A carbon emission intelligent meter

Abstract

The present invention relates to the field of intelligent data control, and provides a carbon emission intelligent meter, which generates a carbon measurement demand object attribute set; takes capability attribute data of multiple existing medical carbon measurement sources as a training data set, trains the historical object attribute set, and then applies the carbon measurement attribute index of the historical object attribute set to the carbon measurement object attribute set to be measured; during the waiting process, inputs the capability attribute data of multiple medical carbon measurement sources received in real time into the trained carbon measurement object attribute set to be measured, and the output result is the query result of the medical carbon measurement source; transmits the control authority of the selected medical carbon measurement source to the carbon measurement demand object; optimizes the model index and weight of the carbon measurement demand object attribute set of the carbon measurement demand object receiving the medical carbon measurement source; and reasonably allocates carbon measurement demand resources and optimally matches carbon measurement demand and carbon measurement provider.

Description

Technical Field

[0001] The present invention relates to the field of intelligent data control, and in particular to an intelligent carbon emission meter. Background Art

[0002] According to the working principle, household carbon emission metering machines are mainly classified into electronic carbon emission metering machines, molecular sieve carbon emission metering machines, chemical agent carbon emission metering machines and oxygen-enriched membrane carbon emission metering machines, etc. Molecular sieve carbon emission metering machines are currently the only mature carbon emission metering machines with international and national standards.

[0003] In particular, since the carbon emission measurement levels and capabilities are different in different regions, the urgency of different carbon emission measurement needs is also different. How to better allocate resources and optimally match carbon emission measurement needs is an urgent problem to be solved. Summary of the invention

[0004] To achieve the above objectives, this application provides the following technical solutions:

[0005] According to a first aspect of the present invention, the present invention claims protection for a carbon emission smart meter, comprising:

[0006] A carbon metering attribute generation module is used to generate a carbon metering demand object attribute set, wherein the carbon metering demand object attribute set includes a historical object attribute set and a to-be-metered object attribute set;

[0007] A carbon metering attribute training module is used to take the capability attribute data of multiple existing medical carbon metering sources as a training data set to train the historical object attribute set, and after the training is completed, apply the carbon metering attribute index of the historical object attribute set to the to-be-metered object attribute set;

[0008] The medical carbon metering source query module is used to input the capability attribute data of multiple medical carbon metering sources received in real time into the attribute set of the carbon metering object to be trained during the waiting process of the carbon metering demand object, and the output result is the query result of the medical carbon metering source;

[0009] A transmission module, used for transmitting the selected medical carbon metering source control authority to the carbon metering demand object through an encrypted communication line;

[0010] The optimization module is used to optimize the model index and weight of the carbon metering demand object attribute set of the carbon metering demand object receiving the medical carbon metering source.

[0011] Furthermore, the historical object attribute set and the to-be-measured object attribute set of the carbon measurement demand object attribute set adopt DOMP neural carbon measurement attributes; wherein the step of training the historical object attribute set includes:

[0012] Collecting medical carbon metering sources and extracting capability attribute data through the collector of the historical object attribute set, dividing the data into a training sample set and a test sample set as training data of the historical object attribute set;

[0013] Generate the historical object attribute set; the historical object attribute set includes two carbon measurement path samples OMP, and synchronously train the two carbon measurement path samples OMP;

[0014] Input the training sample set into two carbon metering path sample OMPs for training, optimize the index of the first carbon metering path sample OMP, find the action with the maximum Q value in the first carbon metering path sample OMP, and take the second carbon metering path sample OMP to calculate the Q value that meets the requirements;

[0015] Generate a residual function of the historical object attribute set, which represents the difference between the Q value of the current carbon measurement demand and the target Q value. When the residual function is minimized, decide that the training of the historical object attribute set is completed, and input the test sample set into the trained historical object attribute set to calibrate the accuracy of the training.

