CN119046801A - Agricultural machinery operation state monitoring method and system based on Internet of things - Google Patents
Agricultural machinery operation state monitoring method and system based on Internet of things Download PDFInfo
- Publication number
- CN119046801A CN119046801A CN202411550164.1A CN202411550164A CN119046801A CN 119046801 A CN119046801 A CN 119046801A CN 202411550164 A CN202411550164 A CN 202411550164A CN 119046801 A CN119046801 A CN 119046801A
- Authority
- CN
- China
- Prior art keywords
- data
- target domain
- agricultural machinery
- running state
- state monitoring
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000012544 monitoring process Methods 0.000 title claims abstract description 108
- 238000000034 method Methods 0.000 title claims abstract description 52
- 238000009826 distribution Methods 0.000 claims abstract description 85
- 230000008859 change Effects 0.000 claims abstract description 16
- 238000010276 construction Methods 0.000 claims abstract description 16
- 238000013507 mapping Methods 0.000 claims abstract description 15
- 238000007781 pre-processing Methods 0.000 claims abstract description 15
- 238000012549 training Methods 0.000 claims description 39
- 238000004140 cleaning Methods 0.000 claims description 13
- 238000010606 normalization Methods 0.000 claims description 13
- 238000006243 chemical reaction Methods 0.000 claims description 11
- 230000004913 activation Effects 0.000 claims description 8
- 238000011176 pooling Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 8
- 230000006872 improvement Effects 0.000 claims description 7
- 238000005070 sampling Methods 0.000 claims description 6
- 238000012423 maintenance Methods 0.000 claims description 4
- 230000000739 chaotic effect Effects 0.000 claims description 3
- 238000009499 grossing Methods 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 239000000284 extract Substances 0.000 abstract description 8
- 238000013461 design Methods 0.000 abstract description 2
- 230000009286 beneficial effect Effects 0.000 description 7
- 230000008569 process Effects 0.000 description 5
- 230000009471 action Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A40/00—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production
- Y02A40/10—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production in agriculture
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an agricultural machinery operation state monitoring method and system based on the Internet of things, wherein the method comprises the following steps: data acquisition, data preprocessing, construction of an agricultural machine running state monitoring model, agricultural machine running state monitoring model parameter searching and real-time running state monitoring. The invention belongs to the technical field of fault monitoring, in particular to an agricultural machinery running state monitoring method and system based on the Internet of things, wherein the scheme extracts the characteristics of source domain and target domain data and completes data reconstruction, calculates the difference of source domain and target domain distribution by reproducing the Hilbert space of a kernel, aligns the data distribution of the two domains by minimizing the difference, and designs a total loss function; and initializing individual positions based on the improved tent mapping, improving convergence factors based on the Gaussian distribution change curve, calculating optimal individual positions based on a proportion weighting strategy, and updating all individual positions by combining the improved convergence factors and the optimal individual positions, so that the monitoring accuracy and efficiency of the running state of the agricultural machinery are improved.
Description
Technical Field
The invention belongs to the technical field of fault monitoring, and particularly relates to an agricultural machinery operation state monitoring method and system based on the Internet of things.
Background
The method for monitoring the running state of the agricultural machine comprehensively utilizes the Internet and artificial intelligence technology to realize real-time monitoring of the running state of the agricultural machine and improve the use efficiency and safety of the agricultural machine. However, the existing agricultural machine running state monitoring method has the problems that the agricultural machine running characteristics are difficult to accurately extract, the agricultural machine running data distribution among different domains is different, the relevance among data categories is ignored during classification, and the classification accuracy is low, and the existing agricultural machine running state monitoring method has the problems that the adaptability to different agricultural machine types and works is poor, the monitoring effect performance is poor under the new environment and the new data, and the generalization capability is weak.
Disclosure of Invention
Aiming at the problems that the agricultural machine operation state monitoring method and system based on the Internet of things are difficult to accurately extract the agricultural machine operation features, the agricultural machine operation data distribution among different domains is different, and the relevance among data categories is ignored during classification, so that the classification accuracy is low, the method aims at solving the problems that the existing agricultural machine operation state monitoring method is difficult to accurately extract the agricultural machine operation features, and the classification accuracy is low; the method comprises the steps of designing a symmetrical structure of a decoder, realizing reconstruction of data, maintaining the integrity and accuracy of the data, calculating the difference of source domain and target domain distribution through a Hilbert space of a reproduction core, aligning the data distribution of two domains through minimizing difference loss, improving the adaptability of a model to different data domains, introducing a smooth tag algorithm, optimizing the coding of data class tags, designing a total loss function based on the reconstruction loss, the difference loss and the classification loss, facilitating more accurate understanding and analysis of the operation state data of the agricultural machinery, improving the monitoring accuracy and efficiency of the operation state of the agricultural machinery, aiming at the problems that the existing operation state monitoring method of the agricultural machinery has poor adaptability to different agricultural machinery types and works, poor monitoring effect under new environment and new data and weak generalization capability, generating the initial position of an individual by introducing a method for improving tent mapping by random variables, facilitating the increase of the diversity of searching, adopting a convergence factor based on a Gaussian distribution change curve, the method is beneficial to balancing the capability of local search and global search in the search process, is beneficial to improving the convergence and stability of an optimization algorithm by designing a proportion weighting strategy based on fitness and first three optimal individual positions, is beneficial to improving the convergence factor and the optimal individual positions based on the improvement of Gaussian distribution change curve, is beneficial to improving the search efficiency and accuracy, improves the comprehensiveness and accuracy of monitoring various agricultural machinery and working operation states, and can better guide the use, maintenance and management of the agricultural machinery.
The technical scheme adopted by the invention is that the agricultural machinery running state monitoring method based on the Internet of things comprises the following steps:
Step S1, data acquisition;
S2, preprocessing data;
S3, constructing an agricultural machinery running state monitoring model;
s4, searching parameters of an agricultural machinery running state monitoring model;
and S5, monitoring the real-time running state.
Further, in step S1, the data acquisition is to deploy the Internet of things equipment to the agricultural machinery, acquire historical operation data and real-time operation data of the agricultural machinery based on the Internet of things equipment, wherein the historical operation data and the real-time operation data comprise agricultural machinery equipment information, environment parameter data, operation track data, sensor data and maintenance records, the historical operation data also comprise agricultural machinery operation states, the agricultural machinery operation states comprise normal and agricultural machinery fault types, the operation states are used as data type labels, and the acquired data are transmitted to a central processing center.
