to the edge gateway in response to the adjustment operation of the dosing amount prediction value by the management personnel in the water treatment process; and the edge gateway is further used to send the adjusted dosing amount value to the first programmable logic controller, so that the first programmable logic controller determines the running frequency of the dosing pump according to the adjusted dosing amount prediction value.

[0126] For example, after determining the coagulant dosage prediction value, the edge gateway 13 can transmit the water inlet parameter feature corresponding to the coagulant dosage prediction value, the water outlet parameter feature, and the floc image corresponding to the floc feature to the remote monitoring platform 16, so that the remote monitoring platform 16 displays the current data, that is, the current feature values corresponding to the water inlet parameter feature and the water outlet parameter feature, and the floc feature corresponding to the floc image and the current coagulant dosage prediction value can be displayed in the graphical user interface of the remote monitoring platform 16. The management personnel can view the relevant data in the graphical user interface of the remote monitoring platform 16, and when the management personnel determines that the current coagulant dosage prediction value needs to be adjusted according to the actual situation, the coagulant dosage prediction value can be adjusted through the remote monitoring platform 16 to obtain an adjusted coagulant dosage value.

[0127] The remote monitoring platform 16 transmits the adjusted coagulant dosage value to the edge gateway 13 in response to the adjustment operation of the coagulant dosage prediction value, and the edge gateway 13 transmits it to the first programmable logic controller 111. The first programmable logic controller 111 can determine the required operating frequency of the first coagulant pump and the second coagulant pump according to the received adjusted coagulant dosage value, and then send the required operating frequency to the second programmable logic controller. The second programmable logic controller adjusts the current operating frequency of the first coagulant pump and the second coagulant pump to the re-determined required operating frequency according to the re-determined required operating frequency.

[0128] In an exemplary embodiment, the water treatment system in the present disclosure further comprises a cloud computing platform. Based on this, the remote monitoring platform is further configured to, in response to the adjustment operation of the coagulant dosage prediction value by the management personnel during the water treatment process according to the water inlet parameter feature, the water outlet parameter feature, and the floc image, generate a first data record according to the water inlet parameter feature, the water outlet parameter feature, the floc image, and the adjusted coagulant dosage value, and send the first data record to the cloud computing platform. The cloud computing platform is configured to, when the number of the first data records received by the remote monitoring platform reaches a preset value, update the preset coagulant dosage prediction model according to the first data record.

[0129] The cloud computing platform can update the preset coagulant dosage prediction model according to the actual coagulant dosage data to improve the accuracy of the coagulant dosage prediction.

[0130] In an example embodiment, the dosing control system in the water treatment process in the present disclosure can further include a smart inspection robot configured with an augmented reality glasses, the smart inspection robot performs fault inspection on each device in the dosing control system in the water treatment process through the augmented reality glasses, and marks the fault inspection result in a real-time picture taken to obtain a fault inspection picture, and transmits the fault inspection picture to the remote monitoring platform through the edge gateway; the remote monitoring platform is further configured to display the fault inspection picture.

[0131] In an example embodiment, the remote monitoring platform is further configured to transmit a fault repair instruction marked in the fault inspection picture by the augmented reality marking tool to the smart inspection robot through the edge gateway, so that the smart inspection robot repairs the fault based on the fault repair instruction.

[0132] For example, the smart inspection robot is configured with an augmented reality glasses, the smart inspection robot can perform fault inspection on each device in the water treatment site through the augmented reality glasses, for example, the robot takes a pipeline image through the augmented reality glasses, identifies the pipeline image to determine whether there is a pipeline leakage problem, etc., the augmented reality module can mark the leakage point in the picture and superimpose text description information to describe the fault information, thereby generating a fault inspection picture, returning the fault inspection picture to the remote monitoring platform, and the fault repair expert can quickly determine the current fault in the dosing control system in the water treatment process through the fault inspection picture displayed by the remote monitoring platform, and then give a fault repair instruction, such as marking the repair instruction of each fault in the fault inspection picture by the augmented reality marking tool, and sending the marked repair instruction to the smart inspection robot, the repair instruction marked by the fault repair expert can be superimposed and displayed in the form of augmented reality at the corresponding repair position of the augmented reality picture, and the on-site staff or the smart inspection robot can see the instruction and repair the fault according to the instruction.

[0133] Through the smart inspection robot configured with the augmented reality glasses, intelligent fault inspection can be realized, at the same time, the expert can guide the fault repair without going to the site, which shortens the fault repair period and improves the fault repair efficiency.

[0134] In the present disclosure, the edge gateway can predict the dosing amount based on a complex machine learning model, improve the accuracy of dosing amount control, at the same time, the edge gateway and the programmable logic controller are used for task splitting, which improves the processing efficiency of each hardware device as a whole, thereby assisting to improve the dosing control efficiency in the water treatment process, speed up the response speed of the system, and reduce the dosing delay of the water treatment process.

