Resistor granulation monitoring method and system based on industrial Internet of Things

Abstract

The invention discloses a resistor granulation monitoring method and system based on industrial Internet of Things, and relates to the technical field of granulation monitoring. A distribution graph algorithm is adopted to eliminate abnormal values ​​in the granulation process data to obtain first data; the first data is fused to obtain second data to improve the accuracy of limited measurements; a more refined granulation process monitoring model is constructed by adopting a set balance equation to comprehensively describe the particle group, reflecting the dynamic change of the particle group over time, which is helpful to realize real-time monitoring and control of the granulation process; the granulation process monitoring model is optimized by rolling optimization and feedback correction, and the optimized granulation process monitoring model is adopted to monitor the subsequent granulation process, which can realize the monitoring of complex processes, and at the same time, the granulation process can be controlled according to the feedback of the monitoring situation, so as to improve the granulation efficiency and quality.

Description

Technical Field

[0001] The present invention relates to the technical field of granulation monitoring, and in particular to a resistor granulation monitoring method and system based on industrial Internet of Things. Background Art

[0002] In the production process of resistors, when processing raw materials, it is necessary to granulate the raw materials, that is, add an appropriate amount of binder to the powder, rely on the coagulation and rotation of the wet powder to form a core of the granule, and the core grows in two ways: agglomeration and cladding, and finally forms a spherical product. Granulators are often used for granulation in existing large-scale production. Granulators have the advantages of a wide range of materials, large processing capacity and high operating rate. In the granulation process, the automatic control system can be set up to realize the automation of powder addition, liquid spraying and fine particle size control, which is of great significance to enterprises to improve production efficiency, save production costs, ensure product quality, and improve the working environment of workers.

[0003] In granulation production, after the raw materials are crushed, the moisture content is tested and meet the granulation requirements, they begin to enter the granulator, and the motor in the granulator drives the pressure wheel to squeeze the slurry in the ring die. However, in actual production, the operation of the motor in the granulator is often affected by factors such as the diversity of slurry raw materials, the difference in ring die compression ratio, and the unevenness of slurry moisture content, and often shows large state fluctuations. For example, due to inconsistent moisture content of slurry during production, the ring die may be blocked during continuous operation, which may lead to motor overload; at the same time, long-term wear of the ring die will also change the motor operation state. The emergence of the above problems will further lead to fluctuations in granulation quality, or unexpected shutdowns due to problems such as overload and stalling, causing losses to the company's continuous, automated, and efficient production.

[0004] Therefore, it is necessary to monitor the granulator module motor in real time. However, it is difficult to establish an accurate mathematical model for the granulation process in the existing monitoring technology. It is difficult for traditional control methods to monitor complex processes, resulting in the inability to monitor the granulation process in real time and the inability to understand the granulation conditions in real time. The current granulation control monitoring technology has shown several significant deficiencies in practical applications. First, there is a one-sided problem of parameter monitoring, which mainly relies on the independent analysis of established parameter indicators, but fails to effectively integrate the relationship between the parameters for comprehensive analysis, which limits the comprehensiveness and depth of monitoring; secondly, although manual inspection and regular maintenance mechanisms are necessary, due to their inherent lag, it is difficult to respond to changes in system status in a timely manner, and over-reliance on manual experience, there is a risk of subjective judgment bias; thirdly, the lack of real-time fault analysis, judgment and diagnosis prediction capabilities makes it difficult to detect and solve potential problems early, increasing uncertainty in production. Summary of the invention

[0005] The technical problem to be solved by the present invention is that it is difficult to establish an accurate mathematical model for the granulation process to monitor the granulation process, and it is impossible to understand the granulation working conditions in real time. The purpose is to provide a resistor granulation monitoring method and system based on the industrial Internet of Things. The abnormal values ​​in the granulation process data are eliminated by using a distribution graph algorithm to obtain first data; the first data are then fused to obtain second data to improve the accuracy of limited measurements; a more refined granulation process monitoring model is constructed by using a set balance equation to comprehensively describe the particle group, reflecting the dynamic changes of the particle group over time, which is helpful to realize real-time monitoring and control of the granulation process; the granulation process monitoring model is optimized by rolling optimization and feedback correction, and the optimized granulation process monitoring model is used to monitor the subsequent granulation process, which can realize the monitoring of complex processes, and at the same time, the granulation process can be controlled according to the feedback of the monitoring situation to improve the granulation efficiency and quality.