Further, in step S2, the data preprocessing is to perform data cleaning, data conversion, data normalization and data set construction processing on the collected data through a central processing center, the data cleaning includes noise processing, repeated values, missing values and abnormal values, the data conversion is to convert the data after the data cleaning into a vector form, the data normalization is to unify a data range based on a maximum and minimum normalization method, and the data set construction is to respectively construct a source domain training set with a data type label and a target domain training set without the data type label based on the historical operation data and the real-time operation data after the data cleaning, the data conversion and the data normalization.
Further, in step S3, the construction of the agricultural machinery operation state monitoring model specifically includes the following steps:
step S31, extracting features, designing an encoder for extracting features from input data, wherein the encoder consists of 5 one-dimensional convolution blocks, each convolution block comprises a convolution layer and a maximum pooling layer, the convolution blocks are activated by RELU functions, the size of a first convolution kernel is 1 multiplied by 64, the sizes of the other four convolution kernels are 1 multiplied by 3, the step size of the maximum pooling layer is 1 multiplied by 2, and the formula for extracting the features is as follows:
;
;
Wherein, Q y and Q m are respectively the source domain feature and the target domain feature extracted from the source domain training set and the target domain training set, enc is the network structure of the encoder, and X y and X m are respectively the source domain training set and the target domain training set;
Step S32, data reconstruction, designing a decoder for data reconstruction, wherein the decoder consists of 5 one-dimensional deconvolution blocks, each deconvolution block comprises a deconvolution layer and an up-sampling layer, the structure of the decoder is completely symmetrical with the encoder, the size of a first deconvolution core is 1 multiplied by 64, the sizes of the other four deconvolution cores are 1 multiplied by 3, and the step length of the up-sampling layer is 1 multiplied by 2, and the method comprises the following steps:
step S321, calculating reconstructed target domain data, wherein the formula is as follows:
;
In the formula, Is the reconstructed target domain data, dec is the network structure of the decoder;
step S322, calculating reconstruction loss of the target domain data, wherein the following formula is adopted:
;
Where L rec is the reconstruction loss of the target domain data, N m is the number of target domain data in the target domain training set, AndThe method comprises the steps of respectively reconstructing the ith target domain data before reconstruction and the ith target domain data after reconstruction, wherein i is a target domain data index;
step S33 of domain alignment, which maps both the source domain data and the target domain data to the hilbert space having the reproduction core, aligns the distributions of the two domains by minimizing the difference between the source domain distribution and the target domain distribution, comprising the steps of:
step S331, calculating the difference between the source domain distribution and the target domain distribution, wherein the formula is as follows:
;
;
;
In the formula, Is the difference between the source domain distribution and the target domain distribution, p （c） and q （c） are the source domain distribution and the target domain distribution, respectively, l is the feature layer index, C is the total number of data class labels, C is the data class label index, N y is the amount of source domain data in the source domain training set,AndThe weights of the ith and jth source domain data belonging to class c data class labels in the source domain training set,AndThe weights of the ith and jth target domain data belonging to class c data class labels in the target domain training set,AndThe number of source domain data and target domain data in the source domain training set and the target domain training set, respectively, which belong to the same class c as the ith source domain data and target domain data, k (·) is a kernel function of the hilbert space with a reproduction kernel,AndThe activation values on the ith and jth source domain data of the first layer feature layer,AndThe activation values on the ith and jth target domain data of the first layer of feature layers respectively;
Step S332, calculating the difference loss between the source domain distribution and the target domain distribution, wherein the following formula is adopted:
;
where L LMMD is the difference loss of source domain distribution and target domain distribution, L max is the number of feature layers;
Step S34, classifying, namely inputting the extracted target domain characteristics into a softmax function to calculate the probability that the target domain data belongs to each data class label, and selecting the data class label with the highest probability as an output label of the target domain data on a model, wherein the step comprises the following steps:
Step S341, introducing a smooth tag algorithm to optimize the one-hot coding of the real data category tag, wherein the following formula is adopted:
;
Where epsilon is the smoothing parameter, Is thatIs used for the identification of the real data tag of (c),Is thatThe one-hot encoding of the real data label of (2), C is the total number of the data class labels, and C is the index of the data class labels;
step S342, classifying the loss, wherein the formula is as follows:
;
where L cls is the classification loss, AndRespectively areBelonging to class c and in the target domain training setThe probability of class data class labels, W and b are weight and bias, respectively;
step S35, designing a total loss function of the model, and combining the reconstruction loss, the difference loss and the classification loss according to the weight to obtain the total loss function of the model, wherein the formula is as follows:
;
Where L is the total loss function, β 1、β2 and β 3 are the first, second and third weights, respectively, and L LMMD is the difference loss of the source domain distribution and the target domain distribution.
Further, in step S4, the searching of the agricultural machinery operation state monitoring model parameters specifically includes the following steps:
Step S41, an initial individual position, a parameter search space is built for a first weight beta 1, a second weight beta 2, a third weight beta 3, a weight W and a bias b in an agricultural machine running state monitoring model, random variable improvement tent mapping is introduced, F individual positions are initialized in the parameter search space based on the improved tent mapping, the individual positions are used as representative of agricultural machine running state monitoring model parameters, a total loss function of the agricultural machine running state monitoring model trained based on the individual positions is used as an individual fitness value, and the initial individual position is represented by the following formula:
;
;
Wherein A w+1 and A w are respectively w+1th and w improved chaotic sequences, v is a control value, lb and ub are respectively the lower limit and the upper limit of a parameter search space, D w is the initial position of a w individual, and rand (·) is a random number generation function;
step S42, improving convergence factor based on Gaussian distribution change curve, wherein the formula is as follows:
;
wherein a (G) is a convergence factor in the G-th iterative search, G is a search number index, G is a total number of searches, and η is a cutoff value;
Step S43, calculating an optimal individual position, updating an individual fitness value, sequencing according to the sequence from small fitness to large fitness, selecting the first three individual positions as a first optimal individual position, a second optimal individual position and a third optimal individual position, and designing a proportional weighting strategy based on the fitness and the first three optimal individual positions to calculate the optimal individual position, wherein the formula is as follows:
;
;
;
Wherein U o (g) and Z o (g) are the fitness ratio and the position ratio of the o-th optimal individual position at the g-th iterative search, U 1（g）、u2 (g) and U 3 (g) are the fitness ratio and the position ratio of the first optimal individual position at the g-th iterative search, the second optimal individual position and the third optimal individual position at the g-th iterative search, U3928 (g) and Z 2 (g) are the fitness ratio and the position ratio of the first optimal individual position at the g-th iterative search, U o (g) is the fitness of the o-th optimal individual position at the g-th iterative search, U3825 (g) is the fitness ratio and the position ratio of the first optimal individual position at the g-th iterative search, U 2 (g) and Z 1 (g) are the fitness ratio and the position ratio of the second optimal individual position at the g-th iterative search, U 2 (g) and the Z 2 (g) are the fitness ratio and the position of the o-th optimal individual position at the g-th iterative search, U o (g) is the fitness ratio and the o-th optimal individual position at the g-th iterative search;
Step S44, updating the individual position according to the following formula:
;
Wherein D w (g+1) and D w (g) are the positions of the w-th individual at the g+1st and g-th iterative searches, and S 1 and S 2 are the first and second random values, respectively;
Step S45, determining optimal parameters of the model, presetting an fitness threshold, updating fitness of an individual, taking the position of the individual with the minimum fitness as the optimal parameters of the model when the minimum fitness is lower than the fitness threshold, constructing an agricultural machinery running state monitoring model based on the optimal parameters of the model, otherwise, turning to step S41 if the total number of searches is reached, otherwise, turning to step S42, and continuing iterative search.