[0016] Furthermore, after the training of the historical object attribute set is completed, the carbon metering attribute index of the historical object attribute set is applied to the object attribute set to be carbon metered; after the object attribute set to be carbon metered queries the medical carbon metering source, it generates experience information and sends it to the historical object attribute set for training iteration of the historical object attribute set.

[0017] Furthermore, the step of receiving capability attribute data of multiple medical carbon metering sources in real time includes:

[0018] The carbon metering demand object sends a detection interest packet to find potential medical carbon metering sources and capability attribute data;

[0019] After receiving the medical carbon measurement source, it is filtered through anomaly denoising to obtain a medical carbon measurement source set as the input of the attribute set of the object to be carbon measured.

[0020] Furthermore, before performing the query of the medical carbon metering source, a step of triggering the query of the medical carbon metering source is included; the reasons for triggering the query of the medical carbon metering source are determined to include at least: the carbon metering demand object queries the medical carbon metering source for the carbon metering demand object when initializing; or the connected carbon metering transmission channel deteriorates in quality or is disconnected, and a new medical carbon metering source is queried for the carbon metering demand object;

[0021] When deciding whether the quality of the connected carbon metering transmission channel deteriorates or the connection is interrupted, the interruption probability of the current carbon metering demand object connecting to the medical carbon metering source is calculated, and the current carbon metering demand priority is used as the random seed, and the hit replacement probability is calculated using a random function, which is expressed as:

[0022] (1);

[0023] Generate a random number below 200 using the rand() function. The formula is:

[0024] (2);

[0025] Among them, H represents the medical carbon measurement source replacement certification mark, Timestamp represents the current timestamp, Indicates that the average interruption probability is less than or equal to the current interruption probability if It means the probability of executing the replacement of medical carbon metering source, p is the current interruption probability; P max It represents the maximum interruption probability, that is, when the interruption probability of the carbon metering transmission channel is greater than the maximum interruption probability, the probability calculation method will no longer be used for replacement identification, and the medical carbon metering source will be replaced directly. Mod100 means taking the remainder after dividing by 100.

[0026] The present invention relates to the field of intelligent data control, and seeks protection for a carbon emission intelligent meter, which generates a carbon metering demand object attribute set; takes capability attribute data of multiple existing medical carbon metering sources as a training data set, trains the historical object attribute set, and then applies the carbon metering attribute index of the historical object attribute set to the carbon metering object attribute set to be waited for; during the waiting process, inputs the capability attribute data of multiple medical carbon metering sources received in real time into the trained carbon metering object attribute set to be waited for, and the output result is the query result of the medical carbon metering source; transmits the control authority of the selected medical carbon metering source to the carbon metering demand object; optimizes the model index and weight of the carbon metering demand object attribute set of the carbon metering demand object receiving the medical carbon metering source; the present invention can better reasonably allocate carbon metering demand resources and optimally match carbon metering demand and carbon metering provider. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 A working flow chart of a carbon emission intelligent meter involved in the present invention;

[0028] Figure 2 A second working flow chart of a carbon emission intelligent meter involved in the present invention;

[0029] Figure 3 This is a structural module diagram of a carbon emission intelligent meter involved in the present invention. DETAILED DESCRIPTION

[0030] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0031] According to the first embodiment of the present invention, referring to Figure 1 The present invention claims a working method of a carbon emission intelligent meter, comprising:

[0032] Generate a carbon metering demand object attribute set, wherein the carbon metering demand object attribute set includes a historical object attribute set and a to-be-metered object attribute set;

[0033] Taking the capability attribute data of multiple existing medical carbon metering sources as a training data set, training the historical object attribute set, and applying the carbon metering attribute index of the historical object attribute set to the carbon metering object attribute set after the training is completed;

[0034] During the waiting process of the carbon metering demand object, the capability attribute data of multiple medical carbon metering sources received in real time are input into the attribute set of the carbon metering object to be trained, and the output result is the query result of the medical carbon metering source;

[0035] The selected medical carbon metering source control authority is transmitted to the carbon metering demand object through an encrypted communication line;

[0036] The model index and weight of the carbon metering demand object attribute set of the carbon metering demand object receiving the medical carbon metering source are optimized.