[0006] The present invention is achieved through the following technical solutions:

[0007] The first aspect of the present invention provides a resistor granulation monitoring method based on industrial Internet of Things, comprising the following specific steps:

[0008] Collecting granulation process data, and removing abnormal values ​​in the granulation process data using a distribution graph algorithm to obtain first data;

[0009] Fusing the first data based on batch estimation to obtain second data;

[0010] Based on the second data, a granulation process monitoring model is constructed using a collective balance equation;

[0011] The granulation process monitoring model is optimized according to rolling optimization and feedback correction to obtain an optimized granulation process monitoring model;

[0012] The optimized granulation process monitoring model is used to monitor the resistor granulation process.

[0013] Furthermore, the use of a distribution graph algorithm to remove abnormal values ​​in the granulation process data specifically includes:

[0014] Quartiles were used to sort the collected granulation process data;

[0015] Performing quartile dispersion calculation on the sorted granulation process data;

[0016] Determine whether there are outliers in the granulation process data based on the dispersion, and if so, remove them.

[0017] Further, determining whether there are abnormal values ​​in the granulation process data according to the discreteness specifically includes:

[0018] Determine the difference between the lower quartile and the upper quartile based on the dispersion;

[0019] According to the difference between the lower quartile and the upper quartile, it is judged whether the difference between the lower quartile and the upper quartile is within a set threshold. If not, it is determined that there are abnormal values ​​in the granulation process data, and the abnormal values ​​are eliminated.

[0020] Furthermore, the fusing the first data based on batch estimation specifically includes:

[0021] Assume that after removing the outliers from the n data measured by n sensors, there are m data left. The m data are divided into equal groups to obtain k groups, where each group contains j data.

[0022] The arithmetic mean and standard deviation of k groups of data are calculated respectively, and the data fusion result based on batch estimation is obtained based on the arithmetic mean and standard deviation of k groups of data.

[0023] Furthermore, the granulation process monitoring model is constructed based on the second data by using the collective balance equation, and the specific steps include:

[0024] Based on the second data, extract the particle spatial position vector and particle distribution characteristic vector within time t, and construct a particle population distribution equation;

[0025] According to the particle population distribution equation, the particle population distribution change rate caused by the particle size change and the particle population distribution change rate based on the spatial coordinate system are calculated;

[0026] A granulation process monitoring model is constructed based on the particle group distribution change rate caused by particle size change and the particle group distribution change rate in the spatial coordinate system.

[0027] Furthermore, the step of extracting the particle spatial position vector and the particle distribution feature vector within the time t includes:

[0028] Based on the second data, extracting the difference between the set granulation amount and the actual granulation amount during the granulation process;

[0029] Determine the difference change according to the difference between the set granulation amount and the granulation amount in the actual granulation process;

[0030] According to the difference and difference change between the set granulation amount and the actual granulation amount during the granulation process, the spatial position of the particles and the particle distribution characteristics within the time t are determined;

[0031] Based on the spatial position of the particle within time t, determine the spatial position vector of the particle within time t;

[0032] Based on the particle distribution characteristics within time t, a particle distribution feature vector within time t is determined.

[0033] Furthermore, the granulation process monitoring model is constructed, and the specific steps include:

[0034] Obtain the rate of change of particle population distribution caused by particle size change and the particle group distribution change rate based on the spatial coordinate system ;

[0035] The granulation process monitoring model is constructed by combining the collective balance equation, including:

[0036] ;

[0037] Where t represents the target granulation time, x represents the particle distribution characteristic vector, and z represents the particle spatial position vector. represents the particle distribution function at the target granulation time t, the particle distribution characteristic vector x and the particle spatial position z, represents the particle generation rate, Represents the extinction rate of particles.