Further, in step S5, the real-time operation state monitoring is based on the output label of the model to obtain the real-time operation state of the agricultural machine, so as to realize the real-time monitoring of the agricultural machine.
The invention provides an agricultural machinery running state monitoring system based on the Internet of things, which comprises a data acquisition module, a data preprocessing module, an agricultural machinery running state monitoring model building module, an agricultural machinery running state monitoring model parameter searching module and a real-time running state monitoring module, wherein the data acquisition module is used for acquiring data of the agricultural machinery running state monitoring model;
The data acquisition module acquires historical operation data and real-time operation data of the agricultural machinery based on the Internet of things equipment and sends the data to the data preprocessing module;
The data preprocessing module performs data cleaning, data conversion, data normalization and data set construction processing on the collected data through the central processing center, and sends the data to the agricultural machinery operation state monitoring model construction module;
The method comprises the steps that an encoder is designed by a model module for constructing an agricultural machine running state monitoring model to extract the characteristics of source domain data and target domain data, a decoder is designed to complete data reconstruction, the difference of source domain distribution and target domain distribution is calculated through a Hilbert space of a reproduction core, the data distribution of the two domains is aligned through minimized difference, a smooth tag algorithm is introduced to optimize the encoding of a data class tag, a total loss function is designed based on reconstruction loss, difference loss and classification loss, and data is sent to a model parameter searching module for agricultural machine running state monitoring;
The agricultural machinery running state monitoring model parameter searching module introduces random variable improvement tent mapping, further initializes individual positions, improves convergence factors based on Gaussian distribution change curves, calculates optimal individual positions based on a proportion weighting strategy, updates all individual positions by combining the improved convergence factors and the optimal individual positions to obtain model optimal parameters, and sends data to the real-time running state monitoring module;
The real-time running state monitoring module obtains the real-time running state of the agricultural machine based on the output label of the model, and realizes real-time monitoring of the agricultural machine.
By adopting the scheme, the beneficial effects obtained by the invention are as follows:
(1) Aiming at the problems that the existing agricultural machine running state monitoring method is difficult to accurately extract the agricultural machine running characteristics, the distribution of agricultural machine running data among different domains is different, and the relevance among data categories is ignored during classification, so that the classification accuracy is low, the method adopts a plurality of one-dimensional convolution blocks, combines a maximum pooling layer and RELU activation functions, specifically extracts the characteristics of data of a source domain and a target domain, maps high-dimensional data into a low-dimensional space, effectively reduces the dimensionality of the data, retains effective information, calculates the distribution difference of the source domain and the target domain through the Hilbert space of a reproduction core, improves the adaptability of a model to different data domains by minimizing the difference loss alignment of the two domains, introduces a smooth tag algorithm, optimizes the coding of the data category tags, and is beneficial to more accurately understanding and analyzing the agricultural machine running state data based on reconstruction loss, difference loss and classification loss design total loss function, and improves the monitoring accuracy and efficiency of the agricultural machine running state.
(2) Aiming at the problems that the existing agricultural machine running state monitoring method is poor in adaptability to different agricultural machine types and works, poor in monitoring effect and weak in generalization capability under new environment and new data, the method for improving tent mapping by introducing random variables is used for generating initial positions of F individuals, increasing searching diversity and reducing possibility of sinking into local optimal solutions, a convergence factor based on Gaussian distribution change curve is adopted, the value of the convergence factor is gradually reduced according to iterative searching times, the capacity of balancing local searching and global searching in the searching process is facilitated, the convergence speed and the searching efficiency are improved, the adaptability and the position information of the individual positions are comprehensively considered by designing a proportional weighting strategy based on the adaptability and the first three optimal individual positions, the optimal individual positions are effectively determined, the convergence and the stability of an optimization algorithm are facilitated to be improved, the improved convergence factor and the optimal individual positions based on Gaussian distribution change curve are combined, the searching efficiency and the accuracy are facilitated to be improved, the comprehensive and the accuracy of monitoring on various agricultural machines and the working running states are improved, and the agricultural machine can be better guided to be used, maintained and managed.
Drawings
FIG. 1 is a schematic flow chart of an agricultural machinery operation state monitoring method based on the Internet of things;
FIG. 2 is a schematic diagram of an agricultural machinery operation state monitoring system based on the Internet of things, which is provided by the invention;
FIG. 3 is a flow chart of step S3;
Fig. 4 is a flow chart of step S4.
The accompanying drawings are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate the invention and together with the embodiments of the invention, serve to explain the invention.
Detailed Description
The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, but not all embodiments, and all other embodiments obtained by those skilled in the art without making any inventive effort based on the embodiments of the present invention are within the scope of protection of the present invention.
In the description of the present invention, it should be understood that the terms "upper," "lower," "front," "rear," "left," "right," "top," "bottom," "inner," "outer," and the like indicate orientation or positional relationships based on those shown in the drawings, merely to facilitate description of the invention and simplify the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus should not be construed as limiting the invention.
Referring to fig. 1, the method for monitoring the running state of an agricultural machine based on the internet of things provided by the invention comprises the following steps:
step S1, data acquisition, namely acquiring historical operation data and real-time operation data of an agricultural machine based on Internet of things equipment;
Step S2, data preprocessing, namely performing data cleaning, data conversion, data normalization and data set construction processing on the acquired data through a central processing center;
S3, constructing an agricultural machinery running state monitoring model, designing an encoder to extract the characteristics of source domain and target domain data, designing a decoder to complete data reconstruction, calculating the difference of source domain and target domain distribution through the Hilbert space of a reproduction core, aligning the data distribution of the two domains through the minimum difference, introducing a smooth tag algorithm to optimize the encoding of a data class tag, and designing a total loss function based on reconstruction loss, difference loss and classification loss;
step S4, searching parameters of an agricultural machinery running state monitoring model, introducing a random variable to improve tent mapping, initializing individual positions, improving convergence factors based on Gaussian distribution change curves, calculating optimal individual positions based on a proportion weighting strategy, and updating all individual positions by combining the improved convergence factors and the optimal individual positions to obtain optimal parameters of the model;
and S5, monitoring the real-time running state, and obtaining the real-time running state of the agricultural machine based on the output label of the model to realize the real-time monitoring of the agricultural machine.
