Industrial furnace energy data monitoring and processing system and method based on machine learning

Abstract

The present invention relates to the technical field of data processing, and discloses a system and method for monitoring and processing industrial kiln energy data based on machine learning. The method installs monitoring equipment on the side wall of the industrial kiln, collects industrial kiln data through the installed monitoring equipment, performs preprocessing and three-dimensional modeling, and predicts the three-dimensional modeling process of the industrial kiln through a support vector machine algorithm to obtain predicted three-dimensional model data; based on the obtained predicted three-dimensional model data, an industrial kiln energy-saving evaluation system is preliminarily established through an industrial kiln transformation energy-saving calculation method, and the obtained predicted three-dimensional model data is transformed to obtain transformed three-dimensional model data; finally, based on the established industrial kiln energy-saving evaluation system, the transformed three-dimensional model data is optimized through a genetic algorithm, and an industrial kiln is produced and monitored in real time according to the optimized three-dimensional model data, thereby improving the energy-saving effect of industrial kiln energy data monitoring.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to an industrial furnace energy data monitoring and processing system and method based on machine learning. Background Art

[0002] According to the current status of industrial enterprises, the energy consumption of kiln firing process accounts for about 60% to 70% of the energy consumption of kiln firing process chain, and the cost loss of industrial production increases. Therefore, it has become a challenge in the industrial production process to transform kilns and reduce the energy consumption of kiln firing process chain.

[0003] The existing Chinese patent application CN110287642A, the invention measures and saves the various parameters of the ceramic kiln in real time by installing intelligent detection equipment, and builds a thermal technology scheme and calculation model for adjusting the air coefficient based on the saved parameter data; at the same time, the air coefficient of the ceramic kiln is adjusted to the optimal value through the thermal technology scheme and calculation model for adjusting the air coefficient, and operates according to the optimal value. However, since the invention only considers the energy-saving adjustment of the ceramic kiln by adjusting the air coefficient, the energy-saving effect is not good, and the reference data is insufficient, it has certain limitations. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention provides an industrial kiln energy data monitoring and processing system and method based on machine learning, which has the advantages of accuracy, reliability, high efficiency, and energy saving, and solves the problem of increased industrial production cost losses.

[0005] In order to solve the above technical problem of increased industrial production cost loss, the present invention provides the following technical solution:

[0006] This embodiment discloses a method for monitoring and processing industrial furnace energy data based on machine learning, which specifically includes the following steps:

[0007] S1. Install monitoring equipment on the side wall of the industrial kiln, collect industrial kiln data through the installed monitoring equipment and pre-process it to obtain pre-processed industrial kiln data;

[0008] S2. Performing three-dimensional modeling of the industrial kiln based on the obtained pre-processed industrial kiln data, predicting the three-dimensional modeling process of the industrial kiln through a support vector machine algorithm, and calculating the error generated in the three-dimensional modeling process of the industrial kiln to obtain predicted three-dimensional model data;

[0009] S3. Based on the predicted three-dimensional model data, a preliminary industrial kiln energy-saving evaluation system is established through the calculation method of energy-saving amount of industrial kiln transformation;

[0010] S4, transforming the obtained predicted three-dimensional model data to obtain transformed three-dimensional model data;

[0011] S5. Based on the established industrial kiln energy-saving evaluation system, the transformed three-dimensional model data is optimized through genetic algorithm to obtain the optimal three-dimensional model data;

[0012] S6. Modify the industrial kiln according to the obtained optimal three-dimensional model data and perform real-time monitoring.

[0013] The present invention installs monitoring equipment on the side wall of an industrial kiln, collects industrial kiln data through the installed monitoring equipment, performs preprocessing and three-dimensional modeling, and predicts the three-dimensional modeling process of the industrial kiln through a support vector machine algorithm to obtain predicted three-dimensional model data; based on the obtained predicted three-dimensional model data, an industrial kiln transformation energy saving calculation method is used to preliminarily establish an industrial kiln energy saving evaluation system, and at the same time, the obtained predicted three-dimensional model data is transformed to obtain transformed three-dimensional model data; finally, based on the established industrial kiln energy saving evaluation system, the transformed three-dimensional model data is optimized through a genetic algorithm, and an industrial kiln is produced according to the optimized three-dimensional model data and real-time monitoring is performed, thereby improving the energy saving effect of industrial kiln energy data monitoring.