[0038] Furthermore, the granulation process monitoring model is optimized according to the rolling optimization and feedback correction, and the specific steps include:

[0039] Conduct fault-tolerant control on the granulation process monitoring model and detect data loss in the granulation simulation process in real time;

[0040] Determine whether the data loss of the granulation simulation process exceeds the data loss threshold of the granulation simulation process. If it exceeds, then:

[0041] Conduct fault diagnosis on the granulation simulation process of the granulation process monitoring model, perform rolling optimization and feedback correction on the faults, and obtain continuous error feedback;

[0042] The granulation process monitoring model is corrected according to continuous error feedback to obtain the optimized granulation process monitoring model.

[0043] The second aspect of the present invention provides a resistor granulation monitoring system based on the industrial Internet of Things, which is applied to a resistor granulation monitoring method based on the industrial Internet of Things, including a granulation monitoring perception control platform, a granulation monitoring sensor network platform and a granulation monitoring management platform;

[0044] The granulation monitoring perception control platform is used to collect granulation process data and transmit it to the granulation monitoring management platform through the granulation monitoring sensor network platform;

[0045] The granulation monitoring and management platform includes a data processing module, a model building module and an analysis module;

[0046] The data processing module is used to process the granulation process data;

[0047] The model building module is used to build a granulation process monitoring model according to the collective balance equation;

[0048] The analysis module is used to optimize the granulation process monitoring model according to rolling optimization and feedback correction to obtain the optimized granulation process monitoring model, and monitor the granulation process according to the optimized granulation process monitoring model.

[0049] Furthermore, the specific steps of the data processing module processing the granulation process data include:

[0050] Using a distribution graph algorithm to remove abnormal values ​​in the granulation process data to obtain first data;

[0051] The first data is fused based on the batch estimation to obtain the second data.

[0052] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0053] The first data is obtained by eliminating abnormal values ​​in the granulation process data by using a distribution graph algorithm; the first data is fused to obtain second data to improve the accuracy of limited measurements; a more refined granulation process monitoring model is constructed by using a set equilibrium equation to comprehensively describe the particle population and reflect the dynamic changes of the particle population over time, which is helpful to achieve real-time monitoring and control of the granulation process; the granulation process monitoring model is optimized by rolling optimization and feedback correction, and the optimized granulation process monitoring model is used to monitor the subsequent granulation process, which can realize the monitoring of complex processes, and at the same time, the granulation process can be controlled according to the feedback of the monitoring situation, thereby improving the granulation efficiency and quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without creative work. In the drawings:

[0055] Figure 1 A monitoring method in an embodiment of the present invention;

[0056] Figure 2 This is the monitoring system in the embodiment of the present invention. DETAILED DESCRIPTION

[0057] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0058] As a possible implementation, Figure 1 As shown, the first aspect of this embodiment provides a resistor granulation monitoring method based on industrial Internet of Things, by adopting a distribution graph algorithm to eliminate abnormal values ​​in the granulation process data to obtain first data; the first data is fused to obtain second data to improve the accuracy of limited measurements; a more refined granulation process monitoring model is constructed by adopting a set balance equation to comprehensively describe the particle group and reflect the dynamic changes of the particle group over time, which is helpful to realize real-time monitoring and control of the granulation process; the granulation process monitoring model is optimized by rolling optimization and feedback correction, and the optimized granulation process monitoring model is used to monitor the subsequent granulation process, which can realize the monitoring of complex processes, and at the same time, the granulation process can be controlled according to the feedback of the monitoring situation to improve the granulation efficiency and quality.

[0059] In some possible embodiments, the collected granulation process data include: the granulation tower capacity of the granulator, the predicted granulation amount, the actual granulation amount, the raw material and granulation product data during granulation in the granulation tower, wherein the raw material and granulation product data during granulation in the granulation tower include the particle characteristics during granulation: particle size, shape, porosity, strength, solubility, particle size distribution and density.

[0060] In some possible implementations, the distribution graph algorithm removes outliers by using quartiles to reorganize the data structure, and obtains outliers through certain comparison calculations. Quartiles are to arrange all collected granulation process data from small to large and divide them into four equal parts, and the granulation process data at three division points are quartiles. Quartiles mainly include upper quartiles, median, and lower quartiles. First, the granulation process data are arranged in ascending order to obtain the upper quartile, median, and lower quartile. The quartile dispersion is equal to the difference between the lower quartile and the upper quartile. The outliers of the data set are estimated according to the difference threshold, and the outliers are removed.

[0061] In some possible implementations, batch estimation theory is an algorithm derived from recursive estimation theory, which is used to fuse different detection values ​​obtained for the same detection target, and can achieve more ideal results and better accuracy. Fusion of the first data based on batch estimation specifically includes: assuming that after removing abnormal values ​​from n data measured by n sensors, m data remain, and the m data are grouped as equally as possible to obtain k groups, each group of data is j; the arithmetic mean and standard deviation of the k groups of data are calculated respectively, and the data fusion result based on batch estimation is obtained based on the arithmetic mean and standard deviation. The calculation steps include:

[0062] Calculate the arithmetic mean of each group : ;

[0063] Calculate the standard deviation for each group : ;

[0064] Calculate the single group data fusion result based on the arithmetic mean and standard deviation of each group: ;

[0065] in, represents the rth data in the kth group, j is the number of data in each group, represents the arithmetic mean of the i-th group, represents the standard deviation of the i-th group, k is the number of groups, and y represents the data fusion result.

[0066] On the basis of obtaining the single group data fusion result, the data fusion results of all groups are obtained in turn based on batch estimation:

[0067]

[0068] in, and They represent the variance of the current batch estimation data fusion result and the variance of the previous batch estimation data fusion result, respectively. Indicates the last batch estimation data fusion result, It represents the fusion result of the batch estimation data, and y represents the fusion result of a single group of data.

[0069] In some possible implementations, a granulation process monitoring model is constructed based on the second data, and the specific steps include:

[0070] Based on the second data, the particles are tracked to determine the status of the particles in the granulation process, the particle spatial position vector and the particle distribution characteristic vector within the time t are extracted, and the particle population distribution equation is constructed;

[0071] According to the particle population distribution equation, the particle population distribution change rate caused by the particle size change and the particle population distribution change rate based on the spatial coordinate system are calculated;

[0072] According to the particle group distribution change rate caused by particle size change and the particle group distribution change rate based on the spatial coordinate system, a granulation process monitoring model is constructed in combination with the collective balance equation.

[0073] In some possible implementations, the particle spatial position vector and particle distribution feature vector within time t are extracted, and the specific steps include:

[0074] extracting a difference and a difference change between a given granulation amount and an actual granulation amount based on the second data;

[0075] Determine the spatial position of particles and particle distribution characteristics within time t;

[0076] Based on determining the particle spatial position and particle distribution characteristics within time t, the particle spatial position vector and particle distribution characteristic vector within time t are determined respectively.

[0077] In some possible implementations, the calculation step of constructing a granulation process monitoring model includes:

[0078] Obtain the rate of change of particle population distribution caused by particle size change and the particle group distribution change rate based on the spatial coordinate system ;

[0079] The granulation process monitoring model is constructed by combining the collective balance equation, including:

[0080] ;

[0081] Where t represents the target granulation time, x represents the particle distribution characteristic vector, and z represents the particle spatial position vector. represents the particle distribution function at the target granulation time t, the particle distribution characteristic vector x and the particle spatial position z, represents the particle generation rate, Represents the extinction rate of particles.

[0082] In some possible implementations, the granulation process monitoring model is optimized according to rolling optimization and feedback correction, and the specific steps include: performing fault-tolerant control on the granulation process monitoring model to detect data loss in the granulation process in real time; judging whether the data loss in the granulation process exceeds the data loss threshold of the granulation process, and if so: performing fault diagnosis on the granulation process of the granulation process monitoring model, performing rolling optimization and feedback correction on the fault, and obtaining continuous error feedback; and correcting the granulation process monitoring model according to the continuous error feedback to obtain an optimized granulation process monitoring model.

[0083] Among them, the optimization control is to optimize the control in the shortest adjustment time to achieve the stable transition of the granulation system to the production state that can obtain the maximum qualified granule output with the least raw material and energy consumption. Fault-tolerant control refers to the ability to complete its basic functions when the components or subsystems of the system fail. The application of fault-tolerant control theory in granulation system control is mainly for fault-tolerant control of sensors, actuators and controllers. One is to detect and identify faults with a certain accuracy; the other is to avoid or reduce the adverse effects of faults by reconstructing the control system structure on the basis of the former.

[0084] Rolling optimization and feedback correction of faults are performed to obtain continuous error feedback, which has advantages in dealing with multi-variable interactions, process constraints and optimization adjustments, making it an important method for granulation process control. Under ideal conditions, this control method can achieve long-term operation of the granulation process at the optimization target point and reduce downtime caused by instability factors of manual operation. The combination of model predictive control and intelligent control can make up for the defects of traditional predictive control methods such as low accuracy, lack of self-learning and self-organization functions, and weak robustness.

[0085] As a possible implementation, Figure 2 As shown, the second aspect of this embodiment provides a resistor granulation monitoring system based on the industrial Internet of Things. When the system is running, a resistor granulation monitoring method based on the industrial Internet of Things is used to perform granulation monitoring. The system includes a granulation monitoring perception control platform, a granulation monitoring sensor network platform and a granulation monitoring management platform;

[0086] The granulation monitoring perception control platform is used to collect granulation process data and transmit it to the granulation monitoring management platform through the granulation monitoring sensor network platform;

[0087] The granulation monitoring and management platform includes data processing module, model building module and analysis module;

[0088] The data processing module is used to process the granulation process data;

[0089] The model building module is used to build a granulation process monitoring model based on the collective balance equation;

[0090] The analysis module is used to optimize the granulation process monitoring model according to rolling optimization and feedback correction to obtain the optimized granulation process monitoring model, and monitor the granulation process according to the optimized granulation process monitoring model.

[0091] In some possible implementations, the specific steps of processing the granulation process data by the data processing module include:

[0092] A distribution graph algorithm is used to remove abnormal values ​​in the granulation process data to obtain first data; and the first data is fused based on batch estimation to obtain second data.

[0093] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A resistor granulation monitoring method based on industrial Internet of Things, characterized in that: The specific steps include: Collecting granulation process data, and removing abnormal values ​​in the granulation process data using a distribution graph algorithm to obtain first data; Fusing the first data based on batch estimation to obtain second data; Based on the second data, a granulation process monitoring model is constructed using a collective balance equation; The granulation process monitoring model is optimized according to rolling optimization and feedback correction to obtain an optimized granulation process monitoring model; The optimized granulation process monitoring model is used to monitor the resistor granulation process; Wherein, based on the second data, a granulation process monitoring model is constructed by using a collective equilibrium equation, and the specific steps include: Based on the second data, extract the particle spatial position vector and particle distribution characteristic vector within time t, and construct a particle population distribution equation; According to the particle population distribution equation, the particle population distribution change rate caused by the particle size change and the particle population distribution change rate based on the spatial coordinate system are calculated; According to the particle group distribution change rate caused by the particle size change and the particle group distribution change rate of the spatial coordinate system, a granulation process monitoring model is constructed. The granulation process monitoring model includes: Obtain the rate of change of particle population distribution caused by particle size change and the particle group distribution change rate based on the spatial coordinate system ; The granulation process monitoring model is constructed by combining the collective balance equation, including: ； Where t represents the target granulation time, x represents the particle distribution characteristic vector, and z represents the particle spatial position vector. represents the particle distribution function under the target granulation time t, particle distribution characteristic vector x and particle space position vector z, represents the particle generation rate, Represents the extinction rate of particles.