In step S1, the data collection is to deploy the internet of things device to the agricultural machine, collect historical operation data and real-time operation data of the agricultural machine based on the internet of things device, where the historical operation data and the real-time operation data include information of the agricultural machine, environmental parameter data, operation track data, sensor data and maintenance records, the historical operation data includes operation states of the agricultural machine, the operation states include normal and agricultural machine fault types, and the operation states are used as data type labels, and the collected data are transmitted to a central processing center.
Referring to fig. 1, in step S2, the data cleaning includes processing noise, a repetition value, a missing value, and an outlier, the data conversion is to convert the data after the data cleaning into a vector form, the data normalization is to unify a data range based on a maximum and minimum normalization method, and the data set construction is to respectively construct a source domain training set with a data type label and a target domain training set without a data type label based on the data cleaning, the data conversion, the historical operation data after the data normalization, and the real-time operation data.
In a fourth embodiment, referring to fig. 1 and 3, the method for constructing an agricultural machine operation state monitoring model in step S3 specifically includes the following steps:
step S31, extracting features, designing an encoder for extracting features from input data, wherein the encoder consists of 5 one-dimensional convolution blocks, each convolution block comprises a convolution layer and a maximum pooling layer, the convolution blocks are activated by RELU functions, the size of a first convolution kernel is 1 multiplied by 64, the sizes of the other four convolution kernels are 1 multiplied by 3, the step size of the maximum pooling layer is 1 multiplied by 2, and the formula for extracting the features is as follows:
;
;
Wherein, Q y and Q m are respectively the source domain feature and the target domain feature extracted from the source domain training set and the target domain training set, enc is the network structure of the encoder, and X y and X m are respectively the source domain training set and the target domain training set;
Step S32, data reconstruction, designing a decoder for data reconstruction, wherein the decoder consists of 5 one-dimensional deconvolution blocks, each deconvolution block comprises a deconvolution layer and an up-sampling layer, the structure of the decoder is completely symmetrical with the encoder, the size of a first deconvolution core is 1 multiplied by 64, the sizes of the other four deconvolution cores are 1 multiplied by 3, and the step length of the up-sampling layer is 1 multiplied by 2, and the method comprises the following steps:
step S321, calculating reconstructed target domain data, wherein the formula is as follows:
;
In the formula, Is the reconstructed target domain data, dec is the network structure of the decoder;
step S322, calculating reconstruction loss of the target domain data, wherein the following formula is adopted:
;
Where L rec is the reconstruction loss of the target domain data, N m is the number of target domain data in the target domain training set, AndThe method comprises the steps of respectively reconstructing the ith target domain data before reconstruction and the ith target domain data after reconstruction, wherein i is a target domain data index;
step S33 of domain alignment, which maps both the source domain data and the target domain data to the hilbert space having the reproduction core, aligns the distributions of the two domains by minimizing the difference between the source domain distribution and the target domain distribution, comprising the steps of:
step S331, calculating the difference between the source domain distribution and the target domain distribution, wherein the formula is as follows:
;
;
;
In the formula, Is the difference between the source domain distribution and the target domain distribution, p （c） and q （c） are the source domain distribution and the target domain distribution, respectively, l is the feature layer index, C is the total number of data class labels, C is the data class label index, N y is the amount of source domain data in the source domain training set,AndThe weights of the ith and jth source domain data belonging to class c data class labels in the source domain training set,AndThe weights of the ith and jth target domain data belonging to class c data class labels in the target domain training set,AndThe number of source domain data and target domain data in the source domain training set and the target domain training set, respectively, which belong to the same class c as the ith source domain data and target domain data, k (·) is a kernel function of the hilbert space with a reproduction kernel,AndThe activation values on the ith and jth source domain data of the first layer feature layer,AndThe activation values on the ith and jth target domain data of the first layer of feature layers respectively;
Step S332, calculating the difference loss between the source domain distribution and the target domain distribution, wherein the following formula is adopted:
;
where L LMMD is the difference loss of source domain distribution and target domain distribution, L max is the number of feature layers;
Step S34, classifying, namely inputting the extracted target domain characteristics into a softmax function to calculate the probability that the target domain data belongs to each data class label, and selecting the data class label with the highest probability as an output label of the target domain data on a model, wherein the step comprises the following steps:
Step S341, introducing a smooth tag algorithm to optimize the one-hot coding of the real data category tag, wherein the following formula is adopted:
;
Where epsilon is the smoothing parameter, Is thatIs used for the identification of the real data tag of (c),Is thatThe one-hot encoding of the real data label of (2), C is the total number of the data class labels, and C is the index of the data class labels;
step S342, classifying the loss, wherein the formula is as follows:
;
where L cls is the classification loss, AndRespectively areBelonging to class c and in the target domain training setThe probability of class data class labels, W and b are weight and bias, respectively;
step S35, designing a total loss function of the model, and combining the reconstruction loss, the difference loss and the classification loss according to the weight to obtain the total loss function of the model, wherein the formula is as follows:
;
Where L is the total loss function, β 1、β2 and β 3 are the first, second and third weights, respectively, and L LMMD is the difference loss of the source domain distribution and the target domain distribution.
By executing the operation, aiming at the problems that the existing agricultural machine operation state monitoring method has the defects that the agricultural machine operation characteristics are difficult to accurately extract, the distribution of agricultural machine operation data among different domains is different, and the relevance among data categories is ignored during classification, so that the classification accuracy is low, the method adopts a plurality of one-dimensional convolution blocks, combines a maximum pooling layer and RELU activation functions, specifically extracts the characteristics of source domain and target domain data, maps high-dimensional data to a low-dimensional space, effectively reduces the dimensionality of the data, and retains effective information; the method comprises the steps of calculating the difference of distribution of a source domain and a target domain by reproducing the Hilbert space of a kernel, aligning the data distribution of the source domain and the target domain by minimizing the difference loss, improving the adaptability of a model to different data domains, introducing a smooth tag algorithm, optimizing the coding of data type tags, designing a total loss function based on the reconstruction loss, the difference loss and the classification loss, being beneficial to more accurately understanding and analyzing the running state data of the agricultural machinery, and improving the monitoring accuracy and efficiency of the running state of the agricultural machinery.