[0014] Preferably, the step of installing a monitoring device on the side wall of the industrial kiln, collecting industrial kiln data through the installed monitoring device and preprocessing the data, and obtaining the preprocessed industrial kiln data comprises the following steps:

[0015] The equipment is set to collect industrial kiln data including: industrial kiln image data and industrial kiln energy data;

[0016] S11. Install monitoring equipment on the side wall of the industrial kiln, collect industrial kiln data through the installed monitoring equipment, and set the collected industrial kiln data to include industrial kiln image data and industrial kiln energy data;

[0017] S12, preprocessing the industrial kiln energy data to obtain preprocessed industrial kiln energy data;

[0018] S13, based on the obtained preprocessed industrial kiln energy data, preprocess the industrial kiln image data to determine the temperature conditions of each area in the industrial kiln;

[0019] The pre-processed industrial furnace energy data and the determined temperature conditions of each area in the industrial furnace are set as pre-processed industrial furnace data.

[0020] Preferably, the preprocessing of the industrial furnace energy data to obtain the preprocessed industrial furnace energy data comprises the following steps:

[0021] Set the industrial furnace energy data including temperature data and pressure data;

[0022] The temperature data and the pressure data are standardized by means of data standardization;

[0023] The temperature data standardization processing formula is as follows:

[0024] ;

[0025] Among them, X α Indicates the collected temperature data, Indicates the temperature data after standardization; X αmax Indicates the maximum value in the temperature data; X α min Indicates the minimum value in the temperature data;

[0026] The formula for normalizing pressure data is as follows:

[0027] ;

[0028] Among them, X β represents the collected pressure data, Indicates the pressure data after normalization; X β max Indicates the maximum value in the pressure data; X β min Indicates the minimum value in the pressure data;

[0029] The set of standardized temperature data and pressure data is set as the pre-processed industrial furnace energy data.

[0030] Preferably, the preprocessing of the industrial kiln image data based on the obtained preprocessed industrial kiln energy data to determine the temperature conditions of each area in the industrial kiln comprises the following steps:

[0031] The fire point pixel and the non-fire point pixel are determined by the principle that the outer flame temperature has the highest brightness and the flame center temperature has the lowest brightness;

[0032] Since the industrial kiln image data is the industrial kiln thermal radiation image data, the fire point pixel and the non-fire point pixel are determined according to the temperature range and brightness of each area of ​​the industrial kiln thermal radiation image data;

[0033] The area with the highest temperature is set as the boundary of the fire pixel, the area below the area with the highest temperature is the fire pixel, and the area above the area with the highest temperature is the non-fire pixel;

[0034] The temperature difference between the fire point pixel and the non-fire point pixel in the collected industrial kiln image data is calculated, and the calculation formula is as follows:

[0035] ;

[0036] Where, ΔT iIt represents the temperature difference between the fire point pixel and the non-fire point pixel in the pixel channel i in the industrial kiln image data, T imix It represents the temperature of the fire pixel in pixel channel i in the industrial kiln image data, T iB Represents the temperature of the non-fire point pixel in pixel channel i in the industrial kiln image data, V i The wavelength of pixel channel i in the industrial kiln image data, C2, C1 are radiation constants, indicating the degree of thermal radiation attenuation, N imix The brightness of the fire point pixel in pixel channel i in the industrial kiln image data, N iB Represents the brightness of non-fire point pixels in pixel channel i in industrial kiln image data;

[0037] The pre-processed industrial furnace energy data and the determined temperature conditions of each area in the industrial furnace are set as pre-processed industrial furnace data.

[0038] The present invention performs standardized processing on industrial kiln energy data, and based on the standardized industrial kiln energy data, pre-processes industrial kiln image data to determine the temperature conditions of various areas in the industrial kiln, thereby realizing monitoring and analysis of various data of the industrial kiln, and improving the accuracy of industrial kiln energy data monitoring.

[0039] Preferably, the process of predicting the three-dimensional modeling of the industrial kiln by a support vector machine algorithm, calculating the error generated in the three-dimensional modeling of the industrial kiln, and obtaining the predicted three-dimensional model data comprises the following steps:

[0040] The collected industrial kiln data is set as c, and the corresponding output 3D modeling is set as e;

[0041] The mathematical expression of the support vector machine algorithm is as follows:

[0042] ;

[0043] Among them, w represents the weight, b represents the bias value, represents a nonlinear mapping from low-dimensional to high-dimensional space, and f(c) represents the predicted value;

[0044] The loss function of the support vector machine algorithm is as follows:

[0045] ;

[0046] Among them, ε represents the loss function, and L(f(c), e, ε) represents the relationship function between f(c), e and ε.