[0135] Moreover, it is noted that the above-described figures are only a schematic representation of the processes comprised in the method according to the exemplary embodiments of the present application, and are not intended to limit purposes. It is readily understood that the processes shown in the above-described figures do not indicate or limit the chronological order of these processes. In addition, it is readily understood that these processes can be executed, for example, synchronously or asynchronously in a plurality of modules.

[0136] It should be noted that, although several modules or units of the device for action execution are mentioned in the above detailed description, such a division is not mandatory. Indeed, according to the exemplary embodiments of the present disclosure, the features and functionalities of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functionalities of one module or unit described above can be further divided into a plurality of modules or units.

[0137] Moreover, although the various steps of the methods of the present disclosure are described in a particular order in the figures, this is not required or implied, nor is it required that all of the steps shown be performed in order to achieve the desired result. Additionally or alternatively, certain steps can be omitted, multiple steps can be combined into one step, one step can be broken into multiple steps, etc.

[0138] The exemplary embodiments of the present disclosure also provide a computer program product. The computer program product comprises a computer program which, when executed by a processor, implements the above-described method of reagent dosage prediction or the method of determining the required operating frequency of a reagent pump.

[0139] In an embodiment, the computer program product can be a tangible product containing the computer program, such as a computer-readable storage medium storing the computer program. The readable storage medium can be a storage medium based on electrical, magnetic, optical, electromagnetic, infrared, etc. signals, including but not limited to: random access memory (RAM), read-only memory (ROM), magnetic tape, floppy disk, flash memory (Flash), mechanical hard disk (HDD), solid state disk (SSD), etc. For example, the computer program product can be implemented as a non-volatile storage medium storing the computer program, such as a read-only memory, a Nand flash memory, etc.

[0140] In an embodiment, the computer program product can be an intangible product containing the computer program. For example, the computer program product can be implemented as a virtual digital product, such as an executable file, an installation package, etc. digital file storing the computer program.

[0141] The code of a computer program can be written in any form of programming language, including compiled or interpreted languages. Programming languages such as C, Java, C++, Python, etc. The program code can execute entirely on the user's computing device, partly on the user's computing device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device or entirely on the remote computing device or server. In the latter scenario, the remote computing device can be connected to the user's computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computing device, for example, through the Internet using an Internet Service Provider.

[0142] The computer program can be carried by or transmitted via electrical, magnetic, optical, electromagnetic, infrared or other signals. An electronic device can convert the signal carrying the computer program into a digital signal and then run the computer program. When the computer program is run on the electronic device, its code is used to make the electronic device perform (more specifically, the processor of the electronic device can be made to perform) the method steps of various exemplary embodiments of the present disclosure.

[0143] It should be understood that the present disclosure is not limited to the specific method steps or structural aspects already described and shown in the drawings, and various modifications and changes can be made without departing from the scope of the present disclosure. Based on the specific embodiments provided by the present disclosure, those skilled in the art will easily think of other embodiments. Therefore, the specific embodiments provided by the present disclosure are only exemplary, the scope and spirit of the present disclosure are indicated by the claims, and any variations, uses or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or conventional technical means in the technical field not disclosed by the present disclosure should be covered.

Claims

1. A chemical feed control system in a water treatment process, characterized by, The application relates to a water treatment system, which comprises the following components: a programmable logic controller, configured to transmit current characteristic values of water inlet parameter characteristics and water outlet parameter characteristics collected by sensors to an edge gateway and drive a camera to take flocculation images; the camera, configured to take flocculation images in response to driving instructions sent by the programmable logic controller and transmit the taken flocculation images to the edge gateway; the edge gateway, configured to obtain a dosing amount prediction value based on a preset dosing amount prediction model according to the water inlet parameter characteristics, the water outlet parameter characteristics and the flocculation images, and feed back the obtained dosing amount prediction value to the programmable logic controller; the programmable logic controller is further configured to control a dosing pump to perform dosing operation in a water treatment process according to the dosing amount prediction value fed back by the edge gateway; wherein the programmable logic controller comprises a first programmable logic controller and a second programmable logic controller, the first programmable logic controller is configured to transmit current characteristic values of water inlet parameter characteristics and water outlet parameter characteristics collected by sensors to an edge gateway and drive a camera to take flocculation images; the first programmable logic controller is further configured to receive the dosing amount prediction value fed back by the edge gateway, determine a required running frequency of the dosing pump based on the dosing amount prediction value, and transmit the required running frequency of the dosing pump to the second programmable logic controller; and the second programmable logic controller is configured to adjust a current running frequency of the dosing pump according to the required running frequency of the dosing pump to control the dosing pump to perform dosing operation in a water treatment process.