Fifth embodiment referring to fig. 1 and 4, based on the above embodiment, in step S4, the agricultural machinery operation state monitoring model parameter search specifically includes the following steps:
Step S41, an initial individual position, a parameter search space is built for a first weight beta 1, a second weight beta 2, a third weight beta 3, a weight W and a bias b in an agricultural machine running state monitoring model, random variable improvement tent mapping is introduced, F individual positions are initialized in the parameter search space based on the improved tent mapping, the individual positions are used as representative of agricultural machine running state monitoring model parameters, a total loss function of the agricultural machine running state monitoring model trained based on the individual positions is used as an individual fitness value, and the initial individual position is represented by the following formula:
;
;
Wherein A w+1 and A w are respectively w+1th and w improved chaotic sequences, v is a control value, lb and ub are respectively the lower limit and the upper limit of a parameter search space, D w is the initial position of a w individual, and rand (·) is a random number generation function;
step S42, improving convergence factor based on Gaussian distribution change curve, wherein the formula is as follows:
;
wherein a (G) is a convergence factor in the G-th iterative search, G is a search number index, G is a total number of searches, and η is a cutoff value;
Step S43, calculating an optimal individual position, updating an individual fitness value, sequencing according to the sequence from small fitness to large fitness, selecting the first three individual positions as a first optimal individual position, a second optimal individual position and a third optimal individual position, and designing a proportional weighting strategy based on the fitness and the first three optimal individual positions to calculate the optimal individual position, wherein the formula is as follows:
;
;
;
Wherein U o (g) and Z o (g) are the fitness ratio and the position ratio of the o-th optimal individual position at the g-th iterative search, U 1（g）、u2 (g) and U 3 (g) are the fitness ratio and the position ratio of the first optimal individual position at the g-th iterative search, the second optimal individual position and the third optimal individual position at the g-th iterative search, U3928 (g) and Z 2 (g) are the fitness ratio and the position ratio of the first optimal individual position at the g-th iterative search, U o (g) is the fitness of the o-th optimal individual position at the g-th iterative search, U3825 (g) is the fitness ratio and the position ratio of the first optimal individual position at the g-th iterative search, U 2 (g) and Z 1 (g) are the fitness ratio and the position ratio of the second optimal individual position at the g-th iterative search, U 2 (g) and the Z 2 (g) are the fitness ratio and the position of the o-th optimal individual position at the g-th iterative search, U o (g) is the fitness ratio and the o-th optimal individual position at the g-th iterative search;
Step S44, updating the individual position according to the following formula:
;
Wherein D w (g+1) and D w (g) are the positions of the w-th individual at the g+1st and g-th iterative searches, and S 1 and S 2 are the first and second random values, respectively;
Step S45, determining optimal parameters of the model, presetting an fitness threshold, updating fitness of an individual, taking the position of the individual with the minimum fitness as the optimal parameters of the model when the minimum fitness is lower than the fitness threshold, constructing an agricultural machinery running state monitoring model based on the optimal parameters of the model, otherwise, turning to step S41 if the total number of searches is reached, otherwise, turning to step S42, and continuing iterative search.
According to the method, the initial positions of F individuals are generated, the diversity of searching is increased, the possibility of sinking into a local optimal solution is reduced, the convergence factor based on a Gaussian distribution change curve is adopted, the value of the convergence factor is gradually reduced according to the iterative searching times, the capacity of balancing local searching and global searching in the searching process is facilitated, the convergence speed and the searching efficiency are improved, the fitness and the position information of the individual positions are comprehensively considered, the optimal individual position is effectively determined, the convergence and the stability of an optimization algorithm are improved, the improved convergence factor based on the Gaussian distribution change curve and the optimal individual position are combined, the comprehensive and the accuracy of searching for various agricultural machines and the working states are improved, and the agricultural machine monitoring and the agricultural machine management can be better guided and maintained.
An embodiment six, referring to fig. 2, based on the above embodiment, the system for monitoring the running state of an agricultural machine based on the internet of things provided by the invention comprises a data acquisition module, a data preprocessing module, a module for constructing an agricultural machine running state monitoring model, a parameter searching module for the agricultural machine running state monitoring model and a real-time running state monitoring module;
The data acquisition module acquires historical operation data and real-time operation data of the agricultural machinery based on the Internet of things equipment and sends the data to the data preprocessing module;
The data preprocessing module performs data cleaning, data conversion, data normalization and data set construction processing on the collected data through the central processing center, and sends the data to the agricultural machinery operation state monitoring model construction module;
The method comprises the steps that an encoder is designed by a model module for constructing an agricultural machine running state monitoring model to extract the characteristics of source domain data and target domain data, a decoder is designed to complete data reconstruction, the difference of source domain distribution and target domain distribution is calculated through a Hilbert space of a reproduction core, the data distribution of the two domains is aligned through minimized difference, a smooth tag algorithm is introduced to optimize the encoding of a data class tag, a total loss function is designed based on reconstruction loss, difference loss and classification loss, and data is sent to a model parameter searching module for agricultural machine running state monitoring;
The agricultural machinery running state monitoring model parameter searching module introduces random variable improvement tent mapping, further initializes individual positions, improves convergence factors based on Gaussian distribution change curves, calculates optimal individual positions based on a proportion weighting strategy, updates all individual positions by combining the improved convergence factors and the optimal individual positions to obtain model optimal parameters, and sends data to the real-time running state monitoring module;
The real-time running state monitoring module obtains the real-time running state of the agricultural machine based on the output label of the model, and realizes real-time monitoring of the agricultural machine.
It is noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made hereto without departing from the spirit and principles of the present invention.
The invention and its embodiments have been described above with no limitation, and the actual construction is not limited to the embodiments of the invention as shown in the drawings. In summary, if one of ordinary skill in the art is informed by this disclosure, a structural manner and an embodiment similar to the technical solution should not be creatively devised without departing from the gist of the present invention.
Claims (8)
1. The agricultural machinery operation state monitoring method based on the Internet of things is characterized by comprising the following steps of:
Step S1, data acquisition;
S2, preprocessing data;
S3, constructing an agricultural machinery running state monitoring model, designing an encoder to extract the characteristics of source domain and target domain data, designing a decoder to complete data reconstruction, calculating the difference of source domain and target domain distribution through the Hilbert space of a reproduction core, aligning the data distribution of the two domains through the minimum difference, introducing a smooth tag algorithm to optimize the encoding of a data class tag, and designing a total loss function based on reconstruction loss, difference loss and classification loss;
step S4, searching parameters of an agricultural machinery running state monitoring model, introducing a random variable to improve tent mapping, initializing individual positions, improving convergence factors based on Gaussian distribution change curves, calculating optimal individual positions based on a proportion weighting strategy, and updating all individual positions by combining the improved convergence factors and the optimal individual positions to obtain optimal parameters of the model;
step S5, monitoring the real-time running state;
step S4 includes a step S42 of improving the convergence factor based on the Gaussian distribution change curve, wherein the formula is as follows:
;
Where a (G) is a convergence factor at the time of the G-th iterative search, G is a search number index, G is a total number of searches, and η is a cutoff value.
