Evaporator maintenance method, device and terminal equipment based on joint learning framework

Abstract

The present invention is applicable to the technical field of energy equipment maintenance, and provides an evaporator maintenance method, device and terminal device based on a joint learning framework, wherein the method includes: obtaining target data of the evaporator in each descaling cycle; using multiple regression prediction models to perform prediction calculations on the target data respectively to obtain multiple prediction results; establishing a joint learning task for the multiple prediction results based on the minimum sum of squares of the combined prediction errors within the descaling cycle as the goal; obtaining the weight coefficient of each prediction result according to the joint learning task; combining the weight coefficient and the corresponding prediction result to obtain the target prediction parameters for the descaling maintenance of the evaporator in the next descaling cycle. The present invention uses the establishment of a joint learning task for multiple increasing function models to obtain a more accurate new combined model, and uses the new combined model to output the target prediction parameters to improve the accuracy of the prediction and the feasibility of applicability.

Description

Technical Field

[0001] The present invention belongs to the technical field of energy equipment maintenance, and in particular relates to an evaporator maintenance method, device, terminal equipment and storage medium based on a joint learning framework. Background Art

[0002] The evaporator uses heating method to heat the solution containing non-volatile solute in the raw liquid to boiling state, so that part of the solvent is vaporized and removed, thereby increasing the solute concentration in the solvent. It is widely used in chemical industry, food industry, pharmaceutical industry and other industries.

[0003] Based on the development of Industry 4.0, predictive maintenance has become an important research field in industrial production and life. Whether it is possible to effectively and accurately judge the scaling of the evaporator and remove the scale in time is an important topic now. Predictive maintenance for evaporator scaling not only saves a lot of energy, but also improves the efficiency of the system and equipment. Reasonable arrangement of descaling avoids emergency shutdowns in extreme cases and brings benefits to the factory. At present, the site is based on post-event judgment. After the energy consumption increases significantly, it begins to arrange maintenance such as cleaning of the key parts of the evaporator according to the predetermined plan. In this way, when performing post-event maintenance, the current situation has caused the evaporator heat exchange efficiency to decrease, energy waste has become a fait accompli, and it has caused certain damage to production and operation. Therefore, there is a certain proportion of energy-saving optimization space in this link.

[0004] In addition, there are some prediction software products in the prior art, which mainly adopt an incremental model. However, in actual applications, it is found that once the evaporator is scaled, based on field experience, the steam consumption will increase significantly, but it is not known which incremental model it is. In addition, the input of many prediction software requires the composition analysis of the original liquid and the generated liquid. According to the feedback from the on-site production situation, it is difficult to perform a detailed composition analysis of the input / output solution.

[0005] Therefore, how to improve the accuracy and feasibility of predictive maintenance of evaporators is a current technical challenge. Summary of the invention

[0006] In view of this, the embodiments of the present invention provide an evaporator maintenance method, apparatus and terminal device based on a joint learning framework to solve the problem of how to improve the accuracy and feasibility of predictive maintenance of the evaporator.

[0007] According to a first aspect of an embodiment of the present invention, there is provided an evaporator maintenance method based on a joint learning framework, comprising: obtaining target data of the evaporator in each descaling cycle; using a plurality of regression prediction models to perform prediction calculations on the target data respectively to obtain a plurality of prediction results; establishing a joint learning task for the plurality of prediction results based on the goal of minimizing the sum of squares of the combined prediction errors within the descaling cycle; obtaining a weight coefficient of each of the prediction results according to the joint learning task; and combining the weight coefficient and the corresponding prediction result to obtain target prediction parameters for descaling maintenance of the evaporator in the next descaling cycle.

[0008] In some optional embodiments, a joint learning task is established for the multiple prediction results based on the goal of minimizing the sum of squares of combined prediction errors within the descaling cycle, including: obtaining the prediction error of each prediction result; and establishing a joint learning task with minimizing the sum of squares of combined prediction errors as the objective function for the prediction errors of all regression prediction models according to a joint learning algorithm.

[0009] In some optional embodiments, obtaining a weight coefficient of each of the prediction results according to the joint learning task includes: calculating an optimal solution of the objective function; and determining a weight coefficient of each regression prediction model according to the optimal solution.

[0010] In some optional embodiments, it also includes: obtaining parameter data generated when the evaporator is descaled for the first time according to the target prediction parameters; when the parameter data does not meet the preset standard value, calculating the percentage of the parameter data to the preset standard value; modifying the descaling critical value of each regression prediction model participating in the joint learning task according to the percentage to obtain a modified regression prediction model; and updating the joint learning task using the modified regression prediction model.