[0047] The present invention performs three-dimensional modeling of an industrial kiln based on the obtained preprocessed industrial kiln data, calculates the errors generated in the three-dimensional modeling process of the industrial kiln through the mathematical expression of a support vector machine algorithm based on the obtained three-dimensional modeling, and outputs the predicted three-dimensional model data, thereby ensuring the reliability of the three-dimensional modeling process of the industrial kiln.

[0048] Preferably, the method of initially establishing an industrial kiln energy-saving evaluation system based on the obtained predicted three-dimensional model data and through an industrial kiln transformation energy-saving calculation method comprises the following steps:

[0049] The energy saving calculation formula for industrial kiln transformation is as follows:

[0050] ;

[0051] Among them, ΔE represents the energy saving of industrial kiln transformation, E0 represents the energy consumption of industrial kiln before transformation, E1 represents the energy consumption of industrial kiln after transformation, G0 represents the output of industrial kiln before transformation, G1 represents the output of industrial kiln after transformation, and G represents the smaller value of the output of industrial kiln before and after transformation;

[0052] Based on the calculated energy saving amount of industrial kiln renovation, the threshold value of each level of energy saving amount of industrial kiln renovation is set, and at the same time, based on the set threshold value of each level of energy saving amount of industrial kiln renovation, an industrial kiln energy saving evaluation system is established.

[0053] The present invention adopts an industrial kiln transformation energy saving calculation method, establishes an industrial kiln energy saving evaluation system by calculating the energy consumption of the industrial kiln before and after the transformation and the output of the industrial kiln before and after the transformation, and calculates the energy saving of the industrial kiln transformation based on the established industrial kiln energy saving evaluation system, thereby improving the accuracy of the established industrial kiln energy saving evaluation system.

[0054] Preferably, the step of transforming the obtained predicted three-dimensional model data to obtain the transformed three-dimensional model data comprises the following steps:

[0055] The modification of the obtained predicted three-dimensional model data includes modifying the volume of the predicted three-dimensional model data and adjusting the position of the vent in the predicted three-dimensional model data;

[0056] Set to save the transformed 3D model data after each transformation is completed.

[0057] Preferably, the process of optimizing the transformed three-dimensional model data by genetic algorithm based on the established industrial furnace energy-saving evaluation system to obtain the optimal three-dimensional model data comprises the following steps:

[0058] S51, set the genetic population size, number of iterations and chromosome encoding;

[0059] Set each population as n transformed three-dimensional model data;

[0060] S52, randomly selecting a three-dimensional model data from the three-dimensional model data after transformation of various populations to generate an initial population S gen ;

[0061] S53, using the energy-saving calculation formula of industrial kiln transformation in the industrial kiln energy-saving evaluation system as the fitness function of the genetic algorithm, and calculating the fitness of each individual in the population;

[0062] Each individual is assumed to represent the energy saving of industrial furnace transformation of a set of transformed three-dimensional model data;

[0063] The fitness calculation formula for each individual is as follows:

[0064] ;

[0065] Among them, Fit(k) represents the fitness of the kth individual, and ΔE represents the energy saving of industrial furnace transformation;

[0066] S54, selecting the best individuals from all individuals based on the roulette wheel method;

[0067] S55. Cross the selected excellent individuals through sequential crossover to generate a new population ;

[0068] S56. Randomly select an individual in the population to mutate with a set probability to produce a mutated population ;

[0069] S57, comparing the fitness difference ΔFit between the initial population and the population after crossover mutation by the genetic algorithm;

[0070] When ΔFit<0, it means that the fitness of the mutated population is higher than that of the initial population, and the population is accepted. When ΔFit≥0, it means that the fitness of the mutated population is lower than that of the initial population, and the population is rejected.

[0071] S58, judging whether the maximum number of iterations has been reached according to the number of iterations of the algorithm, outputting the optimal solution when the maximum number of iterations has been reached, and continuing to execute steps S52-S57 when the maximum number of iterations has not been reached;

[0072] The optimal solution to be output is set to be the optimal 3D model data.

[0073] Preferably, the roulette-based method of selecting excellent individuals from all individuals comprises the following steps:

[0074] The calculation formula for selecting the best individuals among all individuals by roulette is as follows:

[0075] ;

[0076] Wherein, p(k) represents the probability of the kth individual being selected, and K represents the amount of the transformed three-dimensional model data.

[0077] The present invention optimizes the transformed three-dimensional model data through a genetic algorithm based on the established industrial kiln energy-saving evaluation system, sets the population size, number of iterations and fitness function, and searches for the optimal energy-saving three-dimensional model data based on a roulette wheel selection method, a sequential crossover method and a random mutation method, thereby ensuring the high efficiency of industrial kiln energy data monitoring.

[0078] This embodiment also discloses an industrial furnace energy data monitoring and processing system based on machine learning, including: a data acquisition module, a data processing module, a three-dimensional modeling module, an energy-saving evaluation module, and a three-dimensional model optimization module;

[0079] The data acquisition module is used to collect industrial kiln data and transmit the collected industrial kiln data to the data processing module;

[0080] The data processing module is used to process the received industrial kiln data and transmit the processed industrial kiln data to the three-dimensional modeling module;

[0081] The three-dimensional modeling module is used to perform three-dimensional modeling on the received processed industrial kiln data and obtain three-dimensional model data;

[0082] The energy-saving assessment module is used to establish an industrial furnace energy-saving assessment system by calculating the energy-saving amount according to the obtained three-dimensional model data;

[0083] The three-dimensional model optimization module is used to optimize the transformed three-dimensional model data through a genetic algorithm according to the establishment of an industrial kiln energy-saving evaluation system.

[0084] Beneficial Effects

[0085] Compared with the prior art, the present invention provides an industrial furnace energy data monitoring and processing system and method based on machine learning, which has the following beneficial effects:

[0086] 1. The invention installs monitoring equipment on the side wall of an industrial kiln, collects industrial kiln data through the installed monitoring equipment, performs preprocessing and three-dimensional modeling, and predicts the three-dimensional modeling process of the industrial kiln through a support vector machine algorithm to obtain predicted three-dimensional model data; based on the obtained predicted three-dimensional model data, an industrial kiln transformation energy saving calculation method is used to preliminarily establish an industrial kiln energy saving evaluation system, and at the same time, the obtained predicted three-dimensional model data is transformed to obtain transformed three-dimensional model data; finally, based on the established industrial kiln energy saving evaluation system, the transformed three-dimensional model data is optimized through a genetic algorithm, and an industrial kiln is produced and real-time monitoring is performed based on the optimized three-dimensional model data, thereby improving the energy-saving effect of industrial kiln energy data monitoring.

[0087] 2. The invention standardizes the industrial kiln energy data, and based on the standardized industrial kiln energy data, pre-processes the industrial kiln image data, determines the temperature conditions of each area in the industrial kiln, realizes the monitoring and analysis of various data of the industrial kiln, and improves the accuracy of industrial kiln energy data monitoring.

[0088] 3. The invention performs three-dimensional modeling of the industrial kiln based on the preprocessed industrial kiln data, calculates the errors generated in the three-dimensional modeling process of the industrial kiln through the mathematical expression of the support vector machine algorithm based on the obtained three-dimensional modeling, and outputs the predicted three-dimensional model data, thereby ensuring the reliability of the three-dimensional modeling process of the industrial kiln.

[0089] 4. The invention adopts an energy-saving calculation method for industrial kiln transformation, establishes an industrial kiln energy-saving evaluation system by calculating the energy consumption of the industrial kiln before and after the transformation and the output of the industrial kiln before and after the transformation, and calculates the energy-saving amount of the industrial kiln transformation based on the established industrial kiln energy-saving evaluation system, thereby improving the accuracy of the established industrial kiln energy-saving evaluation system.

[0090] 5. This invention optimizes the transformed three-dimensional model data through genetic algorithms based on the established industrial kiln energy-saving evaluation system, sets the population size, number of iterations and fitness function, and searches for the optimal energy-saving three-dimensional model data based on roulette selection, sequential crossover and random mutation methods, thereby ensuring the high efficiency of industrial kiln energy data monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0091] Figure 1 It is a schematic diagram of the industrial furnace energy data monitoring process structure of the present invention. DETAILED DESCRIPTION

[0092] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0093] Example 1

[0094] See also Figure 1 This embodiment discloses a method for monitoring and processing industrial furnace energy data based on machine learning, which specifically includes the following steps:

[0095] S1. Install monitoring equipment on the side wall of the industrial kiln, collect industrial kiln data through the installed monitoring equipment and pre-process it to obtain pre-processed industrial kiln data;

[0096] Monitoring equipment is installed on the side wall of the industrial kiln. The industrial kiln data is collected and preprocessed through the installed monitoring equipment. The preprocessed industrial kiln data is obtained by the following steps:

[0097] The equipment is set to collect industrial kiln data including: industrial kiln image data and industrial kiln energy data;

[0098] S11. Install monitoring equipment on the side wall of the industrial kiln, collect industrial kiln data through the installed monitoring equipment, and set the collected industrial kiln data to include industrial kiln image data and industrial kiln energy data;