2. The method for monitoring the operation state of the agricultural machinery based on the Internet of things according to claim 1 is characterized in that in the step S3, the construction of the operation state monitoring model of the agricultural machinery specifically comprises the following steps:
step S31, extracting features, designing an encoder for extracting features from input data, wherein the encoder consists of 5 one-dimensional convolution blocks, each convolution block comprises a convolution layer and a maximum pooling layer, the convolution blocks are activated by RELU functions, the size of a first convolution kernel is 1 multiplied by 64, the sizes of the other four convolution kernels are 1 multiplied by 3, the step size of the maximum pooling layer is 1 multiplied by 2, and the formula for extracting the features is as follows:
;
;
Wherein, Q y and Q m are respectively the source domain feature and the target domain feature extracted from the source domain training set and the target domain training set, enc is the network structure of the encoder, and X y and X m are respectively the source domain training set and the target domain training set;
Step S32, data reconstruction, designing a decoder for data reconstruction, wherein the decoder consists of 5 one-dimensional deconvolution blocks, each deconvolution block comprises a deconvolution layer and an up-sampling layer, the structure of the decoder is completely symmetrical with the encoder, the size of a first deconvolution core is 1 multiplied by 64, the sizes of the other four deconvolution cores are 1 multiplied by 3, and the step length of the up-sampling layer is 1 multiplied by 2, and the method comprises the following steps:
step S321, calculating reconstructed target domain data, wherein the formula is as follows:
;
In the formula, Is the reconstructed target domain data, dec is the network structure of the decoder;
step S322, calculating reconstruction loss of the target domain data, wherein the following formula is adopted:
;
Where L rec is the reconstruction loss of the target domain data, N m is the number of target domain data in the target domain training set, AndThe method comprises the steps of respectively reconstructing the ith target domain data before reconstruction and the ith target domain data after reconstruction, wherein i is a target domain data index;
Step S33, domain alignment, mapping the source domain data and the target domain data to a Hilbert space with a reproduction core, and aligning the distribution of the two domains by minimizing the difference between the source domain distribution and the target domain distribution;
Step S34, classifying, namely inputting the extracted target domain characteristics into a softmax function to calculate the probability that the target domain data belongs to each data class label, and selecting the data class label with the highest probability as an output label of the target domain data on a model, wherein the step comprises the following steps:
Step S341, introducing a smooth tag algorithm to optimize the one-hot coding of the real data category tag, wherein the following formula is adopted:
;
Where epsilon is the smoothing parameter, Is thatIs used for the identification of the real data tag of (c),Is thatThe one-hot encoding of the real data label of (2), C is the total number of the data class labels, and C is the index of the data class labels;
step S342, classifying the loss, wherein the formula is as follows:
;
where L cls is the classification loss, AndRespectively areBelonging to class c and in the target domain training setThe probability of class data class labels, W and b are weight and bias, respectively;
step S35, designing a total loss function of the model, and combining the reconstruction loss, the difference loss and the classification loss according to the weight to obtain the total loss function of the model, wherein the formula is as follows:
;
Where L is the total loss function, β 1、β2 and β 3 are the first, second and third weights, respectively, and L LMMD is the difference loss of the source domain distribution and the target domain distribution.
3. The method for monitoring the operation state of an agricultural machine based on the Internet of things according to claim 2, wherein in step S33, the domain alignment specifically comprises the following steps:
step S331, calculating the difference between the source domain distribution and the target domain distribution, wherein the formula is as follows:
;
;
;
In the formula, Is the difference between the source domain distribution and the target domain distribution, p （c） and q （c） are the source domain distribution and the target domain distribution, respectively, l is the feature layer index, C is the total number of data class labels, C is the data class label index, N y is the amount of source domain data in the source domain training set,AndThe weights of the ith and jth source domain data belonging to class c data class labels in the source domain training set,AndThe weights of the ith and jth target domain data belonging to class c data class labels in the target domain training set,AndThe number of source domain data and target domain data in the source domain training set and the target domain training set, respectively, which belong to the same class c as the ith source domain data and target domain data, k (·) is a kernel function of the hilbert space with a reproduction kernel,AndThe activation values on the ith and jth source domain data of the first layer feature layer,AndThe activation values on the ith and jth target domain data of the first layer of feature layers respectively;
Step S332, calculating the difference loss between the source domain distribution and the target domain distribution, wherein the following formula is adopted:
;
Where L LMMD is the difference penalty of the source domain distribution and the target domain distribution, and L max is the number of feature layers.
4. The method for monitoring the operation state of the agricultural machinery based on the Internet of things according to claim 1, wherein in step S4, the search of the operation state monitoring model parameters of the agricultural machinery specifically comprises the following steps:
Step S41, an initial individual position, a parameter search space is built for a first weight beta 1, a second weight beta 2, a third weight beta 3, a weight W and a bias b in an agricultural machine running state monitoring model, random variable improvement tent mapping is introduced, F individual positions are initialized in the parameter search space based on the improved tent mapping, the individual positions are used as representative of agricultural machine running state monitoring model parameters, a total loss function of the agricultural machine running state monitoring model trained based on the individual positions is used as an individual fitness value, and the initial individual position is represented by the following formula:
;
;
Wherein A w+1 and A w are respectively w+1th and w improved chaotic sequences, v is a control value, lb and ub are respectively the lower limit and the upper limit of a parameter search space, D w is the initial position of a w individual, and rand (·) is a random number generation function;
Step S42, improving convergence factors based on Gaussian distribution change curves;
Step S43, calculating an optimal individual position, updating an individual fitness value, sequencing according to the sequence from small fitness to large fitness, selecting the first three individual positions as a first optimal individual position, a second optimal individual position and a third optimal individual position, and designing a proportional weighting strategy based on the fitness and the first three optimal individual positions to calculate the optimal individual position, wherein the formula is as follows:
;
;
;
Wherein U o (g) and Z o (g) are the fitness ratio and the position ratio of the o-th optimal individual position at the g-th iterative search, U 1（g）、u2 (g) and U 3 (g) are the fitness ratio and the position ratio of the first optimal individual position at the g-th iterative search, the second optimal individual position and the third optimal individual position at the g-th iterative search, U3928 (g) and Z 2 (g) are the fitness ratio and the position ratio of the first optimal individual position at the g-th iterative search, U o (g) is the fitness of the o-th optimal individual position at the g-th iterative search, U3825 (g) is the fitness ratio and the position ratio of the first optimal individual position at the g-th iterative search, U 2 (g) and Z 1 (g) are the fitness ratio and the position ratio of the second optimal individual position at the g-th iterative search, U 2 (g) and the Z 2 (g) are the fitness ratio and the position of the o-th optimal individual position at the g-th iterative search, U o (g) is the fitness ratio and the o-th optimal individual position at the g-th iterative search;
Step S44, updating the individual position according to the following formula:
;
Wherein D w (g+1) and D w (g) are the positions of the w-th individual at the g+1st and g-th iterative searches, and S 1 and S 2 are the first and second random values, respectively;
Step S45, determining optimal parameters of the model, presetting an fitness threshold, updating fitness of an individual, taking the position of the individual with the minimum fitness as the optimal parameters of the model when the minimum fitness is lower than the fitness threshold, constructing an agricultural machinery running state monitoring model based on the optimal parameters of the model, otherwise, turning to step S41 if the total number of searches is reached, otherwise, turning to step S42, and continuing iterative search.