[0011] In some optional embodiments, calculating the percentage of the parameter data to the preset standard value includes: calculating the percentage of each parameter data to the corresponding preset standard value respectively; averaging all the calculated percentages, and taking the average value as the target percentage.

[0012] In some optional embodiments, the target data includes at least one IoT data of valve opening, environmental data, steam flow, pipeline diameter and completed liquid volume.

[0013] In some optional embodiments, the regression prediction model includes a combination of multiple models including a regression model, a linear regression model, an exponential regression model, a parametric regression model, and an elastic network regression model.

[0014] According to a second aspect of an embodiment of the present invention, there is provided an evaporator maintenance device based on a joint learning framework, comprising: a target data acquisition module, used to acquire target data of the evaporator in each descaling cycle; a parallel regression prediction module, used to use a plurality of regression prediction models to perform prediction calculations on the target data respectively to obtain a plurality of prediction results; a joint learning module, in which a user establishes a joint learning task for the plurality of prediction results based on the goal of minimizing the sum of squares of the combined prediction errors within the descaling cycle; a weight coefficient calculation module, used to acquire the weight coefficient of each of the prediction results according to the joint learning task; and a maintenance prediction module, used to combine the weight coefficient and the corresponding prediction result to obtain the target prediction parameters for the descaling maintenance of the evaporator in the next descaling cycle.

[0015] According to a third aspect of an embodiment of the present invention, a terminal device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of the method described in any one of the first aspects are implemented.

[0016] According to a fourth aspect of an embodiment of the present invention, a storage medium is provided, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method described in any one of the first aspects are implemented.

[0017] The beneficial effect of the embodiments of the present invention compared with the prior art is that the present invention uses a joint learning task for multiple incremental function models to obtain a more accurate new combination model, and then uses the new combination model to output target prediction parameters, thereby not only improving the accuracy of predictive maintenance, but also being able to add or delete the incremental function model according to actual on-site conditions, thereby improving the feasibility range of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 is a flow chart of an evaporator maintenance method based on a joint learning framework provided in Example 1;

[0020] Figure 2 is a structural schematic diagram of an evaporator maintenance device based on a joint learning framework provided in Example 2;

[0021] Figure 3It is a schematic diagram of a terminal device to which some embodiments of the evaporator maintenance method based on the joint learning framework or the evaporator maintenance device based on the joint learning framework can be applied. DETAILED DESCRIPTION

[0022] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.

[0023] In order to illustrate the technical solution of the present invention, a specific embodiment is provided below for illustration.

[0024] Embodiment 1

[0025] Figure 1 This is a flow chart of the evaporator maintenance method based on the joint learning framework provided in Example 1.

[0026] Among them, Figure 1 As shown, the evaporator maintenance method based on the joint learning framework specifically includes the following steps S101-S105.

[0027] S101, obtaining target data of the evaporator in each descaling cycle.

[0028] For example, the descaling cycle may start with the first day of normal operation of the evaporator after descaling, and end with the first day of normal operation of the evaporator after the next descaling. It should be understood that the present application does not limit the specific time start and end of the descaling cycle.

[0029] Among them, the target data may include normal operating data of the evaporator, or may also include data obtained by monitoring the evaporator, for example, data collected by monitoring the evaporator based on the Internet of Things, including but not limited to at least one of valve opening, environmental data, steam flow, pipeline diameter and completed liquid volume.

[0030] S102, using a plurality of regression prediction models to perform prediction calculations on the target data respectively to obtain a plurality of prediction results.

[0031] The regression prediction model may include, but is not limited to, a combination of multiple models in the regression model, linear regression model, exponential regression model, parametric regression model, and elastic network regression model. Of course, in practice, it may also be a combination of one or more of the aforementioned models with other regression models. The target data collected in the above steps are respectively input into each determined regression prediction model to obtain multiple different prediction results.

[0032] The regression prediction model can be initialized using the historical data of the evaporator. Exemplarily, before using multiple regression prediction models to perform prediction calculations on the target data respectively, the steps further include: collecting historical data of the evaporator; initializing parameters of the determined regression prediction model according to the historical data. For example, the example data may include but is not limited to steam consumption, finished liquid generation, descaling critical value, etc. It should be understood that the regression model that performs the prediction task is a model that has been trained in advance using the historical data of the evaporator.

[0033] S103, establishing a joint learning task for the multiple prediction results based on the goal of minimizing the sum of squares of combined prediction errors within the descaling cycle.