2. The chemical feed control system in a water treatment process according to claim 1, wherein The dosing pump comprises a first dosing pump and a second dosing pump; and the first programmable logic controller determines the required running frequency of the dosing pump based on the dosing amount prediction value by performing the following processes: determining a target running frequency corresponding to the dosing amount prediction value according to a first mapping relationship between the dosing amount and the running frequency of the dosing pump; solving a multi-objective optimization function pre-constructed with the dosing pump life and the dosing pump energy consumption as optimization objectives, taking the total equivalent running frequency of the first dosing pump and the second dosing pump as the target running frequency, taking the adjustment priority of a single pump higher than the adjustment priority of double pumps at the same time as a constraint condition, and taking the required running frequencies of the first dosing pump and the second dosing pump in respective safe running frequency intervals as constraint conditions, and determining a first running frequency change value of the first dosing pump and a second running frequency change value of the second dosing pump according to a solving result; determining a first required running frequency of the first dosing pump according to the first running frequency change value and determining a second required running frequency of the second dosing pump according to the second running frequency change value.

3. The dosing control system in a water treatment process according to claim 2, wherein The first programmable logic controller determines the required running frequency of the dosing pump based on the dosing amount prediction value by performing the following processes: determining a first characteristic value of a remaining life prediction characteristic corresponding to the first dosing pump according to first historical running data of the first dosing pump, and obtaining a first predicted remaining life of the first dosing pump based on the first characteristic value and a dosing pump remaining life prediction model; determine a second feature value of a second characteristic corresponding to the second chemical dosing pump according to second historical operation data of the second chemical dosing pump, and obtain a second predicted remaining service life of the second chemical dosing pump based on the second feature value and the chemical dosing pump remaining service life prediction model; in a case where a difference between the first predicted remaining service life and the second predicted remaining service life is greater than a second preset value, taking a minimum difference between the first predicted remaining service life and the second predicted remaining service life as a new added optimization target to update the multi-objective optimization function, and determining a required operation frequency of the first chemical dosing pump and the second chemical dosing pump according to a solving result of the updated multi-objective optimization function.

4. The dosing control system in a water treatment process according to claim 1, wherein The second programmable logic controller is further configured to collect an actual chemical dosing amount of the chemical dosing pump, and send the actual chemical dosing amount to the first programmable logic controller, and the first programmable logic controller is configured to determine a correction coefficient of the required operation frequency of the chemical dosing pump according to a deviation between the actual chemical dosing amount and the chemical dosing amount prediction value, and correct the required operation frequency of the chemical dosing pump in subsequent calculation according to the correction coefficient.

5. The dosing control system in a water treatment process according to claim 1, wherein The system further comprises a remote monitoring platform, and the edge gateway is further configured to transmit the water inlet parameter feature, the water outlet parameter feature, the flocculation image corresponding flocculation feature, and the chemical dosing amount prediction value to the remote monitoring platform. The remote monitoring platform is configured to display the water inlet parameter feature, the water outlet parameter feature, the flocculation image corresponding flocculation feature, and the chemical dosing amount prediction value.

6. The chemical feed control system in a water treatment process according to claim 5, wherein The remote monitoring platform is further configured to send an adjusted chemical dosing amount value to the edge gateway in response to an adjustment operation on the displayed chemical dosing amount prediction value by a manager in the water treatment process. The edge gateway is further configured to send the received adjusted chemical dosing amount value to the first programmable logic controller, so that the first programmable logic controller determines the operation frequency of the chemical dosing pump according to the adjusted chemical dosing amount prediction value.

7. The dosing control system in a water treatment process according to claim 5, wherein The system further comprises a cloud computing platform, and the remote monitoring platform is further configured to generate a first data record according to the water inlet parameter feature, the water outlet parameter feature, the flocculation image, and the adjusted chemical dosing amount value in response to an adjustment operation on the chemical dosing amount prediction value by a manager in the water treatment process according to the water inlet parameter feature, the water outlet parameter feature, and the flocculation image, and send the first data record to the cloud computing platform. The cloud computing platform is configured to update the preset chemical dosing amount prediction model according to the first data record in a case where a quantity of the first data record received by the remote monitoring platform reaches a preset value.

8. The dosing control system in a water treatment process according to claim 5, wherein The system further comprises an intelligent inspection robot configured with an augmented reality glasses, the intelligent inspection robot performs fault inspection on each device in the chemical dosing control system in the water treatment process through the augmented reality glasses, marks a fault inspection result in a real-time picture taken to obtain a fault inspection picture, and transmits the fault inspection picture to the remote monitoring platform through the edge gateway, and the remote monitoring platform is further configured to display the fault inspection picture.

9. The dosing control system in a water treatment process according to claim 8, wherein, The remote monitoring platform is further configured to transmit, to the intelligent inspection robot, a fault repair instruction marked in the fault inspection picture by the augmented reality marking tool through the edge gateway, so that the intelligent inspection robot performs fault repair based on the fault repair instruction.

Images