5. The method for monitoring the running state of the agricultural machinery based on the Internet of things according to claim 1 is characterized in that in the step S1, the Internet of things equipment is deployed on the agricultural machinery, historical running data and real-time running data of the agricultural machinery are collected based on the Internet of things equipment, the historical running data and the real-time running data comprise agricultural machinery equipment information, environment parameter data, running track data, sensor data and maintenance records, the historical running data further comprise the running state of the agricultural machinery, the running state of the agricultural machinery comprises normal and agricultural machinery fault types, the running state is used as a data type label, and the collected data are transmitted to a central processing center.
6. The method for monitoring the running state of the agricultural machinery based on the Internet of things according to claim 1, wherein in the step S2, the data preprocessing is to perform data cleaning, data conversion, data normalization and data set construction processing on collected data through a central processing center, and the data set construction is to respectively construct a source domain training set with a data type label and a target domain training set without the data type label based on historical running data and real-time running data after data cleaning, data conversion and data normalization.
7. The method for monitoring the running state of the agricultural machinery based on the Internet of things according to claim 1, wherein in step S5, the real-time running state monitoring is based on an output label of a model to obtain the real-time running state of the agricultural machinery, so that the real-time monitoring of the agricultural machinery is realized.
8. An agricultural machinery operation state monitoring system based on the Internet of things is used for realizing the agricultural machinery operation state monitoring method based on the Internet of things according to any one of claims 1-7, and is characterized by comprising a data acquisition module, a data preprocessing module, an agricultural machinery operation state monitoring model building module, an agricultural machinery operation state monitoring model parameter searching module and a real-time operation state monitoring module;
The data acquisition module acquires historical operation data and real-time operation data of the agricultural machinery based on the Internet of things equipment and sends the data to the data preprocessing module;
The data preprocessing module performs data cleaning, data conversion, data normalization and data set construction processing on the collected data through the central processing center, and sends the data to the agricultural machinery operation state monitoring model construction module;
The method comprises the steps that an encoder is designed by a model module for constructing an agricultural machine running state monitoring model to extract the characteristics of source domain data and target domain data, a decoder is designed to complete data reconstruction, the difference of source domain distribution and target domain distribution is calculated through a Hilbert space of a reproduction core, the data distribution of the two domains is aligned through minimized difference, a smooth tag algorithm is introduced to optimize the encoding of a data class tag, a total loss function is designed based on reconstruction loss, difference loss and classification loss, and data is sent to a model parameter searching module for agricultural machine running state monitoring;
The agricultural machinery running state monitoring model parameter searching module introduces random variable improvement tent mapping, further initializes individual positions, improves convergence factors based on Gaussian distribution change curves, calculates optimal individual positions based on a proportion weighting strategy, updates all individual positions by combining the improved convergence factors and the optimal individual positions to obtain model optimal parameters, and sends data to the real-time running state monitoring module;
The real-time running state monitoring module obtains the real-time running state of the agricultural machine based on the output label of the model, and realizes real-time monitoring of the agricultural machine.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411550164.1A CN119046801B (en) | 2024-11-01 | 2024-11-01 | Agricultural machinery operation state monitoring method and system based on Internet of things |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411550164.1A CN119046801B (en) | 2024-11-01 | 2024-11-01 | Agricultural machinery operation state monitoring method and system based on Internet of things |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN119046801A true CN119046801A (en) | 2024-11-29 |
| CN119046801B CN119046801B (en) | 2025-01-21 |
Family
ID=93582264
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202411550164.1A Active CN119046801B (en) | 2024-11-01 | 2024-11-01 | Agricultural machinery operation state monitoring method and system based on Internet of things |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN119046801B (en) |
Citations (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6671673B1 (en) * | 2000-03-24 | 2003-12-30 | International Business Machines Corporation | Method for integrated supply chain and financial management |
| US20040167721A1 (en) * | 2001-07-27 | 2004-08-26 | Masahiro Murakawa | Optimal fitting parameter determining method and device, and optimal fitting parameter determining program |
| WO2018072351A1 (en) * | 2016-10-20 | 2018-04-26 | 北京工业大学 | Method for optimizing support vector machine on basis of particle swarm optimization algorithm |
| CN114676634A (en) * | 2022-03-28 | 2022-06-28 | 广东广宇科技发展有限公司 | Water level prediction method and system based on NSGA-III optimized LSTM neural network |
| CN114693972A (en) * | 2022-03-29 | 2022-07-01 | 电子科技大学 | A Reconstruction-Based Domain Adaptation Method for Intermediate Domains |
| CN115115389A (en) * | 2022-03-11 | 2022-09-27 | 南京邮电大学 | A method for predicting express customer churn based on value segmentation and integrated forecasting |
| CN116975741A (en) * | 2023-09-13 | 2023-10-31 | 山东理工昊明新能源有限公司 | Internet of things-based energy equipment fault prediction method and device and electronic equipment |
| CN117584136A (en) * | 2024-01-18 | 2024-02-23 | 泰山学院 | Robot fault detection method and system based on artificial intelligence |
| CN117914701A (en) * | 2024-01-23 | 2024-04-19 | 青岛理工大学 | A blockchain-based building Internet of Things performance optimization system and method |
| CN117975138A (en) * | 2024-01-31 | 2024-05-03 | 南京邮电大学 | Pipeline defect detection method and system based on improved hawk optimization algorithm |