[0034] As for the multiple prediction results obtained by predicting the target data by the above-mentioned multiple regression prediction models, since the prediction results generated by each regression prediction model are different, that is, the difference between the prediction results of some regression prediction models and the actual value is small, that is, the error is small, while the difference between the prediction results of some regression prediction models and the actual value is large, that is, the error is large. Therefore, the errors of the prediction results of each regression prediction model are combined to minimize the sum of squares of the combined error obtained after the combination, so that the prediction results can be more accurate and closer to the objective results.

[0035] Specifically, in an example 1, the above step S103 may specifically include:

[0036] S131, obtaining the prediction error of each prediction result;

[0037] S132, establishing a joint learning task with the minimum sum of squares of combined prediction errors as the objective function for the prediction errors of all regression prediction models according to the joint learning algorithm.

[0038] Among them, the prediction results generated by each regression prediction model are equivalent to an independent data set. Therefore, the joint learning algorithm can be used to take the data set formed by each prediction result as a joint learning participant, and send the prediction error of the prediction result to the joint party (ie, the joint central server). Then, a joint learning task with the minimum sum of squares of combined prediction errors as the objective function can be established on the joint party.

[0039] S104: Obtain a weight coefficient of each prediction result according to the joint learning task.

[0040] Continuing with the above steps, the joint learning task is to establish an objective function based on the prediction errors generated by each regression prediction model to minimize the square of the combined prediction error, wherein the objective function includes the weight coefficients of each regression prediction model.

[0041] Specifically, in a second example, the step S104 may specifically include the following steps:

[0042] S141, calculating the optimal solution of the objective function;

[0043] S142: Determine the weight coefficients of each regression prediction model according to the optimal solution.

[0044] The above steps S103-S104 are described in detail below using a specific example.

[0045] For example, assuming that r different regression prediction models are used for prediction, the combined prediction model composed of these r single regression prediction models can be expressed as:

[0046]

[0047] in, is the combined forecast value at time t, y i,t is the predicted value of the i-th model at time t, w i is the weight of the i-th prediction model, which is expressed as:

[0048]

[0049] Specifically, the prediction error e of the i-th prediction model at time t is i,t It can be expressed as:

[0050] e it =y t -y it ,

[0051] Among them, y t is the actual value at time t, and the prediction error vector of the i-th prediction model is:

[0052] F i =[e i1 , e i2 ,…,e in ] T ,

[0053] Therefore, the error matrix e can be further obtained as:

[0054] e=[F1,F2,…,F r ];

[0055] And, the information error matrix E r for:

[0056]

[0057] According to the above results, we further set R r =[1, 1, …, 1] T is an r-dimensional vector whose elements are all 1, that is, the weight vector of r prediction models. Therefore, we can establish the objective function S with the minimum sum of squares of combined prediction errors:

[0058]

[0059] From this, we can get the linear programming:

[0060] min S=W T E r W

[0061]

[0062] Among them, requesting the objective function can get the optimal solution:

[0063]

[0064] Among them, W is the weight of each prediction model result.

[0065] This example obtains the weight coefficients of each regression prediction model based on the principle of minimizing the sum of squares of combined prediction errors over a period of time. It should be understood that the above example is only used to illustrate the establishment of a joint learning task with the minimum sum of squares of combined prediction errors as the objective function, and this application does not limit the specific implementation method of combined prediction.

[0066] S105, combining the weight coefficient and the corresponding prediction result to obtain a target prediction parameter for descaling maintenance of the evaporator in the next descaling cycle.

[0067] Among them, according to the weight coefficients determined for each regression prediction model in the above steps and the prediction results generated by it, the prediction results of all regression prediction models can be combined to obtain a target prediction result with the minimum combined error, that is, the target prediction parameter. Compared with a single model, this embodiment uses a joint learning algorithm to combine the prediction results generated by multiple regression prediction models, so that the error of the target prediction parameter obtained by the final combination is minimized, thereby improving the accuracy of the prediction.

[0068] In practical applications, after obtaining the target prediction parameters, the parameters of each regression prediction model can be automatically adjusted or corrected according to the maintenance results to ensure the accuracy of the prediction results.

[0069] Specifically, in a third example, the evaporator maintenance method based on the joint learning framework may further include the following steps:

[0070] S106, obtaining parameter data generated by the first use of the evaporator after descaling according to the target prediction parameter;

[0071] S107, when the parameter data does not meet the preset standard value, calculating the percentage of the parameter data to the preset standard value;

[0072] S108, modifying the descaling critical value of each regression prediction model participating in the joint learning task according to the percentage to obtain a modified regression prediction model;

[0073] S109: Update the joint learning task using the modified regression prediction model.

