Storage medium, and leakage early warning method, device and equipment of microwave heating equipment

Abstract

The invention discloses a memory, and a leakage early warning method, device and equipment for microwave heating equipment, and the method comprises the steps: obtaining equipment historical data of the microwave heating equipment, and generating a modeling database for machine learning according to the time corresponding relation of the equipment historical data; determining the number of early warning levels of microwave leakage and the corresponding relation between the early warning levels and the radiation illumination range according to the corresponding relation between the radiation illumination of microwave leakage and the radiation pollution hazard degree; establishing a correlation coefficient matrix of the characteristic variables, and respectively establishing a data set corresponding to each characteristic variable group according to a calculation result of the correlation coefficient matrix; and performing data training and testing by taking the data set of the feature group as modeling data, and generating an early warning level prediction model for generating a microwave leakage early warning level according to the current data of the preset parameter item based on a random forest classification model. According to the invention, radiation pollution caused by microwave leakage of large-scale industrial microwave heating equipment can be effectively prevented.

Description

Technical Field

[0001] The invention relates to the field of microwave heating, and in particular to a leakage early warning method, device and equipment for storage devices and microwave heating equipment. Background Art

[0002] Microwave heating technology refers to a heating technology that uses the electromagnetic field generated by electromagnetic waves with a frequency between 300MHz and 300GHz to cause the polar molecules in the heated material to change their motion trajectory as the polarity of the electromagnetic field changes. In the process of changing the motion trajectory, many polar molecules rub and collide with each other, thereby converting electromagnetic energy into thermal energy.

[0003] Microwave heating is a common heating method for rapid pyrolysis of biomass. Compared with other pyrolysis methods, it has the characteristics of fast heating rate, short residence time, and moderate pyrolysis temperature. Therefore, it has research and development prospects in the field of chemical applications.

[0004] The microwave radiation generated during the microwave heating process is a physical pollution source, and its leakage will cause long-term and irreversible damage to the health of microwave operators. In order to prevent the radiation pollution caused by microwave leakage from microwave heating equipment, the existing technology usually uses the electromagnetic shielding principle to separate the area where the electromagnetic field is generated from other areas with a metal mesh.

[0005] The inventors have found through research that the microwave leakage prevention measures of microwave heating equipment in the prior art have at least the following defects:

[0006] The limitation of this method is that it is only applicable to small-scale laboratory-scale microwave equipment that does not operate for a long period or continuously, and it is difficult to apply it to large-scale industrial equipment that operates continuously for a long period. Therefore, large-scale industrial microwave heating equipment is at risk of radiation contamination due to microwave leakage.

[0007] The information disclosed in this background technology section is only intended to enhance the understanding of the overall background of the invention and should not be regarded as an acknowledgment or any form of suggestion that the information constitutes the prior art already known to a person skilled in the art. Summary of the invention

[0008] The purpose of the present invention is to prevent radiation pollution caused by microwave leakage of large-scale industrial microwave heating equipment.

[0009] The present invention provides a leakage warning method for microwave heating equipment, comprising the steps of:

[0010] S11, acquiring equipment historical data of the microwave heating equipment, and generating a modeling database for machine learning with parameter items as feature variables according to the time correspondence of each of the equipment historical data; the parameter items of the equipment historical data include equipment and working condition parameter items, and the radiation illumination of microwave leakage;

[0011] S12. Determine the number of warning levels for microwave leakage and the corresponding relationship between the warning level and the range of radiation illumination according to the corresponding relationship between the radiation illumination of microwave leakage and the degree of radiation pollution hazard;

[0012] S13, establishing a correlation coefficient matrix of features, and establishing data sets corresponding to each feature group according to calculation results of the correlation coefficient matrix;

[0013] S14. Performing data training and testing using the data set of the feature group as modeling data, and generating a warning level prediction model for generating a microwave leakage warning level according to current data of preset parameter items based on a random forest classification model.

[0014] Preferably, the present invention further comprises:

[0015] S15, taking the parameter items corresponding to the preset features in the feature group corresponding to the warning level prediction model as monitoring items, and obtaining the current monitoring data of each monitoring item;

[0016] S16. Generate input data of the warning level prediction model according to the current monitoring data, and generate a prediction result of the warning level after a preset time step through the warning level prediction model.

[0017] In another aspect of the present invention, a leakage warning device for microwave heating equipment is also provided, comprising:

[0018] A data processing unit, used to obtain equipment historical data of the microwave heating equipment, and generate a modeling database for machine learning with parameter items as characteristic variables according to the time correspondence between each of the equipment historical data; the parameter items of the equipment historical data include equipment and working condition parameter items, and the radiation illumination of microwave leakage;

[0019] The warning level setting unit is used to determine the number of warning levels of microwave leakage and the corresponding relationship between the warning level and the range of radiation illumination according to the corresponding relationship between the radiation illumination of microwave leakage and the degree of radiation pollution hazard;

[0020] A data set division unit is used to establish a correlation coefficient matrix of features, and to establish data sets corresponding to each feature group according to calculation results of the correlation coefficient matrix;

[0021] A modeling unit is used to perform data training and testing using the data set of the feature group as modeling data, and to generate a warning level prediction model for generating a microwave leakage warning level according to current data of preset parameter items based on a random forest classification model.

[0022] Preferably, the present invention further comprises:

[0023] A real-time data acquisition unit, used to obtain current monitoring data of each monitoring item by taking the parameter item corresponding to the preset feature in the feature group corresponding to the warning level prediction model as the monitoring item;

[0024] The warning result generating unit is used to generate input data of the warning level prediction model according to the current monitoring data, and generate a prediction result of the warning level after a preset time step through the warning level prediction model.

[0025] On the other hand, an embodiment of the present invention further provides a leakage warning device for microwave heating equipment, which includes a computer program stored on a medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer executes the methods described in the above aspects and achieves the same technical effects.

[0026] In another aspect of the embodiment of the present invention, a storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, each step of the leakage warning method of the microwave heating device as described in any one of the above items is implemented.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] The present invention is applicable to large-scale industrial microwave heating equipment that cannot be installed with a metal mesh for electromagnetic shielding. The inventor has found through research that, for large-scale industrial microwave heating equipment, the severity of microwave leakage is correlated with operating conditions and environmental factors. Therefore, in the present invention, a variety of operating conditions and environmental factors are used as feature items, and historical data of operating conditions and environmental factors are used as modeling data to construct a prediction model for predicting the warning level of microwave leakage (i.e., a warning level prediction model).

[0029] Since the present invention can predict whether there is a risk of microwave leakage and the location of microwave leakage based on the operating conditions and usage of large-scale industrial microwave heating equipment, it can effectively prevent radiation pollution caused by microwave leakage of large-scale industrial microwave heating equipment.

[0030] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention and implement it according to the contents of the specification, and to make the above and other purposes, technical features and advantages of the present invention easier to understand, one or more preferred embodiments are listed below and described in detail with the accompanying drawings as follows. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 is a step diagram of the leakage warning method of the microwave heating equipment of the present invention;

[0032] Figure 2 It is a structural schematic diagram of the leakage warning device of the microwave heating equipment of the present invention;

[0033] Figure 3 It is a structural schematic diagram of the leakage early warning control device of the microwave heating equipment in the present invention. DETAILED DESCRIPTION

[0034] The specific implementation modes of the present invention are described in detail below in conjunction with the accompanying drawings, but it should be understood that the protection scope of the present invention is not limited by the specific implementation modes.

[0035] Unless explicitly stated otherwise, throughout the specification and claims, the term “comprise” or variations such as “include” or “comprising”, etc., will be understood to include the stated elements or components but not to exclude other elements or components.

[0036] In this document, for the convenience of description, spatial relative terms such as "below", "below", "down", "above", "above", "upper", etc. may be used to describe the relationship between one element or feature and another element or feature in the accompanying drawings. It should be understood that the spatial relative terms are intended to include different orientations of the object in use or operation in addition to the orientation depicted in the figure. For example, if the object in the figure is turned over, the element described as being "below" or "below" other elements or features will be oriented "above" the element or feature. Therefore, the exemplary term "below" can include both below and above. Objects may also have other orientations (rotated 90 degrees or other orientations) and the spatial relative terms used in this document should be interpreted accordingly.

[0037] In this document, the terms "first", "second", etc. are used to distinguish two different elements or parts, and are not used to limit a specific position or relative relationship. In other words, in some embodiments, the terms "first", "second", etc. can also be interchangeable.

[0038] Embodiment 1

[0039] In order to prevent radiation pollution caused by microwave leakage from large industrial microwave heating equipment, such as Figure 1 As shown, in an embodiment of the present invention, a leakage warning method for microwave heating equipment is provided, comprising the steps of:

[0040] S11, acquiring equipment historical data of the microwave heating equipment, and generating a modeling database for machine learning with parameter items as feature variables according to the time correspondence of each of the equipment historical data; the parameter items of the equipment historical data include equipment and working condition parameter items, and the radiation illumination of microwave leakage;

[0041] In an embodiment of the present invention, the parameter items in the equipment and operating condition parameter items may specifically include: reactor structure type, operation mode, feed processing scale, microwave source information, microwave transmission classification, sealing method and number, mechanical parts connection method and number, microwave cavity material, material dielectric, operating conditions, environmental information, continuous stable operation cycle, cooling method, number of equipment exposures, number of equipment valves, equipment service life, the difference between the equipment service life and the design life, number of historical failures, inlet and outlet radiation illumination information, cavity electromagnetic field distribution information, whether there is an absorbing agent inside the equipment, whether the equipment is close to an electromagnetic interference source, whether the equipment storage location is ventilated, whether a cutoff waveguide is placed at the equipment exposure or an electromagnetic shielding measure is set, whether the microwave process involves a chemical reaction, and whether the equipment is grounded.

[0042] According to the time correspondence, the modeling database can establish the correspondence between the data of various equipment and working condition parameter items and the radiation illumination of microwave leakage.

[0043] In the embodiments of the present invention, the reactor structure type refers to the reactor structure type used in the microwave process device, specifically including: fixed bed reactor, bubbling bed reactor, turbulent bed reactor, moving bed reactor, tubular reactor, kettle reactor and tower reactor, etc.;

[0044] In the embodiments of the present invention, the operation mode refers to the operation mode of the microwave process device, including intermittent operation, semi-continuous operation and continuous operation, etc.

[0045] In the embodiments of the present invention, the feed processing scale refers to the real-time raw material feed amount that the microwave process device can process under normal operating conditions, generally expressed in kg / h, t / a;

[0046] In the embodiment of the present invention, the microwave source information refers to the relevant information of the microwave source in the microwave process device, including: the rated power of the magnetron (unit: W), the rated input voltage (unit: V), the rated input current (unit: A), the maximum input withstand voltage (unit: V), the preheating time (unit: s), the cooling air volume (unit: L / min), the cooling air direction (for example: cross type), the waveguide type (for example: BJ22, BJ26), the microwave rated frequency (unit: MHz), and the working altitude (unit: m);

[0047] In the embodiment of the present invention, the microwave transmission classification refers to the transmission mode in which the magnetron in the microwave process device generates microwaves and transmits them to the microwave cavity, including: coaxial line, suspended microstrip line, microstrip line, rectangular waveguide, strip line, circular waveguide, planar waveguide, fin line waveguide, slot line, dielectric waveguide optical fiber;

[0048] In the embodiments of the present invention, the sealing method and number, the sealing method refers to the physical barrier measures taken between adjacent joint surfaces of the mechanical structure to prevent internal materials from leaking to the outside and / or prevent external impurities from invading the interior of the equipment, including: packing seal, O-type seal, Y-type seal, mechanical seal, oil seal, split-petal sliding ring seal, labyrinth seal, floating ring seal, centrifugal seal, spiral seal, magnetic fluid seal; the number refers to the number of seals using the above sealing methods;

[0049] In the embodiments of the present invention, the connection mode and number of mechanical parts, the connection mode of mechanical parts refers to the process of connecting two separate profiles or parts into a complex part or component through the mechanical bite action of the connecting parts or the pressure bearing action of the end faces of the connecting parts, including: rivet connection, screw connection, hole riveting, TOX riveting, hook connection, hinge connection, welding, self-riveting, adhesive connection, expansion connection, bite seam connection, flange connection, clamp connection, threaded connection; the number refers to the number of connections made by the above mechanical connection modes;

[0050] In the embodiment of the present invention, the microwave cavity material refers to the material used for the microwave cavity in the microwave process device, including: a metal material of one or more alloys of copper, iron and aluminum;

[0051] In the embodiments of the present invention, the dielectric property of the material refers to the microwave dielectric property of the material of the main body of microwave heating performed by the microwave process device, including the dielectric constant, dielectric loss and loss tangent value of the material at 20°C, 100°C, 200°C, 300°C, 400°C, 500°C, 600°C, 700°C, 800°C, 900°C and 1000°C respectively;

[0052] In the embodiment of the present invention, the operating conditions refer to the real-time operating conditions of the microwave process device during continuous and stable operation, including: real-time reaction temperature, real-time pressure, real-time microwave power, real-time microwave frequency, real-time material input, real-time product output, and three waste discharge;

[0053] In the embodiment of the present invention, environmental information refers to relevant information about the environment in which the microwave process device is located during continuous and stable operation, including: ambient temperature, ambient humidity, dew point, light intensity, photosynthetically active radiation, wind direction, wind speed, rainfall, CO2 concentration, surface temperature, surface water content, and atmospheric pressure;

[0054] In the embodiment of the present invention, the continuous stable operation period refers to the time from the start-up of the microwave process device to the current stage of continuous operation, which is generally measured in weeks, days, hours, and can also be measured in minutes or seconds;

[0055] In the embodiment of the present invention, the cooling method refers to the cooling measures taken by the microwave-generating magnetron to stabilize at a suitable working temperature, including: natural cooling, heat sink cooling, fan cooling, water cooling, and liquid nitrogen cooling;

[0056] In the embodiment of the present invention, the number of equipment openings refers to the number of internal spaces of the microwave process device and external spaces that are directly connected without valves;

[0057] In the embodiment of the present invention, the number of equipment valves refers to the number of valves provided between the internal space of the microwave process device and the external space, which are directly connected to each other;

[0058] In the embodiment of the present invention, the service life of the equipment refers to the time span from the first start of the equipment to the present;

[0059] In the embodiment of the present invention, the difference between the service life of the equipment and the design life refers to the difference between the time span from the first start-up of the equipment to the present and the design life of the equipment;

[0060] In the embodiment of the present invention, the number of historical faults refers to the number of faults that have occurred since the equipment was first started, including the number of faults that do not involve microwave leakage;

[0061] In the embodiment of the present invention, the radiation illumination information of the inlet and outlet refers to the radiation illumination obtained by the microwave process device through the microwave detector and other equipment, that is, the radiation flux per unit area of ​​the irradiated surface, in mW / cm 2 Design; including: feed inlet and discharge outlet of microwave process equipment;

[0062] In the embodiment of the present invention, the cavity electromagnetic field distribution information refers to the electromagnetic field strength data in each cavity of the microwave process device when no material is added, which is simulated by using three-dimensional electromagnetic field simulation technology;

[0063] In the embodiment of the present invention, whether there is an absorbing agent inside the equipment refers to whether there are other gaseous, liquid and solid components that are obviously different from the material and can absorb microwave radiation in each cavity in the microwave process device, including but not limited to each cavity material itself that can absorb microwave radiation, coatings that can absorb microwave radiation on the inner wall surface of each cavity material, substances that can absorb microwave radiation that are pre-placed in each cavity in a fixed form before start-up, and substances that can absorb microwave radiation that enter each cavity in a mixed manner with the material after start-up;

[0064] In the embodiment of the present invention, whether the device is near an electromagnetic interference source refers to whether there is an interference source within the protection distance of the microwave process device, and the interference source includes but is not limited to power transmission lines, substations, electrified railways, welding equipment, power equipment, high-frequency furnaces, overfrequency physiotherapy equipment, etc.;

[0065] In the embodiment of the present invention, whether the storage location of the equipment is ventilated refers to whether the microwave process device is stored in a non-enclosed environment, or whether it is stored in an environment with natural or mechanical methods to ensure that the wind flow is unobstructed and can freely penetrate the outside world.

[0066] In the embodiments of the present invention, the open part of the equipment refers to a place where there is no physical barrier between the internal space of the microwave process device and the external environment space and can be freely connected (such as a feed inlet, a discharge outlet, etc.);

[0067] In the embodiment of the present invention, whether the microwave process involves a chemical reaction refers to whether a chemical reaction is involved in the microwave process device;

[0068] In the embodiment of the present invention, whether the equipment is grounded refers to whether the exposed conductive part of the microwave process device is connected to the ground via a conductor.

[0069] In practical applications, in order to filter out invalid data and reduce the amount of model calculation; or, in order to avoid overfitting of the prediction model and improve the prediction accuracy of the prediction model, in an embodiment of the present invention, the device historical data may also be preprocessed.

[0070] Specific preprocessing methods may include: one or more of one-hot encoding, statistic filling, average special value filling, class mean replacement, class random interpolation, regression interpolation, maximum likelihood estimation interpolation and multiple interpolation.

[0071] To facilitate data reading by the random forest classification model, an n-bit state register can be used to encode n states, and each state has an independent register bit to ensure that only one bit is valid at any time. Therefore, in an embodiment of the present invention, a one-hot encoding method can be used to compile data, and the specific steps may include:

[0072] S21, read all characteristic variables in the machine learning database of the microwave leakage prevention warning mechanism, use the OneHotEncoder function in the Sklearn algorithm library to filter out the part containing characteristic variables in the original data, and convert it into an array form;

[0073] S22, read the Onehot class (Categories), and re-transfer the object pointer in the Onehot class to the new list (list) according to the reading order of the feature variables;

[0074] S23, use the pandas library to convert the array in S21 into DataFrame format, and assign the table header name using the list in S23;

[0075] S24. Use the pandas library to merge the DataFrame in S23 with the part of the original data that does not contain the feature variables to complete the one-hot encoding.

[0076] In practical applications, when the feature variable has only a yes / no state (i.e., a binary feature variable), such as the feature variable "whether the equipment is grounded", it can be written in the following table (Table 1):

[0077] Table 1:

[0078] Serial number Device grounding Device not grounded 0 0 1 1 1 0

[0079] When the feature variable contains multiple states or is a discontinuous value (i.e., a multi-class feature variable), such as the feature variables "operation mode" and "equipment service life", they can be written in the following tables (Table 2 and Table 3) respectively:

[0080] Table 2:

[0081] Serial number Intermittent operation Semi-continuous operation Continuous operation 0 0 0 1 1 0 1 0 2 1 0 0

[0082] Table 3:

[0083] Serial number 1 year 2 years 3 years 4 years 5 years 0 0 0 0 0 1 1 0 0 1 0 0 2 0 1 0 0 0 3 1 0 0 0 0 4 0 0 0 1 0

[0084] When the feature variable is a continuous value, it is necessary to manually set the interval range, convert it into a multi-classification variable, and then perform one-hot encoding. For example, the feature variable "continuous stable operation period" can be written in the following table (Table 4):

[0085] Table 4:

[0086] Serial number 0<h≤200 200<h≤400 400<h≤600 600<h≤800 h>800 0 0 0 0 0 1 1 0 0 1 0 0 2 0 1 0 0 0 3 1 0 0 0 0 4 0 0 0 1 0

[0087] S12. Determine the number of warning levels for microwave leakage and the corresponding relationship between the warning level and the range of radiation illumination according to the corresponding relationship between the radiation illumination of microwave leakage and the degree of radiation pollution hazard;

[0088] The magnitude of the radiation illumination of microwave leakage is positively correlated with its degree of hazard; thus, the number of warning levels for microwave leakage and the correspondence between the radiation illumination of microwave leakage and the degree of radiation pollution hazard can be determined, as well as the correspondence between the warning level and the radiation illumination range. Specifically, several warning levels can be set in total, as well as the radiation illumination range corresponding to each warning level can be set; thus, the corresponding warning level can be determined based on the radiation illumination of microwave leakage.

[0089] In one example, the warning level may be set to include four warning levels, namely, A, B, C, and D; then, the corresponding relationship between each warning level and the radiation illumination range may be as follows:

[0090] When the radiation intensity UV is ≤1mw / cm at a distance of 5cm from the microwave process device 2 , it is determined that the current level of the microwave leakage warning mechanism is D;

[0091] When the irradiance at a distance of 5 cm from the microwave process device is 1 mw / cm 2 <UV≤5mw / cm 2 , and determined that the current level of the microwave leakage warning mechanism is C;

[0092] When the irradiance at a distance of 5 cm from the microwave process device is 5 mw / cm 2 <UV≤300mw / cm 2 , and determined that the current level of the microwave leakage warning mechanism is B;

[0093] When the irradiance at 5cm away from the microwave process device is >300mw / cm 2 , and determined that the current microwave leakage warning mechanism level is A.

[0094] S13, establishing a correlation coefficient matrix of features, and establishing data sets corresponding to each feature variable group according to the calculation results of the correlation coefficient matrix;

[0095] In practical applications, it is necessary to establish a correlation coefficient matrix of characteristic variables, so as to find characteristic variables whose characteristic correlation is not strongly correlated in statistics, and form a data set of several independent group characteristic features;

[0096] Specifically, in an embodiment of the present invention, the corr() function in the python language can be used to solve the correlation between different feature variables. The larger the correlation value, the greater the correlation between the feature variables. Generally, when the absolute value of the correlation value is greater than 0.8, the correlation between the two feature variables is considered to be strongly correlated. Therefore, according to the calculation results, the data set composed of the feature variable group of feature variables whose correlations are less than 0.8 is called an independent data set. If n such different data sets can be constructed, each data set can be recorded as Di (i = 1, 2, 3...n).

[0097] In practical applications, there will be strong correlations between some feature variables. If there are strongly correlated feature variables in a feature variable group, it is easy to cause overfitting of the prediction model. Therefore, if there are too many feature variables with strong correlations, the number of feature variables in the feature variable group will be too small, so the data set corresponding to the feature variable group will also be small, which will affect the prediction accuracy of the prediction model.

[0098] To this end, in an embodiment of the present invention, when preprocessing data, the correlation between two parameter items can also be reduced by feature transformation to avoid the problem of overfitting caused by excessive correlation. Specifically:

[0099] Assume that the correlation between parameter item A and parameter item B is greater than a preset threshold. At this time, feature transformation can be performed on parameter item A and parameter item B, specifically including:

[0100] S31, generate the corresponding covariance matrix C according to the data of parameter item A and parameter item B respectively A and the covariance matrix C B ;

[0101] C A =(1 / n)*(A-mean(A))*(A-mean(A)) T ;

[0102] C B =(1 / n)*(B-mean(B))*(B-mean(B)) T ;

[0103] Where n represents the size of the data set, mean(A) and mean(B) represent the means of data sets A and B respectively, and T represents the transposition operation;

[0104] S32, for C A and C B Perform eigenvalue decomposition and obtain C A and C B The eigenvalues ​​and eigenvectors of include:

[0105] C A =Q A *D A *Q A -1 ; C B =Q B *D B *Q B -1 ;

[0106] Among them, Q A and Q B is an orthogonal matrix, D A and D B It is a diagonal matrix, and the elements on the diagonal are the eigenvalues; Q A -1 and Q B -1 Q A and Q B The inverse matrix of .

[0107] S33, according to C B The eigenvalues ​​and C A The eigenvectors generate a new covariance matrix C new ,include:

[0108] C new =Q B *D B *Q B -1 *Q A *(D A +D I ) -1 *Q A -1 ;

[0109] Among them, D I It is a diagonal matrix after eigenvalue decomposition of the covariance matrix of the "optimal pyrolysis power" determined based on experience or equipment requirements; in the embodiment of the present invention, the optimal pyrolysis power refers to the power determined by the process based on experience or equipment requirements, for example, it may be 1000w.

[0110] S34. Obtain data set A through calculation of the new covariance matrix new and B new :in,

[0111] A new =A*C new T ; B new =B*C new T .

[0112] In the embodiment of the present invention, by performing eigenvalue decomposition, the eigenvalues ​​and eigenvectors of the covariance matrix can be obtained, so as to better understand the main features of the data and how they are related. Then, by constructing a new covariance matrix, the eigenvectors from C A , the eigenvalues ​​come from C B Construct a new covariance matrix C in the following way new , we can find that data set A new and B new The correlation between them is reduced, that is, the apparent linear relationship between the two parameter items is stripped away, and the non-numerical relationship between the two is magnified, thereby achieving the purpose of dimensionality reduction.

[0113] In practical applications, the parameter item A and the parameter item B in the embodiment of the present invention may be the two parameter items of microwave power and heating rate, respectively;

[0114] In a specific example, the correlation between the data of the two parameter items of pyrolysis power and heating rate is 0.9. After performing the above-mentioned feature transformation, the correlation between the two data items can be reduced to below 0.6.

[0115] It can be seen from the above that through the above feature transformation processing, data set A is generated based on the original data sets of parameter item A and parameter item B. new and B new ; Since the new data set A new and B new , through calculation, it can be found that their correlation has been effectively reduced. Therefore, in the subsequent model learning, these two data sets (A new and B new ) can be used as feature variables at the same time to better correct the prediction results of the model.

[0116] In addition, in an embodiment of the present invention, the number of feature variables can be increased by generating indirect parameter items. Since the newly generated indirect parameter items are newly added feature variables, the correlation between them and other feature variables may not be a strong correlation. Therefore, the number of feature variables in the feature variable group can be effectively increased, thereby increasing the data in the data set corresponding to the feature variable group. In this way, the prediction accuracy of the prediction model can be effectively improved.

[0117] Specifically in the application scenario of the embodiment of the present invention, the parameter items suitable for generating indirect parameter items mainly include some working condition parameter items and environmental information parameter items. In this way, indirect parameter items corresponding to preset working condition parameter items and / or environmental information parameter items can be generated. Specifically:

[0118] The preset working condition parameter items include: one or more of real-time reaction temperature, real-time pressure, microwave real-time power, microwave real-time frequency, real-time material input amount, real-time product output amount and three waste discharge amount; the environmental information parameter items include: one or more of ambient temperature, ambient humidity, dew point, light intensity, photosynthetically active radiation, wind direction, wind speed, rainfall, CO2 concentration, surface temperature, surface water content and atmospheric pressure;

[0119] The data value of the indirect parameter item is generated according to the data value of the corresponding preset working condition parameter or the environmental information parameter item, including:

[0120] Calculate the difference between the data value of the preset operating condition parameter or the environmental information parameter item and the corresponding reference data, and use the difference as the data value of the corresponding indirect parameter item; the reference data is the median or mean of the historical data of the preset operating condition parameter or the environmental information parameter item during the continuous and stable operation of the microwave process device.

[0121] Taking ambient temperature as an example, a reference data of ambient temperature can be set. For example, based on experience or equipment requirements, the optimal ambient temperature can be set to 25°C; in this way, a brief parameter item can be derived based on the ambient temperature data recorded in the historical data: "ambient temperature deviation"; for example, if the original data value of a certain ambient temperature is 28°C, then a data value of "+3°C" can be generated in the indirect parameter item "ambient temperature deviation" accordingly; if the original data value of a certain ambient temperature is 24°C, then a data value of "-1°C" can be generated in the indirect parameter item "ambient temperature deviation" accordingly, and so on. All data values ​​in the indirect parameter item "ambient temperature deviation" can be generated from the original data of the ambient temperature in the historical data; other operating condition parameters or the environmental information parameter items can also be used to generate corresponding indirect parameter items.

[0122] It should be noted that in the embodiments of the present invention, the reference data corresponding to each operating condition parameter and environmental information parameter item can be set by those skilled in the art based on experience or process requirements of the equipment, and is not specifically limited here.

[0123] S14. Performing data training and testing using the data set of the feature group as modeling data, and generating a warning level prediction model for generating a microwave leakage warning level according to current data of preset parameter items based on a random forest classification model.

[0124] In an embodiment of the present invention, the specific steps of constructing a warning level prediction model based on a random forest classification algorithm may include:

[0125] S41, obtaining a set of feature-independent data sets, and dividing them into a training set and a test set according to a preset ratio;

[0126] In the embodiment of the present invention, a data set Di is randomly read and set as the influence factor X subset. At the same time, the radiant illumination of the microwave leakage corresponding to the data set Di is read and set as the target value Y subset. Among them, the influence factor X subset may include all or part of the parameters of the machine learning database of the microwave leakage warning mechanism (this mainly depends on the number of correlated and non-strongly correlated parameters contained in the data set Di), and the number of parameters is recorded as Num; the target value Y subset is the radiant illumination of the microwave leakage, and the number of parameters is recorded as 1; the number of parameters of the data set Z jointly established by the influence factor X subset and the target value Y subset is Num+1;

[0127] In an embodiment of the present invention, the influencing factor X subset and the target value Y subset can be divided into a training set and a test set according to a preset ratio. The training set is used to train the random forest classification model, and the test set is used to test the prediction accuracy of the random forest classification model;

[0128] In the embodiment of the present invention, the preset ratio refers to the percentage of the number of data groups selected by the test set to the number of data groups selected by the test set and the training set; the value range of the preset ratio is 20%-40%, and more preferably 20%;

[0129] Specifically, assuming that the acquired data set Z has N groups, and each group of data contains Num+1 parameters, the data dimension of the influencing factor X subset is Dx=[N,Num], and the data dimension of the target value Y subset is Dy=[N,1]. When the preset ratio is 20%, the data dimension of the training set X_train is D X_train =[0.8N,Num], the data dimension of the test set X_test is D X_test =[0.2N,Num], the data dimension of the training set Y_train is D Y_train =[0.8N,1], the data dimension of the test set Y_test is D Y_test =[0.2N,1].

[0130] S42, randomly acquiring several groups of data in the training set, and establishing several decision trees according to a preset training method to establish a pre-random forest classification model Model_pre; the established pre-random forest classification model Model_pre constructs a coupling mapping relationship between independent features related to the microwave leakage warning mechanism and the preset level of the microwave leakage prevention warning mechanism;

[0131] The establishment of the pre-random forest classification model Model_pre provides a basis for the subsequent random forest classification model Model parameter adjustment. Therefore, its essence is to build several decision trees that are as different from each other as possible through training, and finally determine the prediction value of the model through the classification results of the decision trees according to the principle of "the minority obeys the majority". Therefore, it is an integrated learning method, and its scope belongs to supervised learning in machine learning.

[0132] The pre-random forest classification model Model_pre contains several decision trees, each of which is trained based on randomly acquired data using a preset training method to present different characteristics. The construction of each decision tree requires random data capture with replacement from the training set. Generally speaking, to ensure that the differences between decision trees are large, the number of data groups in the training set and the number of data groups required to capture each decision tree must be maintained at a certain number.

[0133] In the embodiment of the present invention, in order to ensure the versatility of the model, the number of data groups in the training set should not be less than 500 groups. At the same time, in order to ensure the differences between decision trees, the number of data groups that each decision tree needs to capture accounts for 40% to 60% of the number of data groups in the training set.

[0134] In the embodiment of the present invention, the pre-random forest classification model Model_pre contains N decision trees;

[0135] In an embodiment of the present invention, the preset training method may include the following sub-steps:

[0136] S411, according to the set number of decision trees N, each decision tree randomly grabs data with replacement from the training sets X_train and Y_train according to a preset ratio as the training data set T of the decision tree i ;

[0137] S412, for the i-th decision tree, the training data set T i The parameters in are sorted by the importance of their features using the Xgboost algorithm, and the first several items are stored in the subset Param_i-1 according to the preset values, and the remaining items are stored in the subset Param_i-2;

[0138] In an embodiment of the present invention, the feature variables are integratedly learned through the XGBRFClassifer model in the Xgboost algorithm. The advantage of this is that the feature variables that have a significant impact on the radiation illumination information of microwave leakage can be preferentially extracted to the subset Param_i-1 for storage, and binary partitioning is performed in sequence, thereby reducing the complexity of the decision tree and the response time of the system for calculating the Gini index, and improving the prediction accuracy.

[0139] In the embodiment of the present invention, the preset value is generally set to 5, that is, the first 5 items are stored in the subset Param_i-1, and the remaining items are stored in the subset Param_i-2.

[0140] S413, for the i-th decision tree, when performing binary partition on its j-th layer, determine whether the number of samples in the node is less than the threshold or whether the Gini index of the sample set is less than the threshold; if yes, proceed to step S414; if no, determine whether there are feature variables in the subset Param_i-1 that have not generated nodes; if no, proceed to step S414; if yes, generate nodes for the feature variables in the subset Param_i-1 in sequence (the feature variables are the optimal features of the j-th layer), and perform binary partition; then, repeat step S413;

[0141] In the embodiment of the present invention, if the feature variable is a binary variable, a node is directly generated and a binary partition is performed, and the feature variable itself is the optimal split point (yes / no);

[0142] In the embodiment of the present invention, if the feature variable is a multi-classification variable, it is necessary to calculate the Gini index of each classification case respectively, and select the classification with the smallest Gini index as the optimal split point under the feature. Afterwards, when generating nodes and performing binary partitioning, the optimal split point is used as the basis for partitioning;

[0143] In the embodiment of the present invention, the multi-classification variable feature variable x k , when the selected split point is S n When , the data set corresponding to the current feature variable is divided into two subsets D1 and D2, where:

[0144] D1={x k ∈S n}

[0145]

[0146] In the embodiment of the present invention, the probability that a sample point in the data set belongs to the subset D1 is p, and the probability that it belongs to the subset D2 is q. The calculation formula of the Gini index of the two subsets D1 and D2 is:

[0147] Gini(D1)=2p(1-p)

[0148] Gini(D2)=2q(1-q)

[0149] In the embodiment of the present invention, the multi-classification variable feature variable x k , when the selected split point is S n The calculation formula of the Gini index is:

[0150]

[0151] In the embodiment of the present invention, the threshold value of the number of samples in the node is 2% of the total number of samples;

[0152] In the embodiment of the present invention, the threshold value of the Gini index of the sample set in the node is 0.02;

[0153] S414, for the i-th decision tree, when performing binary partition on its j-th layer, determine whether the number of samples in the node is less than the threshold or whether the Gini index of the sample set is less than the threshold; if yes, proceed to step S415; if no, determine whether there are feature variables in the subset Param_i-2 that have not generated nodes, if no, proceed to step S415; if yes, calculate the optimal split point of each feature variable, and compare the Gini index of each feature variable under the optimal split, select the feature variable with the smallest Gini index to generate a node, and use this feature as the optimal feature of the j-th layer, and its optimal split point as the basis for binary partition of the node; then, repeat step S414;

[0154] S415. When all N decision trees are established, the training of the pre-random forest classification model Model_pre is completed.

[0155] S43, obtaining independent feature data related to the microwave leakage warning mechanism in the test set, and using the trained pre-random forest classification model Model_pre to predict the level of the microwave leakage warning mechanism in the test set, and after the prediction, obtaining the level of the microwave leakage warning mechanism in the test set, and calculating the confusion matrix Mc and the corresponding F1 value;

[0156] In an embodiment of the present invention, predicting the test set is a key step in verifying the training effect of the pre-random forest classification model Model_pre, and is also an important reference for the subsequent parameter adjustment steps. Since the pre-random forest classification model Model_pre contains N decision trees with large differences, for each set of data in the test set, each decision tree established after training can output corresponding prediction results. The pre-random forest classification model Model_pre then determines the prediction results of the model based on the output results of all decision trees and the principle of "the minority obeys the majority".

[0157] In the embodiment of the present invention, the confusion matrix Mc is a visualization tool for judging the training effect of a certain category. By comparing the classification results and the actual prediction values ​​together, the prediction ability of the model for a certain category can be intuitively reflected. The specific expression form is as follows:

[0158] Prediction value = (A, B, C, D) class Prediction value = non-(A,B,C,D) class True value = (A, B, C, D) class TP FN True value = non-(A,B,C,D) class FP TN

[0159] Among them, TP refers to the number of true samples that are correctly classified, that is, they are predicted to be true and actually true; FP refers to the number of false samples that are incorrectly marked as true samples, that is, they are actually false but predicted to be true; TN refers to the number of false samples that are correctly classified, that is, they are predicted to be false and actually false; FN refers to the number of true samples that are incorrectly marked as false samples, that is, they are actually true but predicted to be false.

[0160] Specifically, in the early warning mechanism for predicting Class A microwave leakage, when the actual value is Class A and the predicted value is also Class A, it is classified as TP; when the actual value is Class A and the predicted value is Class B, C or D, it is classified as FN; when the actual value is Class B, C or D and the predicted value is Class A, it is classified as FP; when the actual value is Class B, C or D and the predicted value is Class B, C or D, it is classified as TN; it is worth noting that the classification of TN depends on the currently specified category: for example, when the actual value is Class B and the predicted value is Class C, it is also classified as TN, that is, although the model predicts the actual classification incorrectly, from the perspective of whether it is Class A, the model's judgment is not wrong.

[0161] In the embodiment of the present invention, the F1 value is an important indicator for measuring classification problems and a key parameter for judging the quality of training effects. It is the harmonic mean of precision (Pre) and recall (Recall). The closer it is to 1, the better the prediction effect is, and the closer it is to 0, the worse the prediction effect is. Precision (Pre) is the proportion of correctly classified true samples in the predicted true samples, and recall (Recall) is the proportion of correctly classified true samples in the actual true samples. Their specific calculation formulas are as follows:

[0162]

[0163]

[0164]

[0165] S44, adjusting the parameters of the pre-random forest classification model Model_pre according to the preset parameter adjustment method, and repeating steps S42-S43 until the confusion matrix Mc and the corresponding F1 value of all corresponding parameter adjustment results are obtained;

[0166] In an embodiment of the present invention, the preset parameter adjustment method may include the following sub-steps:

[0167] S441, changing the Gini index as the calculation standard of the optimal feature, using information gain and gain rate to select the optimal feature of the decision tree, and performing branching and partitioning operations on all the cut points under the optimal feature (that is, the decision tree at this time is no longer a binary partition);

[0168] In the embodiment of the present invention, information gain is used as the calculation standard of the optimal feature, and the feature variable x k The specific formula is as follows:

[0169]

[0170] Among them, when there are K feature variables, |C k | is the kth feature variable x k The number of samples; when the feature variable x k When there are n split points, |D i | is the number of samples contained in the nth splittable point; |D ik | is the k-th feature variable x in the sample contained in the n-th split point k The number of samples.

[0171] In the embodiment of the present invention, the feature with the largest information gain is selected as the optimal feature of the jth layer;

[0172] In the embodiment of the present invention, the gain rate is used as the calculation standard of the optimal feature, and the feature variable x k The specific formula is as follows:

[0173]

[0174] In the embodiment of the present invention, the feature with the largest gain rate is selected as the optimal feature of the jth layer;

[0175] S442, change the set number of decision trees, taking the existing number of decision trees N as the minimum value N min , with the preset ratio of the existing number of decision trees N as the maximum value N max , followed by [N min ,N max ] as the selection range of the number of decision trees in the parameter adjustment method;

[0176] In the embodiment of the present invention, the preset ratio range of the number of decision trees is 10% to 100%, and is generally preferably 50%.

[0177] In an embodiment of the present invention, the selection range of the number of decision trees means that after the selection range is set, all possibilities within the range need to be traversed. For example, when the number of decision trees N is 100 by default, 100 can be used as the minimum number of decision trees, and 50% can be used as the preset ratio, then 150 is the maximum number of decision trees under the current step size. In this way, [100, 150] is used as the selection range of the number of decision trees, that is, 100, 101, 102, 103, 104, ..., 150 are used as the selection range of the number of decision trees in the parameter adjustment method.

[0178] S443, looping through the parameters in S441 and S442, and calculating the confusion matrix Mc and the corresponding F1 value under all permutations and combinations;

[0179] In the embodiment of the present invention, there is no specific requirement for the order of S441 and S442, but it is determined by the actual process conditions and the actual computing power of the computer, and it is sufficient that all permutations and combinations that meet the calculation criteria of the optimal features and the number of decision trees are calculated. In other words, these combinations can be calculated sequentially by a single-thread processor, in parallel by a multi-thread processor, or by coupling calculations of multiple threads and multiple processors.

[0180] S45, select the parameter adjustment result corresponding to the pre-random forest classification model with an F1 value closest to 1, then perform K-fold cross validation on the random classification model Model, use the parameter adjustment result and train according to the preset training method in S32, and calculate the comprehensive mixing matrix Mc_f and the corresponding comprehensive F1 value after K-fold cross validation;

[0181] In an embodiment of the present invention, the data set Z is equally divided into K groups of data sets, and each time one group of data sets is selected as a test set, and the remaining K-1 groups of data sets are used as training sets and trained using the parameter adjustment results and the preset training method in S42, so that K models can be obtained. Afterwards, the accuracy of the prediction results of the K models is averaged to obtain the comprehensive mixing matrix Mc_f and the corresponding comprehensive F1 value.

[0182] In the embodiment of the present invention, the value of K-fold cross validation is generally not less than 3, preferably, 10;

[0183] S46, judging whether the comprehensive F1 value is greater than a preset value, if not, returning to step S41 and reselecting another set of feature-independent data sets, and revising the parameter adjustment method in S44 until the preset value is met;

[0184] In the embodiment of the present invention, the preset value of the comprehensive F1 value is generally not less than 0.9, preferably, 0.95;

[0185] In an embodiment of the present invention, the parameter adjustment method in S44 is revised, including increasing the threshold of the number of samples in the node, increasing the threshold of the Gini index of the sample set in the node, increasing the selection range of the number of decision trees, etc.

[0186] As can be seen from the above, the embodiment of the present invention establishes a method for preventing microwave leakage warning mechanism based on the random forest classification model, and uses the algorithm to eliminate the risk of microwave leakage in actual processes. Compared with human experience judgment, the random forest classification model is based on real data in the past, so it is more versatile and timely.

[0187] By establishing a machine learning database for microwave leakage warning mechanism based on random forest classification model and training the random forest classification model on this basis, we can make full use of multiple influencing factors to predict the microwave leakage that may exist in the working conditions, so as to carry out human or mechanical intervention earlier to avoid personal safety and property losses. This method can also effectively avoid investing a large amount of repetitive manpower for detection, and has made outstanding contributions in the field of artificial intelligence manufacturing.

[0188] In the preset training method in the embodiment of the present invention, by establishing a pre-random forest classification model, all possible adjustment parameters involved in the training are trained one by one within the range, and the confusion matrix Mc and the corresponding F1 value corresponding to all the parameter adjustment results are compared to obtain the optimal adjustment result of the parameters. The advantage of this is that the parameters of the random forest classification model to be trained can be screened before the device is started. In this way, when the random forest classification model is subjected to K-fold cross validation, the training pass rate can be improved, that is, the response time of the algorithm can be improved. At the same time, when the confusion matrix Mc and the corresponding F1 value of the trained model fail to pass the preset value, when tracing back to the parameter adjustment stage, the parameter range that has been screened can also be skipped, so that the parameter adjustment is more targeted.

[0189] It can be seen from the above that the embodiment of the present invention relies on a random forest classification model established with high-dimensional parameters and large amounts of data to predict the impact of working conditions on products, and relies on the prediction results to perform human intervention or mechanical intervention according to different warning levels, thereby preventing potential risks brought about by microwave leakage and preventing the serious consequences brought about by the gradual expansion of microwave leakage in advance, thereby strengthening production safety and improving the feasibility of the process.

[0190] Furthermore, in the embodiment of the present invention, the steps of using the warning level prediction model may also be included, which are specifically as follows:

[0191] S15, taking the parameter items corresponding to the preset features in the feature group corresponding to the warning level prediction model as monitoring items, and obtaining the current monitoring data of each monitoring item;

[0192] S16. Generate input data of the warning level prediction model according to the current monitoring data, and generate a prediction result of the warning level after a preset time step through the warning level prediction model.

[0193] In practical applications, the preset actual step size can be used as the calculation cycle, and the prediction result for the next time step can be obtained according to the random forest classification model Model.

[0194] In the embodiment of the present invention, the warning level prediction model may first generate a radiation illumination prediction value, and then generate a warning level prediction value according to the corresponding relationship between the warning level and the radiation illumination range. In this way, the microwave heating equipment is intervened according to the predicted microwave leakage prevention warning level using corresponding rules.

[0195] In summary, the embodiments of the present invention are applicable to large-scale industrial microwave heating equipment that cannot be installed with metal mesh for electromagnetic shielding. The inventors have found through research that, for large-scale industrial microwave heating equipment, the severity of microwave leakage is correlated with operating conditions and environmental factors. Therefore, in the embodiments of the present invention, a variety of operating conditions and environmental factors are used as feature items, and the historical data of the operating conditions and environmental factors are used as modeling data to construct a prediction model for predicting the warning level of microwave leakage (i.e., a warning level prediction model).

[0196] Since the embodiment of the present invention can predict whether there is a risk of microwave leakage and the location of microwave leakage based on the operating conditions and usage of the large-scale industrial microwave heating equipment, it is also possible to effectively prevent radiation pollution caused by microwave leakage of the large-scale industrial microwave heating equipment.

[0197] Embodiment 2

[0198] In another aspect of the embodiment of the present invention, a leakage warning device for microwave heating equipment is also provided. Figure 2 A schematic diagram showing the structure of a leakage warning device for a microwave heating device provided in an embodiment of the present invention is shown. The leakage warning device for a microwave heating device is Figure 1 The device corresponding to the leakage warning method of the microwave heating device described in the corresponding embodiment, that is, implemented by means of a virtual device Figure 1 In the corresponding embodiment, the leakage warning method of the microwave heating device, each virtual module constituting the leakage warning device of the microwave heating device can be executed by an electronic device, such as a network device, a terminal device, or a server. Specifically, the leakage warning device of the microwave heating device in the embodiment of the present invention includes:

[0199] The data processing unit 01 is used to obtain the equipment historical data of the microwave heating equipment, and generate a modeling database for machine learning with parameter items as feature variables according to the time correspondence between each of the equipment historical data; the parameter items of the equipment historical data include equipment and working condition parameter items, and the radiation illumination of microwave leakage;

[0200] The warning level setting unit 02 is used to determine the number of warning levels for microwave leakage and the corresponding relationship between the warning level and the range of radiation illumination according to the corresponding relationship between the radiation illumination of microwave leakage and the degree of radiation pollution hazard;

[0201] The data set division unit 03 is used to establish a correlation coefficient matrix of features, and to establish data sets corresponding to each feature group according to the calculation results of the correlation coefficient matrix;

[0202] The modeling unit 04 is used to perform data training and testing using the data set of the feature group as modeling data, and to generate a warning level prediction model for generating a microwave leakage warning level according to current data of preset parameter items based on a random forest classification model.

[0203] Preferably, the present invention may also include:

[0204] A real-time data acquisition unit, used to obtain current monitoring data of each monitoring item by taking the parameter item corresponding to the preset feature in the feature group corresponding to the warning level prediction model as the monitoring item;

[0205] The warning result generating unit is used to generate input data of the warning level prediction model according to the current monitoring data, and generate a prediction result of the warning level after a preset time step through the warning level prediction model.

[0206] Since the working principle and beneficial effects of the leakage warning device of the microwave heating equipment in the embodiment of the present invention have been Figure 1 The corresponding leakage warning method of the microwave heating equipment is also recorded and explained, so they can be referenced by each other and will not be repeated here.

[0207] Embodiment 3

[0208] Corresponding to the method embodiment, in the embodiment of the present invention, a leakage warning device for microwave heating equipment is also provided, such as a terminal, a server, etc. Among them, the server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc., but is not limited thereto.

[0209] An example diagram of a hardware structure block diagram of a leakage warning device for a microwave heating device provided in an embodiment of the present invention is shown in FIG. Figure 3 As shown, it may include:

[0210] Processor 1, communication interface 2, memory 3 and communication bus 4;

[0211] The processor 1, the communication interface 2, and the memory 3 communicate with each other via the communication bus 4;

[0212] Optionally, the communication interface 2 may be an interface of a communication module, such as an interface of a GSM module;

[0213] The processor 1 may be a central processing unit CPU, or an application-specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present application.

[0214] The memory 3 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0215] The processor 1 is specifically used to execute the computer program stored in the memory 3 to perform the following steps:

[0216] S11, acquiring equipment historical data of the microwave heating equipment, and generating a modeling database for machine learning with parameter items as feature variables according to the time correspondence of each of the equipment historical data; the parameter items of the equipment historical data include equipment and working condition parameter items, and the radiation illumination of microwave leakage;

[0217] S12. Determine the number of warning levels for microwave leakage and the corresponding relationship between the warning level and the range of radiation illumination according to the corresponding relationship between the radiation illumination of microwave leakage and the degree of radiation pollution hazard;

[0218] S13, establishing a correlation coefficient matrix of features, and establishing data sets corresponding to each feature group according to calculation results of the correlation coefficient matrix;

[0219] S14. Performing data training and testing using the data set of the feature group as modeling data, and generating a warning level prediction model for generating a microwave leakage warning level according to current data of preset parameter items based on a random forest classification model.

[0220] Preferably, the present invention may also include:

[0221] S15, taking the parameter items corresponding to the preset features in the feature group corresponding to the warning level prediction model as monitoring items, and obtaining the current monitoring data of each monitoring item;

[0222] S16. Generate input data of the warning level prediction model according to the current monitoring data, and generate a prediction result of the warning level after a preset time step through the warning level prediction model.

[0223] The above product can execute the method provided by the embodiment of the present invention, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not described in detail in this embodiment, please refer to the leakage warning method of the microwave heating device provided by the embodiment of the present invention.

[0224] Embodiment 4

[0225] In an embodiment of the present invention, a storage medium is further provided, which may store a program suitable for execution by a processor, wherein the program is used to:

[0226] S11, acquiring equipment historical data of the microwave heating equipment, and generating a modeling database for machine learning with parameter items as feature variables according to the time correspondence of each of the equipment historical data; the parameter items of the equipment historical data include equipment and working condition parameter items, and the radiation illumination of microwave leakage;

[0227] S12. Determine the number of warning levels for microwave leakage and the corresponding relationship between the warning level and the range of radiation illumination according to the corresponding relationship between the radiation illumination of microwave leakage and the degree of radiation pollution hazard;

[0228] S13, establishing a correlation coefficient matrix of features, and establishing data sets corresponding to each feature group according to calculation results of the correlation coefficient matrix;

[0229] S14. Performing data training and testing using the data set of the feature group as modeling data, and generating a warning level prediction model for generating a microwave leakage warning level according to current data of preset parameter items based on a random forest classification model.

[0230] Preferably, the present invention may also include:

[0231] S15, taking the parameter items corresponding to the preset features in the feature group corresponding to the warning level prediction model as monitoring items, and obtaining the current monitoring data of each monitoring item;

[0232] S16. Generate input data of the warning level prediction model according to the current monitoring data, and generate a prediction result of the warning level after a preset time step through the warning level prediction model.

[0233] Optionally, the detailed functions and extended functions of the program may refer to the above description.

[0234] The above product can execute the method provided by the embodiment of the present invention, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not described in detail in this embodiment, please refer to the methods provided by other embodiments of the present invention.

[0235] Those of ordinary skill 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 this application.

[0236] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interface, device or unit, which can be electrical, mechanical or other forms.

[0237] 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.

[0238] In addition, each functional unit in each embodiment of the present application 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.

[0239] It should be understood that in the embodiments of the present application, the various embodiments and features can be combined with each other to solve the aforementioned technical problems.

[0240] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0241] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A leakage warning method for microwave heating equipment, characterized in that: Includes steps: S11, acquiring equipment historical data of the microwave heating equipment, and generating a modeling database for machine learning with parameter items as feature variables according to the time correspondence of each of the equipment historical data; the parameter items of the equipment historical data include equipment and working condition parameter items, and the radiation illumination of microwave leakage; S12. Determine the number of warning levels for microwave leakage and the corresponding relationship between the warning level and the range of radiation illumination according to the corresponding relationship between the radiation illumination of microwave leakage and the degree of radiation pollution hazard; S13, establishing a correlation coefficient matrix of characteristic variables, and establishing data sets corresponding to each characteristic variable group according to the calculation results of the correlation coefficient matrix; S14. Performing data training and testing using the data set of the feature group as modeling data, and generating a warning level prediction model for generating a microwave leakage warning level according to current data of preset parameter items based on a random forest classification model.

2. The leakage warning method of microwave heating equipment according to claim 1, characterized in that: Also includes: S15, taking the parameter items corresponding to the preset features in the feature group corresponding to the warning level prediction model as monitoring items, and obtaining the current monitoring data of each monitoring item; S16. Generate input data of the warning level prediction model according to the current monitoring data, and generate a prediction result of the warning level after a preset time step through the warning level prediction model.

3. The leakage warning method for microwave heating equipment according to claim 1 or 2, characterized in that: Before generating a modeling database for machine learning using the parameter items as feature variables, the method further includes: Preprocessing the device historical data; The preprocessing includes: one or more of one-hot encoding, statistic filling, average special value filling, class mean replacement, class random interpolation, regression interpolation, maximum likelihood estimation interpolation and multiple interpolation.

4. The leakage warning method of microwave heating equipment according to claim 3, characterized in that: The equipment and working condition parameters include: Reactor structure type, operation mode, feed processing scale, microwave source information, microwave transmission classification, sealing method and number, mechanical parts connection method and number, microwave cavity material, material dielectric, operating conditions, environmental information, continuous stable operation cycle, cooling method, number of equipment exposures, number of equipment valves, equipment service life, difference between equipment service life and design life, number of historical failures, radiation illumination information at the inlet and outlet, cavity electromagnetic field distribution information, whether there are absorbing agents inside the equipment, whether the equipment is close to electromagnetic interference sources, whether the equipment is stored in a ventilated location, whether cutoff waveguides are placed at the equipment exposures or electromagnetic shielding measures are set up, whether the microwave process involves chemical reactions, and whether the equipment is grounded.

5. The leakage warning method of microwave heating equipment according to claim 4, characterized in that: The parameter items of the equipment history data further include: indirect parameter items corresponding to the preset working condition parameter items and / or environmental information parameter items; The preset working condition parameter items include: one or more of real-time reaction temperature, real-time pressure, microwave real-time power, microwave real-time frequency, real-time material input amount, real-time product output amount and three wastes discharge amount; the environmental information parameter items include: one or more of ambient temperature, ambient humidity, dew point, light intensity, photosynthetically active radiation, wind direction, wind speed, rainfall, CO2 concentration, surface temperature, surface water content and atmospheric pressure; The data value of the indirect parameter item is generated according to the data value of the corresponding preset working condition parameter or the environmental information parameter item, including: Calculating the difference between the data value of the preset working condition parameter or the environmental information parameter item and the corresponding reference data, and using the difference as the data value of the corresponding indirect parameter item; the reference data is the median or mean of the historical data of the preset working condition parameter or the environmental information parameter item during the continuous and stable operation of the microwave process device.

6. The leakage warning method of microwave heating equipment according to claim 5, characterized in that: The encoding process of the equipment historical data based on the one-hot encoding method includes: S21. Read all feature variables in the machine learning database of the microwave leakage prevention warning mechanism, and use the OneHotEncoder function in the Sklearn algorithm library to screen out the part containing feature variables in the original data and convert it into an array form; S22. Read the Onehot class (Categories), and re-transmit the object pointer in the Onehot class to a new list according to the reading order of the feature variables; S23. Use the pandas library to convert the array in S21 into a DataFrame form, and assign the header name with the list in S23; S24. Use the pandas library to merge the DataFrame in S23 with the part of the original data that does not contain feature variables to complete the one-hot encoding.

7. The leakage warning method of microwave heating equipment according to claim 3, characterized in that: The preprocessing also includes; When the correlation between parameter item A and parameter item B is greater than the preset threshold, perform feature transformation on parameter item A and parameter item B, including: S31, generate the corresponding covariance matrix C according to the data of parameter item A and parameter item B respectively A and the covariance matrix C B ; S32, for C A and C B Perform eigenvalue decomposition and obtain C A and C B The eigenvalues ​​and eigenvectors of include: C A =Q A *D A *Q A -1 ;C B =Q B *D B *Q B -1 ; Among them, Q A and Q B is an orthogonal matrix, D A and D B It is a diagonal matrix, and the elements on the diagonal are the eigenvalues; S33, according to C B The eigenvalues ​​and C A The eigenvectors of generate the new covariance matrix C new ,include: C new =Q B *D B *Q B -1 *Q A *(D A +D I ) -1 *Q A -1 ; Among them, D I is the diagonal matrix after eigenvalue decomposition of the covariance matrix of the optimal pyrolysis power; S34. Obtain data set A through calculation of the new covariance matrix new and B new :in, A new =A*C new T ;B new =B*C new T 。 8. The leakage warning method of microwave heating equipment according to claim 5, characterized in that: Determining the number of warning levels of microwave leakage and the corresponding relationship between the warning level and the radiation illuminance range according to the corresponding relationship between the radiation illuminance of microwave leakage and the degree of radiation pollution hazard, including: When the radiation illuminance UV at a distance of 5 cm from the microwave-related process device is ≤ 1 mw / cm2, it is determined that the current microwave leakage prevention warning mechanism level is D level; When the radiation illuminance 1 mw / cm2 < UV ≤ 5 mw / cm2 at a distance of 5 cm from the microwave-related process device, it is determined that the current microwave leakage prevention warning mechanism level is C level; When the radiation illuminance 5 mw / cm2 < UV ≤ 300 mw / cm2 at a distance of 5 cm from the microwave-related process device, it is determined that the current microwave leakage prevention warning mechanism level is B level; When the radiation illuminance at a distance of 5 cm from the microwave-related process device > 300 mw / cm2, it is determined that the current microwave leakage prevention warning mechanism level is A level.

9. A leakage warning device for microwave heating equipment, characterized in that: Including: A data processing unit for obtaining the equipment historical data of the microwave heating equipment, and generating a modeling database for machine learning with parameter items as feature variables according to the time correspondence of each equipment historical data; the parameter items of the equipment historical data include equipment and working condition parameter items, and the radiation illuminance of microwave leakage; A warning level setting unit for determining the number of warning levels of microwave leakage and the corresponding relationship between the warning level and the radiation illuminance range according to the corresponding relationship between the radiation illuminance of microwave leakage and the degree of radiation pollution hazard; A data set division unit is used to establish a correlation coefficient matrix of features, and to establish data sets corresponding to each feature group according to calculation results of the correlation coefficient matrix; A modeling unit is used to perform data training and testing using the data set of the feature group as modeling data, and to generate a warning level prediction model for generating a microwave leakage warning level according to current data of preset parameter items based on a random forest classification model.

10. The leakage warning device for microwave heating equipment according to claim 9, characterized in that: Also includes: A real-time data acquisition unit, used to obtain current monitoring data of each monitoring item by taking the parameter item corresponding to the preset feature in the feature group corresponding to the warning level prediction model as the monitoring item; The warning result generating unit is used to generate input data of the warning level prediction model according to the current monitoring data, and generate a prediction result of the warning level after a preset time step through the warning level prediction model.

11. A leakage warning device for microwave heating equipment, characterized in that: include: Memory for storing computer programs; A processor is used to call and execute the computer program to implement the steps of the leakage warning method of the microwave heating equipment according to any one of claims 1 to 8.

12. A storage medium, characterized in that: The method comprises a software program, wherein the software program is suitable for executing, by a processor, the steps of the leakage warning method for the microwave heating device according to any one of claims 1 to 8.

Images