| CN118152920A (en) * | 2024-01-29 | 2024-06-07 | 东莞理工学院 | Bearing fault diagnosis method based on contrastive representation learning and subdomain adaptation |
| CN118285753A (en) * | 2024-04-20 | 2024-07-05 | 广东诺凯科技有限公司 | Internet intelligent remote blood pressure abnormal data monitoring system based on big data |
| CN118350538A (en) * | 2024-04-26 | 2024-07-16 | 广西电网有限责任公司 | Power spatiotemporal data dynamic monitoring method and system based on big data |
| CN118537627A (en) * | 2024-05-06 | 2024-08-23 | 太原科技大学 | Unbalanced multi-source domain self-adaptive model and method based on inter-sample dynamic similarity guidance |
| CN118608106A (en) * | 2024-08-08 | 2024-09-06 | 青岛九天智慧农业集团有限公司 | An intelligent agricultural supply chain supervision method |
| CN118798409A (en) * | 2024-05-15 | 2024-10-18 | 华北电力大学 | A method and system for optimizing regional cold and heat source energy supply network based on dual physical information neural network |
| CN118839228A (en) * | 2024-09-23 | 2024-10-25 | 山东联化新材料有限责任公司 | Industrial flow operation monitoring system based on data analysis |
-
2024
- 2024-11-01 CN CN202411550164.1A patent/CN119046801B/en active Active
Patent Citations (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6671673B1 (en) * | 2000-03-24 | 2003-12-30 | International Business Machines Corporation | Method for integrated supply chain and financial management |
| US20040167721A1 (en) * | 2001-07-27 | 2004-08-26 | Masahiro Murakawa | Optimal fitting parameter determining method and device, and optimal fitting parameter determining program |
| WO2018072351A1 (en) * | 2016-10-20 | 2018-04-26 | 北京工业大学 | Method for optimizing support vector machine on basis of particle swarm optimization algorithm |
| CN115115389A (en) * | 2022-03-11 | 2022-09-27 | 南京邮电大学 | A method for predicting express customer churn based on value segmentation and integrated forecasting |
| CN114676634A (en) * | 2022-03-28 | 2022-06-28 | 广东广宇科技发展有限公司 | Water level prediction method and system based on NSGA-III optimized LSTM neural network |
| CN114693972A (en) * | 2022-03-29 | 2022-07-01 | 电子科技大学 | A Reconstruction-Based Domain Adaptation Method for Intermediate Domains |
| CN116975741A (en) * | 2023-09-13 | 2023-10-31 | 山东理工昊明新能源有限公司 | Internet of things-based energy equipment fault prediction method and device and electronic equipment |
| CN117584136A (en) * | 2024-01-18 | 2024-02-23 | 泰山学院 | Robot fault detection method and system based on artificial intelligence |
| CN117914701A (en) * | 2024-01-23 | 2024-04-19 | 青岛理工大学 | A blockchain-based building Internet of Things performance optimization system and method |
| CN118152920A (en) * | 2024-01-29 | 2024-06-07 | 东莞理工学院 | Bearing fault diagnosis method based on contrastive representation learning and subdomain adaptation |
| CN117975138A (en) * | 2024-01-31 | 2024-05-03 | 南京邮电大学 | Pipeline defect detection method and system based on improved hawk optimization algorithm |
| CN118285753A (en) * | 2024-04-20 | 2024-07-05 | 广东诺凯科技有限公司 | Internet intelligent remote blood pressure abnormal data monitoring system based on big data |
| CN118350538A (en) * | 2024-04-26 | 2024-07-16 | 广西电网有限责任公司 | Power spatiotemporal data dynamic monitoring method and system based on big data |
| CN118537627A (en) * | 2024-05-06 | 2024-08-23 | 太原科技大学 | Unbalanced multi-source domain self-adaptive model and method based on inter-sample dynamic similarity guidance |
| CN118798409A (en) * | 2024-05-15 | 2024-10-18 | 华北电力大学 | A method and system for optimizing regional cold and heat source energy supply network based on dual physical information neural network |
| CN118608106A (en) * | 2024-08-08 | 2024-09-06 | 青岛九天智慧农业集团有限公司 | An intelligent agricultural supply chain supervision method |
| CN118839228A (en) * | 2024-09-23 | 2024-10-25 | 山东联化新材料有限责任公司 | Industrial flow operation monitoring system based on data analysis |
Non-Patent Citations (1)
| Title |
|---|
| 曹宁;江志农;高金吉;: "小样本下基于迁移学习的轴承状态识别方法", 噪声与振动控制, no. 05, 18 October 2020 (2020-10-18) * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN119046801B (en) | 2025-01-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110597240B (en) | A fault diagnosis method for hydro-generator units based on deep learning | |
| CN111967486A (en) | Complex equipment fault diagnosis method based on multi-sensor fusion | |
| CN118709789B (en) | Crop growth prediction method and system based on artificial intelligence and crop growth model | |
| CN109214503B (en) | Power transmission and transformation project cost prediction method based on KPCA-LA-RBM | |
| CN113468803A (en) | Improved WOA-GRU-based flood flow prediction method and system | |
| CN113705396B (en) | Motor fault diagnosis method, system and equipment | |
| Yin et al. | Runoff predictions in ungauged basins using sequence-to-sequence models | |
| CN115829162A (en) | Crop yield prediction method, device, electronic device and medium | |
| CN116050621B (en) | An ultra-short-term power prediction method for offshore wind power based on multi-head self-attention integrated boosting mode | |
| CN115575125A (en) | Bearing fault diagnosis method based on GADF-GAN-AVOA-CNN | |
| CN111783879B (en) | Hierarchical Compression Graph Matching Method and System Based on Orthogonal Attention Mechanism | |
| CN113361559A (en) | Multi-mode data knowledge information extraction method based on deep width joint neural network | |
| CN117574057B (en) | A method for intelligent fault diagnosis of vertical water pump units | |
| CN108470358A (en) | A kind of point cloud registration method for shaking artificial bee colony algorithm based on second order | |
| CN117081063A (en) | A distributed charging load prediction method and system based on GCN-Crossformer model | |
| CN118114985A (en) | Customer service risk assessment method and system based on bidirectional feedback mechanism | |
| CN118656698B (en) | A classification method and device for multiple historical and cultural resources | |
| CN116526478A (en) | Short-term wind power prediction method and system based on improved snake group optimization algorithm | |
| CN117458437A (en) | Short-term wind power prediction method, system, equipment and medium | |
| CN120124007A (en) | Prediction method for wind power of multi-node wind turbine cluster, computer program product and storage medium | |
| CN119046801B (en) | Agricultural machinery operation state monitoring method and system based on Internet of things | |
| CN114254828A (en) | Power load prediction method based on hybrid convolution feature extractor and GRU | |
| CN119398737A (en) | A method for health assessment and life prediction of sludge dewatering machine | |
| CN116405368B (en) | A network fault diagnosis method and system under the condition of high-dimensional unbalanced data | |
| CN113673152A (en) | Digital twin body-based group-level KKS coding intelligent mapping recommendation method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |