CN116880302A - Energy efficiency and carbon emission online monitoring and early warning method based on depth algorithm - Google Patents
Energy efficiency and carbon emission online monitoring and early warning method based on depth algorithm Download PDFInfo
- Publication number
- CN116880302A CN116880302A CN202310933596.XA CN202310933596A CN116880302A CN 116880302 A CN116880302 A CN 116880302A CN 202310933596 A CN202310933596 A CN 202310933596A CN 116880302 A CN116880302 A CN 116880302A
- Authority
- CN
- China
- Prior art keywords
- data
- energy efficiency
- carbon emission
- early warning
- layer
- 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.)
- Pending
Links
- 238000012544 monitoring process Methods 0.000 title claims abstract description 44
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 title claims abstract description 32
- 229910052799 carbon Inorganic materials 0.000 title claims abstract description 32
- 238000000034 method Methods 0.000 title claims abstract description 27
- 238000000605 extraction Methods 0.000 claims abstract description 14
- 238000012549 training Methods 0.000 claims description 28
- 230000006870 function Effects 0.000 claims description 14
- 238000007781 pre-processing Methods 0.000 claims description 14
- 238000012795 verification Methods 0.000 claims description 14
- 238000004519 manufacturing process Methods 0.000 claims description 11
- 238000012360 testing method Methods 0.000 claims description 11
- 238000006243 chemical reaction Methods 0.000 claims description 7
- 238000011176 pooling Methods 0.000 claims description 7
- 239000013598 vector Substances 0.000 claims description 7
- 238000010606 normalization Methods 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 5
- 238000011156 evaluation Methods 0.000 claims description 4
- 238000013507 mapping Methods 0.000 claims description 3
- 230000008859 change Effects 0.000 claims description 2
- 230000008901 benefit Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 230000008447 perception Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000008450 motivation Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/04—Programme control other than numerical control, i.e. in sequence controllers or logic controllers
- G05B19/042—Programme control other than numerical control, i.e. in sequence controllers or logic controllers using digital processors
- G05B19/0428—Safety, monitoring
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B2219/00—Program-control systems
- G05B2219/20—Pc systems
- G05B2219/24—Pc safety
- G05B2219/24024—Safety, surveillance
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention belongs to the technical field of online monitoring, and particularly relates to an energy efficiency and carbon emission online monitoring and early warning method based on a depth algorithm, which aims to solve the problems of incomplete energy efficiency and carbon emission monitoring, incomplete data statistics, low manual extraction efficiency and the like in the prior art.
Description
Technical Field
The invention belongs to the technical field of online monitoring, and particularly relates to an energy efficiency and carbon emission online monitoring and early warning method based on a depth algorithm.
Background
The supervision of energy efficiency and carbon emission is always focused in the production process of enterprises in China, the existing energy efficiency and carbon emission monitoring is not perfect, the relevance of enterprise and process data and main equipment is weak, the data statistics is incomplete, the energy efficiency, carbon emission and operation states of the enterprises cannot be comprehensively known, good planning cannot be made, meanwhile, most of the energy efficiency and carbon emission supervision work in China still is to manually report some energy efficiency and carbon emission data based on the monitored party energy consumption account, errors or omission are easy to occur when the data statistics is carried out in the mode, the data fairness and accuracy cannot be guaranteed, and the enterprise emission condition and the operation states of main equipment cannot be dynamically reflected in real time, so that the function expansion is difficult.
Deep learning is a new field in machine learning research, and the motivation is to build and simulate a neural network for analysis learning of human brain, which is to complete feature extraction by the machine itself without manual extraction, and is widely applied to various industries such as computer vision, natural language processing, fault diagnosis field and the like. The energy efficiency and carbon emission monitoring simulation is carried out according to deep learning by researching various energy efficiency indexes and carbon emission indexes such as resources, energy sources and environment generated in the enterprise production process, and finally the enterprise energy efficiency and carbon emission on-line monitoring and early warning operation is realized.
Disclosure of Invention
The invention aims to provide an energy efficiency and carbon emission online monitoring method based on a depth algorithm.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
an energy efficiency and carbon emission on-line monitoring and early warning method based on a depth algorithm comprises the following steps:
step 1, collecting primary, secondary and tertiary metering appliance data of enterprises and main production process operation data;
step 2, calculating energy efficiency indexes and carbon emission indexes of main production, auxiliary production and main equipment by using related standards and methods, and monitoring the running state of the main equipment;
step 3, carrying out data preprocessing on the energy efficiency index and the carbon emission index, and simultaneously inputting the index subjected to the data preprocessing and the corresponding running state into a database to form a data set;
step 4, dividing the data set into a training set, a verification set and a test set;
step 5, establishing an energy efficiency and carbon emission online monitoring model based on a depth algorithm, and training the model by using a training set;
step 6, evaluating the performance of the online monitoring model by using a verification set and a test set, wherein the evaluation index is an accuracy rate and a loss function;
step 7, inputting the data subjected to data preprocessing into an online monitoring model meeting the monitoring state, and obtaining a monitoring result by adopting a Softmax classifier;
and 8, early warning is carried out according to the obtained monitoring result.
Further, the operation states in the step 2 are defined as five types, including excellent, good, qualified and unqualified.
Further, the data preprocessing in the step 3 comprises data conversion and standardization processing;
the data conversion adopts a single-heat coding mode, discrete states of different data are effectively coded through a single-heat coder, the running states are represented by binary vectors, classification labels are made for each data, classification values are mapped to signal type indexes, the classification values are described by binary vectors, the mapped index positions are marked as 1, the rest positions are marked as zero values, the index marked by mapping is used as a real category for supervised learning, the real category is compared with a prediction category trained through a model, and the accuracy of model training is determined by the distance relation between the two categories;
the normalization process is to scale the features of each column to the [0,1] interval by using a Min-max function, and the Min-maxnormal is adopted to change the dimensionality expression into the dimensionless expression, wherein the formula is as follows:
wherein x is max Is the maximum value in the data set, x min As a minimum value for the data set,is a standardized result.
Further, in the step 4, the data set is divided into a training set, a verification set and a test set, which specifically includes the steps of:
the dataset was first written with 7: the ratio of 3 is divided into a training set and a testing set, and then the training set is divided into the training set and a verification set by adopting k-fold cross verification.
Further, the structure of the monitoring model in the step 5 includes an input layer, a feature extraction layer and a classification layer, wherein the feature extraction layer includes three convolution layers and three pooling layers, the first layer of the convolution layers is a convolution kernel of 7*1, the second layer is a convolution kernel of 5*1, the third layer is a convolution kernel of 3*1, the steps are one step, and the sizes and the steps of the filters of the three pooling layers are 2*1 and 2 steps; the classification layer comprises two full-connection layers, the first full-connection layer is used for flattening the output result of the feature extraction layer to form a one-dimensional sequence form and pass through the linear rectification unit layer, and the second full-connection layer is a Softmax classifier and used for predicting the output of the target class.
Further, the accuracy in the step 6 is Acc accuracy, which is defined as follows:
further, the loss function in the step 6 is a cross entropy function loss, which is defined as follows:
Loss＝-(ylogp+(1-y)log(1-p))
where y is the category and p is the probability of predicting the category
Compared with the prior art, the invention has the following advantages:
the invention completes preprocessing of the original data through data conversion and standardization processing, and the aim of the preprocessing is to strengthen the accuracy of model training, then utilizes feature extraction to extract the features of the input data, then monitors the state of the input data through a Softmax classifier by utilizing accuracy and loss function evaluation indexes, and makes corresponding early warning through the output state.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1
An energy efficiency and carbon emission on-line monitoring and early warning method based on a depth algorithm comprises the following steps:
step 1, collecting primary, secondary and tertiary metering appliance data of enterprises and main production process operation data;
step 2, calculating energy efficiency indexes and carbon emission indexes of main production, auxiliary production and main equipment by using related standards and methods, and monitoring the running state of the main equipment, wherein the running state is defined as five types including excellent, good, qualified and unqualified;
step 3, carrying out data preprocessing on the energy efficiency index and the carbon emission index, and simultaneously inputting the index subjected to the data preprocessing and the corresponding running state into a database to form a data set;
wherein the data preprocessing comprises data conversion and normalization processing;
the data conversion adopts a single-Hot coding mode, and the coding principle of single-Hot coding (One-Hot) is that an N-bit state register is adopted as N-bit state coding, the One-Hot coding can be understood as that each state has unique register bits, and under normal state, a single bit is kept valid, the effective value is 1, and the rest value is 0. The data features are in discrete states before being encoded, are in continuous states after being encoded, and normalization operation is carried out on continuous feature dimensions after being encoded.
Because different data can be regarded as discrete data characteristics, the discrete states of the different data are effectively encoded through the single thermal encoder, the running states are represented by binary vectors, classification labels are made for each data, classification values are mapped to signal type indexes, the classification values are described by the binary vectors, the mapped index positions are marked as 1, the rest positions are marked as zero values, the index marked by the mapping is used as a real category for supervised learning, the real category is compared with a prediction category trained through a model, and the accuracy of model training is determined by the distance relation between the two. The non-numeric data is converted to a finite 0,1 numeric sequence as shown in the table below.
Because the units of different data items of the original data are different, when the data value difference is too large, the original data is subjected to standardized processing, and the method aims to conveniently process the data, and the characteristics of each column are scaled to the [0,1] interval by using a min-max function, so that the data is more convenient and faster to process; in order to eliminate the dimension and accelerate convergence, the dimension expression is changed into a dimensionless expression, so that indexes with different units or magnitudes can be compared and weighted conveniently, and the comparability between the data indexes is solved. The method adopted is Min-max normalization, and the formula is as follows:
wherein x is max Is the maximum value in the data set, x min As a minimum value for the data set,is a standardized result.
Step 4, dividing the data set into a training set, a verification set and a test set, wherein the data set comprises the following specific steps: the dataset was first written with 7:3, dividing the training set into a training set and a testing set, and then dividing the training set into the training set and a verification set by adopting k-fold cross verification, wherein k is 5;
step 5, an energy efficiency and carbon emission online monitoring model based on a depth algorithm is established, the model is trained by using a training set, the structure of the monitoring model comprises an input layer, a feature extraction layer and a classification layer, the feature extraction layer comprises three convolution layers and three pooling layers, wherein the first layer of the convolution layers is a convolution core of 7*1, the second layer is a convolution core of 5*1, the third layer is a convolution core of 3*1, the steps are one step, and the sizes and the steps of the filters of the three pooling layers are 2*1 and 2 steps; the classification layer comprises two full-connection layers, the first full-connection layer is used for flattening the output result of the feature extraction layer to form a one-dimensional sequence form and pass through the linear rectification unit layer, and the second full-connection layer is a Softmax classifier and used for predicting the output of the target class.
The basic parameters were set as follows:
weight sharing, pooled downsampling and regional local perception are adopted in the general training process, so that training parameters are greatly reduced, and the method is also an advantage of a convolutional neural network. Local area perception is actually a sparse connection idea, the latter layer and the former layer are locally connected to obtain local characteristics, and the local information is integrated at the highest layer by layer differentiation to obtain complete characteristic information. Weight sharing is adopted in convolution operation, so that the number of neurons is reduced, training parameters are greatly reduced, the scale of the training parameters is greatly reduced, and meanwhile, the learning rate is also greatly improved. The convolution operation is implemented by a convolution kernel whose parameters include an offset vector and a weight matrix, which are shared in the convolution neural network, while the location of the local feature need not be considered during the extraction process. Downsampling is in essence a pooling operation in convolutional neural networks. In order to further reduce the computational complexity after the convolution operation, the convolved feature data is further scaled and mapped to reduce the dimension of the feature data on the premise of ensuring that the feature is not distorted.
And 6, evaluating the performance of the online monitoring model by using a verification set and a test set, wherein the evaluation indexes are accuracy and a loss function, the accuracy is Acc accuracy, and the method is defined as follows:
the loss function is a cross entropy function loss, which is defined as follows:
Loss＝-(ylogp+(1-y)log(1-p))
where y is the class and p is the probability of predicting the class.
Step 7, inputting the data subjected to data preprocessing into an online monitoring model meeting the monitoring state, and obtaining a monitoring result by adopting a Softmax classifier;
and 8, early warning is carried out according to the obtained monitoring result.
Claims (7)
1. The energy efficiency and carbon emission online monitoring and early warning method based on the depth algorithm is characterized by comprising the following steps of:
step 1, collecting primary, secondary and tertiary metering appliance data of enterprises and main production process operation data;
step 2, calculating energy efficiency indexes and carbon emission indexes of main production, auxiliary production and main equipment by using related standards and methods, and monitoring the running state of the main equipment;
step 3, carrying out data preprocessing on the energy efficiency index and the carbon emission index, and simultaneously inputting the index subjected to the data preprocessing and the corresponding running state into a database to form a data set;
step 4, dividing the data set into a training set, a verification set and a test set;
step 5, establishing an energy efficiency and carbon emission online monitoring model based on a depth algorithm, and training the model by using a training set;
step 6, evaluating the performance of the online monitoring model by using a verification set and a test set, wherein the evaluation index is an accuracy rate and a loss function;
step 7, inputting the data subjected to data preprocessing into an online monitoring model meeting the monitoring state, and obtaining a monitoring result by adopting a Softmax classifier;
and 8, early warning is carried out according to the obtained monitoring result.
2. The method for on-line monitoring and early warning of energy efficiency and carbon emission based on the depth algorithm according to claim 1, wherein the operation states in the step 2 are defined as five types including excellent, good, qualified and unqualified.
3. The method for online monitoring and early warning of energy efficiency and carbon emission based on a depth algorithm according to claim 1, wherein the data preprocessing in the step 3 comprises data conversion and standardization processing;
the data conversion adopts a single-heat coding mode, discrete states of different data are effectively coded through a single-heat coder, the running states are represented by binary vectors, classification labels are made for each data, classification values are mapped to signal type indexes, the classification values are described by binary vectors, the mapped index positions are marked as 1, the rest positions are marked as zero values, the index marked by mapping is used as a real category for supervised learning, the real category is compared with a prediction category trained through a model, and the accuracy of model training is determined by the distance relation between the two categories;
the normalization process is to scale the features of each column to the [0,1] interval by using a Min-max function, and the Min-max normalization is adopted to change the dimensionality expression into a dimensionless expression, wherein the formula is as follows:
wherein x is max Is the maximum value in the data set, x min As a minimum value for the data set,is a standardized result.
4. The method for online monitoring and early warning of energy efficiency and carbon emission based on the depth algorithm according to claim 1, wherein in the step 4, the data set is divided into a training set, a verification set and a test set, and the specific steps are as follows:
the dataset was first written with 7: the ratio of 3 is divided into a training set and a testing set, and then the training set is divided into the training set and a verification set by adopting k-fold cross verification.
5. The method for monitoring and early warning energy efficiency and carbon emission on line based on a depth algorithm according to claim 1, wherein the structure of the monitoring model in the step 5 comprises an input layer, a feature extraction layer and a classification layer, the feature extraction layer comprises three convolution layers and three pooling layers, wherein the first layer of the convolution layers is a convolution kernel of 7*1, the second layer is a convolution kernel of 5*1, the third layer is a convolution kernel of 3*1, the steps are one step, and the filter sizes and the steps of the three pooling layers are 2*1 and 2 steps; the classification layer comprises two full-connection layers, the first full-connection layer is used for flattening the output result of the feature extraction layer to form a one-dimensional sequence form and pass through the linear rectification unit layer, and the second full-connection layer is a Softmax classifier and used for predicting the output of the target class.