[0099] S12, preprocessing the industrial kiln energy data to obtain preprocessed industrial kiln energy data;

[0100] Set the industrial furnace energy data including temperature data and pressure data;

[0101] The temperature data and the pressure data are standardized by means of data standardization;

[0102] The temperature data standardization processing formula is as follows:

[0103] ;

[0104] Among them, X α Indicates the collected temperature data, Indicates the temperature data after standardization; X αmax Indicates the maximum value in the temperature data; X α min Indicates the minimum value in the temperature data;

[0105] The formula for normalizing pressure data is as follows:

[0106] ;

[0107] Among them, X β represents the collected pressure data, Indicates the pressure data after normalization; X β max Indicates the maximum value in the pressure data; X β min Indicates the minimum value in the pressure data;

[0108] The set of standardized temperature data and pressure data is set as pre-processed industrial furnace energy data;

[0109] S13, based on the obtained preprocessed industrial kiln energy data, preprocess the industrial kiln image data to determine the temperature conditions of each area in the industrial kiln;

[0110] The fire point pixel and the non-fire point pixel are determined by the principle that the outer flame temperature has the highest brightness and the flame center temperature has the lowest brightness;

[0111] Since the industrial kiln image data is the industrial kiln thermal radiation image data, the fire point pixel and the non-fire point pixel are determined according to the temperature range and brightness of each area of ​​the industrial kiln thermal radiation image data;

[0112] The area with the highest temperature is set as the boundary of the fire pixel, the area below the area with the highest temperature is the fire pixel, and the area above the area with the highest temperature is the non-fire pixel;

[0113] The temperature difference between the fire point pixel and the non-fire point pixel in the collected industrial kiln image data is calculated, and the calculation formula is as follows:

[0114] ;

[0115] Where, ΔT i It represents the temperature difference between the fire point pixel and the non-fire point pixel in the pixel channel i in the industrial kiln image data, T imix It represents the temperature of the fire pixel in pixel channel i in the industrial kiln image data, T iB Represents the temperature of the non-fire point pixel in pixel channel i in the industrial kiln image data, V i The wavelength of pixel channel i in the industrial kiln image data, C2, C1 are radiation constants, indicating the degree of thermal radiation attenuation, N imix The brightness of the fire point pixel in pixel channel i in the industrial kiln image data, N iB Represents the brightness of non-fire point pixels in pixel channel i in industrial kiln image data;

[0116] The obtained pre-processed industrial furnace energy data and the determined temperature conditions of each area in the industrial furnace are set as pre-processed industrial furnace data;

[0117] S2. Performing three-dimensional modeling of the industrial kiln based on the obtained pre-processed industrial kiln data, predicting the three-dimensional modeling process of the industrial kiln through a support vector machine algorithm, and calculating the error generated in the three-dimensional modeling process of the industrial kiln to obtain predicted three-dimensional model data;

[0118] The three-dimensional modeling process of the industrial kiln is predicted by the support vector machine algorithm, and the error generated in the three-dimensional modeling process of the industrial kiln is calculated to obtain the predicted three-dimensional model data, including the following steps:

[0119] The collected industrial kiln data is set as c, and the corresponding output 3D modeling is set as e;

[0120] The mathematical expression of the support vector machine algorithm is as follows:

[0121] ;

[0122] Among them, w represents the weight, b represents the bias value, represents a nonlinear mapping from low-dimensional to high-dimensional space, and f(c) represents the predicted value;

[0123] The loss function of the support vector machine algorithm is as follows:

[0124] ;

[0125] Among them, ε represents the loss function, and L(f(c), e, ε) represents the relationship function between f(c), e and ε;

[0126] S3. Based on the predicted three-dimensional model data, a preliminary industrial kiln energy-saving evaluation system is established through the calculation method of energy-saving amount of industrial kiln transformation;

[0127] Based on the predicted three-dimensional model data, the energy saving calculation method of industrial kiln transformation is used to preliminarily establish an industrial kiln energy saving evaluation system, which includes the following steps:

[0128] The energy saving calculation formula for industrial kiln transformation is as follows:

[0129] ;

[0130] Among them, ΔE represents the energy saving of industrial kiln transformation, E0 represents the energy consumption of industrial kiln before transformation, E1 represents the energy consumption of industrial kiln after transformation, G0 represents the output of industrial kiln before transformation, G1 represents the output of industrial kiln after transformation, and G represents the smaller value of the output of industrial kiln before and after transformation;