2. The resistor granulation monitoring method based on industrial Internet of Things according to claim 1 is characterized in that: The use of a distribution graph algorithm to remove abnormal values ​​in the granulation process data specifically includes: Quartiles were used to sort the collected granulation process data; Performing quartile dispersion calculation on the sorted granulation process data; Determine whether there are outliers in the granulation process data based on the dispersion, and if so, remove them.

3. The resistor granulation monitoring method based on industrial Internet of Things according to claim 2 is characterized in that: Determining whether there are abnormal values ​​in the granulation process data according to the discreteness specifically includes: Determine the difference between the lower quartile and the upper quartile based on the dispersion; According to the difference between the lower quartile and the upper quartile, it is judged whether the difference between the lower quartile and the upper quartile is within a set threshold. If not, it is determined that there are abnormal values ​​in the granulation process data, and the abnormal values ​​are eliminated.

4. The resistor granulation monitoring method based on industrial Internet of Things according to claim 1 is characterized in that: The fusing the first data based on batch estimation specifically includes: Assume that after removing the outliers from the n data measured by n sensors, there are m data left. The m data are divided into equal groups to obtain k groups, where each group contains j data. The arithmetic mean and standard deviation of k groups of data are calculated respectively, and the data fusion result based on batch estimation is obtained based on the arithmetic mean and standard deviation of k groups of data.

5. The resistor granulation monitoring method based on industrial Internet of Things according to claim 1 is characterized in that: The specific steps of extracting the particle spatial position vector and the particle distribution feature vector within time t include: Based on the second data, extracting the difference between the set granulation amount and the actual granulation amount during the granulation process; The difference change is obtained according to the difference between the set granulation amount and the actual granulation amount during the granulation process; According to the difference and difference change between the set granulation amount and the actual granulation amount during the granulation process, the spatial position of the particles and the particle distribution characteristics within the time t are determined; Based on the spatial position of the particle within time t, determine the spatial position vector of the particle within time t; Based on the particle distribution characteristics within time t, a particle distribution feature vector within time t is determined.

6. The resistor granulation monitoring method based on industrial Internet of Things according to claim 1 is characterized in that: The granulation process monitoring model is optimized according to rolling optimization and feedback correction, and the specific steps include: Conduct fault-tolerant control on the granulation process monitoring model and detect data loss in the granulation simulation process in real time; Determine whether the data loss of the granulation simulation process exceeds the data loss threshold of the granulation simulation process. If it exceeds, then: Conduct fault diagnosis on the granulation simulation process of the granulation process monitoring model, perform rolling optimization and feedback correction on the faults, and obtain continuous error feedback; The granulation process monitoring model is corrected according to continuous error feedback to obtain the optimized granulation process monitoring model.

7. A resistor granulation monitoring system based on industrial Internet of Things, applied to the resistor granulation monitoring method based on industrial Internet of Things according to any one of claims 1 to 6, characterized in that: Including granulation monitoring perception control platform, granulation monitoring sensor network platform and granulation monitoring management platform; The granulation monitoring perception control platform is used to collect granulation process data and transmit it to the granulation monitoring management platform through the granulation monitoring sensor network platform; The granulation monitoring and management platform includes a data processing module, a model building module and an analysis module; The data processing module is used to process the granulation process data; The model building module is used to build a granulation process monitoring model according to the collective balance equation; The analysis module is used to optimize the granulation process monitoring model according to rolling optimization and feedback correction to obtain the optimized granulation process monitoring model, and monitor the granulation process according to the optimized granulation process monitoring model.

8. The resistor granulation monitoring system based on industrial Internet of Things according to claim 7 is characterized in that: The specific steps of the data processing module processing the granulation process data include: Using a distribution graph algorithm to remove abnormal values ​​in the granulation process data to obtain first data; The first data is fused based on the batch estimation to obtain the second data.

Images