6. The method for online monitoring and early warning of energy efficiency and carbon emission based on depth algorithm as claimed in claim 1, wherein the accuracy in the step 6 is Acc accuracy, which is defined as follows:
7. the method for online monitoring and early warning of energy efficiency and carbon emission based on depth algorithm as claimed in claim 1, wherein the loss function in the step 6 is a cross entropy function loss, which is defined as follows:
Loss＝-(ylogp+(1-y)log(1-p))
where y is the class and p is the probability of predicting the class.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310933596.XA CN116880302A (en) | 2023-07-27 | 2023-07-27 | Energy efficiency and carbon emission online monitoring and early warning method based on depth algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310933596.XA CN116880302A (en) | 2023-07-27 | 2023-07-27 | Energy efficiency and carbon emission online monitoring and early warning method based on depth algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116880302A true CN116880302A (en) | 2023-10-13 |
Family
ID=88269760
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310933596.XA Pending CN116880302A (en) | 2023-07-27 | 2023-07-27 | Energy efficiency and carbon emission online monitoring and early warning method based on depth algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116880302A (en) |
-
2023
- 2023-07-27 CN CN202310933596.XA patent/CN116880302A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108320043B (en) | Power distribution network equipment state diagnosis and prediction method based on electric power big data | |
| CN112101480B (en) | Multivariate clustering and fused time sequence combined prediction method | |
| CN111830408A (en) | Motor fault diagnosis system and method based on edge calculation and deep learning | |
| CN110636066B (en) | Network security threat situation assessment method based on unsupervised generative reasoning | |
| CN112668105B (en) | Helicopter transmission shaft abnormity judgment method based on SAE and Mahalanobis distance | |
| CN116300691A (en) | State monitoring method and system for multi-axis linkage numerical control machining | |
| CN113568928B (en) | Data management system applied to energy industry chain early warning system | |
| CN110728320A (en) | Water quality monitoring and early warning method and system based on combination of self-coding and clustering | |
| CN111639850A (en) | Quality evaluation method and system for multi-source heterogeneous data | |
| CN110910026A (en) | Intelligent management and decision-making method and system for route loss of trans-provincial power transmission line | |
| CN115795351A (en) | Elevator big data risk early warning method based on residual error network and 2D feature representation | |
| CN115375026A (en) | Method for predicting service life of aircraft engine in multiple fault modes | |
| CN114386452A (en) | Method for detecting faults of sun wheel of nuclear power circulating water pump | |
| CN117435992A (en) | Fault prediction method and system for hydraulic propulsion system of shield tunneling machine | |
| CN115564075B (en) | Main and auxiliary integrated fault collaborative diagnosis method and system for urban power grid | |
| WO2024045377A1 (en) | Deep-learning-based method for predicting residual life of aero-engine | |
| CN116880302A (en) | Energy efficiency and carbon emission online monitoring and early warning method based on depth algorithm | |
| Hu et al. | Research on the fault identification method of oil pumping unit based on residual network | |
| CN114841063A (en) | Aero-engine residual life prediction method based on deep learning | |
| CN114936703A (en) | Marketing company financial violation prediction method based on improved Transformer model | |
| CN114254828A (en) | Power load prediction method based on hybrid convolution feature extractor and GRU | |
| CN113705695A (en) | Power distribution network fault data identification method based on convolutional neural network | |
| Wang et al. | Research on Identification of Fraud in Financial Reporting Based on Adaptive Group Lasso-Logistic Regression | |
| CN115556099B (en) | Sustainable learning industrial robot fault diagnosis system and method | |
| CN116579632B (en) | Method for monitoring service state of wind power equipment based on deep learning technology |
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 |