[0074] Specifically, in the above step S106, after the evaporator is descaled using the target prediction parameters, parameter data generated when the evaporator is used for the first time is obtained, wherein the parameter data may include but is not limited to values ​​such as steam consumption and finished liquid generation.

[0075] Correspondingly, the preset standard value in step S107 is the standard value or reference value corresponding to the parameter data. Specifically, the parameter data can be set as the standard value based on the value obtained when the new equipment is not scaled. For example, the steam consumption per unit time when the new equipment is in use, the amount of finished liquid generated, etc.

[0076] Specifically, the step S108, modifying the descaling critical value of each regression prediction model participating in the joint learning task according to the percentage, may specifically include: multiplying the percentage by the descaling critical value of the regression prediction model.

[0077] The percentage may be the overall percentage within the descaling cycle, that is, the percentage of the parameter data and the preset standard value calculated, including: calculating the percentage of each parameter data and the corresponding preset standard value respectively; averaging all the calculated percentages, and taking the average value as the target percentage.

[0078] For ease of understanding, an example is given below. For example, assuming that the original evaporator scale thickness is 10 cm, and the data after cleaning is 95% of the standard data, indicating that some scale has not been removed. Then, based on this percentage, the initial descaling threshold (i.e., descaling critical value) should be changed to 10*0.95, i.e., 9.5 cm, and then the modified value is used to update the parameters of the regression prediction model, thereby updating the joint learning task.

[0079] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0080] Embodiment 2

[0081] Under the same inventive concept, this second embodiment also provides an evaporator maintenance device based on a joint learning framework.

[0082] Specifically, Figure 2 The structural schematic diagram of the evaporator maintenance device based on the joint learning framework provided in the second embodiment is shown in the figure. The evaporator maintenance device 200 based on the joint learning framework specifically includes: a target data acquisition module 201, which is used to obtain the target data of the evaporator in each descaling cycle; a parallel regression prediction module 202, which is used to use multiple regression prediction models to perform prediction calculations on the target data respectively to obtain multiple prediction results; a joint learning module 203, in which the user establishes a joint learning task for the multiple prediction results based on the minimum sum of squares of combined prediction errors in the descaling cycle; a weight coefficient calculation module 204, which is used to obtain the weight coefficient of each prediction result according to the joint learning task; a maintenance prediction module 205, which is used to combine the weight coefficient and the corresponding prediction result to obtain the target prediction parameters for the descaling maintenance of the evaporator in the next descaling cycle.

[0083] Since the present implementation and the above-mentioned method belong to the same inventive concept and have the same specific technical features, the technical content related to the device provided in this application can refer to the above-mentioned method embodiment and will not be repeated here.

[0084] Embodiment 3

[0085] This embodiment 3 also provides a schematic diagram of a terminal device to which the above-mentioned evaporator maintenance method based on the joint learning framework or some embodiments of the evaporator maintenance device based on the joint learning framework can be applied, such as Figure 3 As shown, the terminal device 300 includes: a processor 301, a memory 302, and a computer program 303 stored in the memory 302 and executable on the processor 301. When the processor 301 executes the computer program 303, the steps in the above-mentioned embodiments of the evaporator maintenance method based on the joint learning framework are implemented, for example Figure 1 Alternatively, when the processor 301 executes the computer program 303, the functions of each module / unit in the above-mentioned embodiments of the evaporator maintenance device based on the joint learning framework are realized, for example Figure 2 Functions of modules 201 to 205 are shown.

[0086] Exemplarily, the computer program 303 may be divided into one or more modules / units, which are stored in the memory 302 and executed by the processor 301 to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of completing specific functions, which are used to describe the execution process of the computer program 303 in the terminal device 300. For example, the computer program 303 may be divided into a synchronization module, a summary module, an acquisition module, and a return module (a module in a virtual device), and the specific functions of each module are as follows: target data acquisition module 201, parallel regression prediction module 202, joint learning module 203, weight coefficient calculation module 204, and maintenance prediction module 205.

[0087] The terminal device 300 may be a computing device such as a desktop computer, a notebook, a PDA, or a cloud server. The terminal device 300 may include, but is not limited to, a processor 301 and a memory 302. Those skilled in the art will appreciate that Figure 3 It is only an example of the terminal device 300 and does not constitute a limitation of the terminal device 300. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the terminal device 300 may also include input and output devices, network access devices, buses, etc.