[0131] Based on the calculated energy saving amount of industrial kiln transformation, the threshold of each level of energy saving amount of industrial kiln transformation is set, and at the same time, the energy saving evaluation system of industrial kiln is established based on the set threshold of each level of energy saving amount of industrial kiln transformation;

[0132] S4, transforming the obtained predicted three-dimensional model data to obtain transformed three-dimensional model data;

[0133] The predicted three-dimensional model data is transformed to obtain the transformed three-dimensional model data, including the following steps:

[0134] The modification of the obtained predicted three-dimensional model data includes modifying the volume of the predicted three-dimensional model data and adjusting the position of the vent in the predicted three-dimensional model data;

[0135] Set to save the transformed 3D model data after each transformation is completed;

[0136] S5. Based on the established industrial kiln energy-saving evaluation system, the transformed three-dimensional model data is optimized through genetic algorithm to obtain the optimal three-dimensional model data;

[0137] Based on the established industrial kiln energy-saving evaluation system, the transformed three-dimensional model data is optimized through genetic algorithms to obtain the optimal three-dimensional model data, including the following steps:

[0138] S51, set the genetic population size, number of iterations and chromosome encoding;

[0139] Set each population as n transformed three-dimensional model data;

[0140] S52, randomly selecting a three-dimensional model data from the three-dimensional model data after transformation of various populations to generate an initial population S gen ;

[0141] S53, using the energy-saving calculation formula of industrial kiln transformation in the industrial kiln energy-saving evaluation system as the fitness function of the genetic algorithm, and calculating the fitness of each individual in the population;

[0142] Each individual is assumed to represent the energy saving of industrial furnace transformation of a set of transformed three-dimensional model data;

[0143] The fitness calculation formula for each individual is as follows:

[0144] ;

[0145] Among them, Fit(k) represents the fitness of the kth individual, and ΔE represents the energy saving of industrial furnace transformation;

[0146] S54, selecting the best individuals from all individuals based on the roulette wheel method;

[0147] The calculation formula for selecting the best individuals among all individuals by roulette is as follows:

[0148] ;

[0149] Wherein, p(k) represents the probability of the kth individual being selected, and K represents the number of three-dimensional model data after the transformation;

[0150] S55. Cross the selected excellent individuals through sequential crossover to generate a new population ;

[0151] S56. Randomly select an individual in the population to mutate with a set probability to produce a mutated population ;

[0152] S57, comparing the fitness difference ΔFit between the initial population and the population after crossover mutation by the genetic algorithm;

[0153] When ΔFit<0, it means that the fitness of the mutated population is higher than that of the initial population, and the population is accepted. When ΔFit≥0, it means that the fitness of the mutated population is lower than that of the initial population, and the population is rejected.

[0154] S58, judging whether the maximum number of iterations has been reached according to the number of iterations of the algorithm, outputting the optimal solution when the maximum number of iterations has been reached, and continuing to execute steps S52-S57 when the maximum number of iterations has not been reached;

[0155] The optimal solution to be output is set to be the optimal three-dimensional model data;

[0156] S6. Modify the industrial kiln according to the obtained optimal three-dimensional model data and perform real-time monitoring.

[0157] Example 2

[0158] This embodiment also discloses an industrial furnace energy data monitoring and processing system based on machine learning, including: a data acquisition module, a data processing module, a three-dimensional modeling module, an energy-saving evaluation module, and a three-dimensional model optimization module;

[0159] The data acquisition module is used to collect industrial kiln data and transmit the collected industrial kiln data to the data processing module;

[0160] The data processing module is used to process the received industrial kiln data and transmit the processed industrial kiln data to the three-dimensional modeling module;

[0161] The three-dimensional modeling module is used to perform three-dimensional modeling on the received processed industrial kiln data and obtain three-dimensional model data;

[0162] The energy-saving assessment module is used to establish an industrial furnace energy-saving assessment system by calculating the energy-saving amount according to the obtained three-dimensional model data;

[0163] The three-dimensional model optimization module is used to optimize the transformed three-dimensional model data through a genetic algorithm according to the establishment of an industrial kiln energy-saving evaluation system.

[0164] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for monitoring and processing industrial furnace energy data based on machine learning, characterized in that: The following steps are involved: S1. Install monitoring equipment on the side wall of the industrial kiln, collect industrial kiln data through the installed monitoring equipment and pre-process it to obtain pre-processed industrial kiln data; S2. Performing three-dimensional modeling of the industrial kiln based on the obtained pre-processed industrial kiln data, predicting the three-dimensional modeling process of the industrial kiln through a support vector machine algorithm, and calculating the error generated in the three-dimensional modeling process of the industrial kiln to obtain predicted three-dimensional model data; S3. Based on the predicted three-dimensional model data, a preliminary industrial kiln energy-saving evaluation system is established through the calculation method of energy-saving amount of industrial kiln transformation; S4, transforming the obtained predicted three-dimensional model data to obtain transformed three-dimensional model data; The step of transforming the obtained predicted three-dimensional model data to obtain the transformed three-dimensional model data comprises the following steps: The modification of the obtained predicted three-dimensional model data includes modifying the volume of the predicted three-dimensional model data and adjusting the position of the vent in the predicted three-dimensional model data; Set to save the transformed 3D model data after each transformation is completed; S5. Based on the established industrial kiln energy-saving evaluation system, the transformed three-dimensional model data is optimized through genetic algorithm to obtain the optimal three-dimensional model data; S6. Modify the industrial kiln according to the obtained optimal three-dimensional model data and perform real-time monitoring.

2. The method for monitoring and processing industrial furnace energy data based on machine learning according to claim 1 is characterized in that: The step of installing a monitoring device on the side wall of the industrial kiln, collecting industrial kiln data through the installed monitoring device and preprocessing the data to obtain the preprocessed industrial kiln data includes the following steps: S11. Install monitoring equipment on the side wall of the industrial kiln, collect industrial kiln data through the installed monitoring equipment, and set the collected industrial kiln data to include industrial kiln image data and industrial kiln energy data; Among them, the industrial kiln image data is the industrial kiln thermal radiation image data; S12, preprocessing the industrial kiln energy data to obtain preprocessed industrial kiln energy data; S13, based on the obtained preprocessed industrial kiln energy data, preprocess the industrial kiln image data to determine the temperature conditions of each area in the industrial kiln; The pre-processed industrial furnace energy data and the determined temperature conditions of each area in the industrial furnace are set as pre-processed industrial furnace data.

3. The method for monitoring and processing industrial furnace energy data based on machine learning according to claim 2 is characterized in that: The preprocessing of the industrial furnace energy data to obtain the preprocessed industrial furnace energy data comprises the following steps: Set the industrial furnace energy data including temperature data and pressure data; The temperature data and the pressure data are standardized by means of data standardization; The temperature data standardization processing formula is as follows: ； Among them, X α Indicates the collected temperature data, Indicates the temperature data after standardization; X αmax Indicates the maximum value in the temperature data; X α min Indicates the minimum value in the temperature data; The formula for normalizing pressure data is as follows: ； Among them, X β Indicates the collected pressure data, Indicates the pressure data after normalization; X β max Indicates the maximum value in the pressure data; X β min Indicates the minimum value in the pressure data; The set of standardized temperature data and pressure data is set as the pre-processed industrial furnace energy data.

4. The method for monitoring and processing industrial furnace energy data based on machine learning according to claim 2 is characterized in that: The method of preprocessing the industrial kiln image data based on the obtained preprocessed industrial kiln energy data to determine the temperature conditions of each area in the industrial kiln includes the following steps: The fire point pixel and the non-fire point pixel are determined by the principle that the outer flame temperature has the highest brightness and the flame center temperature has the lowest brightness; Since the industrial kiln image data is the industrial kiln thermal radiation image data, the fire point pixel and the non-fire point pixel are determined according to the temperature range and brightness of each area of ​​the industrial kiln thermal radiation image data; The area with the highest temperature is set as the boundary of the fire pixel, the area below the area with the highest temperature is the fire pixel, and the area above the area with the highest temperature is the non-fire pixel; The temperature difference between the fire point pixel and the non-fire point pixel in the collected industrial kiln image data is calculated, and the calculation formula is as follows: ； Where, ΔT i It represents the temperature difference between the fire point pixel and the non-fire point pixel in the pixel channel i in the industrial kiln image data, T imix It represents the temperature of the fire pixel in pixel channel i in the industrial kiln image data, T iB Represents the temperature of the non-fire point pixel in pixel channel i in the industrial kiln image data, V i The wavelength of pixel channel i in the industrial kiln image data, C2, C1 are radiation constants, indicating the degree of thermal radiation attenuation, N imix The brightness of the fire point pixel in pixel channel i in the industrial kiln image data, N iB Represents the brightness of non-fire point pixels in pixel channel i in industrial kiln image data.

5. The method for monitoring and processing industrial furnace energy data based on machine learning according to claim 1 is characterized in that: The method of predicting the three-dimensional modeling process of the industrial kiln by the support vector machine algorithm and calculating the error generated in the three-dimensional modeling process of the industrial kiln to obtain the predicted three-dimensional model data includes the following steps: The collected industrial kiln data is set as c, and the corresponding output 3D modeling is set as e; The mathematical expression of the support vector machine algorithm is as follows: ； Among them, w represents the weight, b represents the bias value, represents a nonlinear mapping from low-dimensional to high-dimensional space, and f(c) represents the predicted value; The loss function of the support vector machine algorithm is as follows: ； Among them, ε represents the loss function, and L(f(c), e, ε) represents the relationship function between f(c), e and ε.

6. The method for monitoring and processing industrial furnace energy data based on machine learning according to claim 1 is characterized in that: Based on the predicted three-dimensional model data, the energy saving calculation method of industrial kiln transformation is used to preliminarily establish an industrial kiln energy saving evaluation system The following steps are involved: The energy saving calculation formula for industrial kiln transformation is as follows: ； Among them, ΔE represents the energy saving of industrial kiln transformation, E0 represents the energy consumption of industrial kiln before transformation, E1 represents the energy consumption of industrial kiln after transformation, G0 represents the output of industrial kiln before transformation, G1 represents the output of industrial kiln after transformation, and G represents the smaller value of the output of industrial kiln before and after transformation; Based on the calculated energy saving amount of industrial kiln renovation, the threshold value of each level of energy saving amount of industrial kiln renovation is set, and at the same time, based on the set threshold value of each level of energy saving amount of industrial kiln renovation, an industrial kiln energy saving evaluation system is established.

7. The method for monitoring and processing industrial furnace energy data based on machine learning according to claim 1 is characterized in that: The method of optimizing the transformed three-dimensional model data by genetic algorithm based on the established industrial kiln energy-saving evaluation system to obtain the optimal three-dimensional model data includes the following steps: S51, set the genetic population size, number of iterations and chromosome encoding; Set each population as n transformed three-dimensional model data; S52, randomly selecting a three-dimensional model data from the three-dimensional model data after transformation of various populations to generate an initial population S gen ; S53, using the energy-saving calculation formula of industrial kiln transformation in the industrial kiln energy-saving evaluation system as the fitness function of the genetic algorithm, and calculating the fitness of each individual in the population; Each individual is assumed to represent the energy saving of industrial furnace transformation of a set of transformed three-dimensional model data; The fitness calculation formula for each individual is as follows: ； Among them, Fit(k) represents the fitness of the kth individual, and ΔE represents the energy saving of industrial furnace transformation; S54, selecting the best individuals from all individuals based on the roulette wheel method; S55. Cross the selected excellent individuals through sequential crossover to generate a new population ; S56. Randomly select an individual in the population to mutate with a set probability to produce a mutated population ; S57, comparing the fitness difference ΔFit between the initial population and the population after crossover mutation by the genetic algorithm; When ΔFit<0, it means that the fitness of the mutated population is higher than that of the initial population, and the population is accepted. When ΔFit≥0, it means that the fitness of the mutated population is lower than that of the initial population, and the population is rejected. S58, judging whether the maximum number of iterations has been reached according to the number of iterations of the algorithm, outputting the optimal solution when the maximum number of iterations has been reached, and continuing to execute steps S52-S57 when the maximum number of iterations has not been reached; The optimal solution to be output is set to be the optimal 3D model data.

8. The method for monitoring and processing industrial furnace energy data based on machine learning according to claim 7 is characterized in that: The roulette-based method for selecting excellent individuals from all individuals comprises the following steps: The calculation formula for selecting the best individuals among all individuals by roulette is as follows: ； Wherein, p(k) represents the probability of the kth individual being selected, and K represents the amount of the transformed three-dimensional model data.

9. A system for implementing the industrial furnace energy data monitoring and processing method based on machine learning as described in any one of claims 1 to 8, characterized in that: include: Data acquisition module, data processing module, 3D modeling module, energy saving assessment module and 3D model optimization module; The data acquisition module is used to collect industrial kiln data and transmit the collected industrial kiln data to the data processing module; The data processing module is used to process the received industrial kiln data and transmit the processed industrial kiln data to the three-dimensional modeling module; The three-dimensional modeling module is used to perform three-dimensional modeling on the received processed industrial kiln data and obtain three-dimensional model data; The energy-saving assessment module is used to establish an industrial furnace energy-saving assessment system by calculating the energy-saving amount according to the obtained three-dimensional model data; The three-dimensional model optimization module is used to optimize the transformed three-dimensional model data through a genetic algorithm according to the establishment of an industrial kiln energy-saving evaluation system.

Images