[0088] The processor 301 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0089] The memory 302 may be an internal storage unit of the terminal device 300, such as a hard disk or memory of the terminal device 300. The memory 302 may also be an external storage device of the terminal device 300, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the terminal device 300. Further, the memory 302 may also include both an internal storage unit of the terminal device 300 and an external storage device. The memory 302 is used to store the computer program and other programs and data required by the terminal device 300. The memory 302 may also be used to temporarily store data that has been output or is to be output.

[0090] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In actual applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the above-mentioned embodiment of the evaporator maintenance method based on the joint learning framework, which will not be repeated here.

[0091] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0092] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0093] In the embodiments provided by the present invention, it should be understood that the disclosed devices / terminal equipment and methods can be implemented in other ways. For example, the device / terminal equipment embodiments described above are only schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0094] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0095] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0096] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the evaporator maintenance method based on the joint learning framework in the above embodiment, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned evaporator maintenance method embodiments based on the joint learning framework can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device that can carry the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media does not include electrical carrier signals and telecommunication signals.

[0097] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. An evaporator maintenance method based on a joint learning framework, characterized in that: include: Obtain target data of the evaporator in each descaling cycle; The target data includes normal operating data of the evaporator or data obtained by monitoring the evaporator; Using multiple regression prediction models to perform prediction calculations on the target data respectively to obtain multiple prediction results; Establishing a joint learning task for the multiple prediction results based on the goal of minimizing the sum of squares of the combined prediction errors within the descaling cycle; Obtaining a weight coefficient of each prediction result according to the joint learning task; Combining the weight coefficient and the corresponding prediction result to obtain a target prediction parameter for descaling maintenance of the evaporator in the next descaling cycle; Also includes: Acquire parameter data generated by the first use of the evaporator after descaling according to the target prediction parameters; the parameter data includes steam consumption and finished liquid generation; When the parameter data does not meet the preset standard value, calculating the percentage of the parameter data to the preset standard value; Modify the descaling critical value of each regression prediction model participating in the joint learning task according to the percentage to obtain a modified regression prediction model; The joint learning task is updated using the modified regression prediction model.

2. The evaporator maintenance method based on the joint learning framework according to claim 1 is characterized in that: A joint learning task is established for the multiple prediction results based on the minimum sum of squares of the combined prediction errors within the descaling cycle as the goal, including: Get the prediction error of each prediction result; According to the joint learning algorithm, a joint learning task is established for the prediction errors of all regression prediction models with the minimum sum of squares of combined prediction errors as the objective function.

3. The evaporator maintenance method based on the joint learning framework according to claim 2 is characterized in that: Obtaining a weight coefficient of each prediction result according to the joint learning task includes: Calculating the optimal solution of the objective function; According to the optimal solution, the weight coefficients of each regression prediction model are determined.

4. The evaporator maintenance method based on the joint learning framework according to claim 1 is characterized in that: The calculating the percentage of the parameter data to the preset standard value includes: Calculate the percentage of each parameter data to the corresponding preset standard value respectively; All calculated percentages were averaged and the average was taken as the target percentage.

5. The evaporator maintenance method based on the joint learning framework according to any one of claims 1 to 4, characterized in that: The target data includes at least one IoT data selected from valve opening, environmental data, steam flow, pipeline diameter and completed liquid volume.

6. The evaporator maintenance method based on the joint learning framework according to any one of claims 1 to 4, characterized in that: The regression prediction model includes a combination of multiple models including a regression model, a linear regression model, an exponential regression model, a parameter regression model and an elastic network regression model.

7. An evaporator maintenance device based on a joint learning framework, characterized in that: include: A target data acquisition module, used to acquire target data of the evaporator in each descaling cycle; the target data includes normal operating data of the evaporator or data acquired by monitoring the evaporator; A parallel regression prediction module is used to use multiple regression prediction models to perform prediction calculations on the target data respectively to obtain multiple prediction results; A joint learning module, in which a user establishes a joint learning task for the multiple prediction results based on the minimum sum of squares of the combined prediction errors within the descaling cycle as the goal; A weight coefficient calculation module, used to obtain the weight coefficient of each prediction result according to the joint learning task; A maintenance prediction module, used for combining the weight coefficient and the corresponding prediction result to obtain a target prediction parameter for descaling maintenance of the evaporator in the next descaling cycle; The maintenance prediction module is also used to: obtain parameter data generated when the evaporator is descaled for the first time according to the target prediction parameters, the parameter data including steam consumption and finished liquid generation; when the parameter data does not meet the preset standard value, calculate the percentage of the parameter data to the preset standard value; modify the descaling critical value of each regression prediction model participating in the joint learning task according to the percentage to obtain a modified regression prediction model; and update the joint learning task using the modified regression prediction model.

8. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images