[0037] Further, see Figure 2 The historical object attribute set of the carbon measurement demand object attribute set and the to-be-measured object attribute set adopt DOMP neural carbon measurement attributes; wherein the step of training the historical object attribute set includes:

[0038] Collecting medical carbon metering sources and extracting capability attribute data through the collector of the historical object attribute set, dividing the data into a training sample set and a test sample set as training data of the historical object attribute set;

[0039] Generate the historical object attribute set; the historical object attribute set includes two carbon measurement path samples OMP, and synchronously train the two carbon measurement path samples OMP;

[0040] Input the training sample set into two carbon metering path sample OMPs for training, optimize the index of the first carbon metering path sample OMP, find the action with the maximum Q value in the first carbon metering path sample OMP, and take the second carbon metering path sample OMP to calculate the Q value that meets the requirements;

[0041] Generate a residual function of the historical object attribute set, which represents the difference between the Q value of the current carbon measurement demand and the target Q value. When the residual function is minimized, decide that the training of the historical object attribute set is completed, and input the test sample set into the trained historical object attribute set to calibrate the accuracy of the training.

[0042] Furthermore, after the training of the historical object attribute set is completed, the carbon metering attribute index of the historical object attribute set is applied to the object attribute set to be carbon metered; after the object attribute set to be carbon metered queries the medical carbon metering source, it generates experience information and sends it to the historical object attribute set for training iteration of the historical object attribute set.

[0043] Furthermore, the step of receiving capability attribute data of multiple medical carbon metering sources in real time includes:

[0044] The carbon metering demand object sends a detection interest packet to find potential medical carbon metering sources and capability attribute data;

[0045] After receiving the medical carbon measurement source, it is filtered through anomaly denoising to obtain a medical carbon measurement source set as the input of the attribute set of the object to be carbon measured.

[0046] Furthermore, before the query of the medical carbon metering source is performed, a step of triggering the query of the medical carbon metering source is included; the reasons for triggering the query of the medical carbon metering source are determined to include at least: the carbon metering demand object queries the medical carbon metering source for the carbon metering demand object when initializing; or the connected carbon metering transmission channel deteriorates in quality or is disconnected, and a new medical carbon metering source is queried for the carbon metering demand object;

[0047] When deciding whether the quality of the connected carbon metering transmission channel deteriorates or the connection is interrupted, the interruption probability of the current carbon metering demand object connecting to the medical carbon metering source is calculated, and the current carbon metering demand priority is used as the random seed, and the hit replacement probability is calculated using a random function, which is expressed as:

[0048] (1);

[0049] Generate a random number below 200 using the rand() function. The formula is:

[0050] (2);

[0051] Among them, H represents the medical carbon measurement source replacement certification mark, Timestamp represents the current timestamp, Indicates that the average interruption probability is less than or equal to the current interruption probability if It means the probability of executing the replacement of medical carbon metering source, p is the current interruption probability; P max It represents the maximum interruption probability, that is, when the interruption probability of the carbon metering transmission channel is greater than the maximum interruption probability, the probability calculation method will no longer be used for replacement identification, and the medical carbon metering source will be replaced directly. Mod100 means taking the remainder after dividing by 100.

[0052] Further, after receiving the medical carbon metering source, the step of filtering by abnormal denoising includes:

[0053] The acquisition module receives the capability attribute data s of the i-th medical carbon metering source at time t t,i =(N t,i ,M t,i ,V t,i ,D t,i ), where N t,i M represents the round-trip time between the medical carbon metering source and the carbon metering demand object of the i-th medical carbon metering source at time t, and the molecular sieve life; t,i represents the time offset of the data packet sent by the i-th medical carbon metering source at time t; V t,i represents the waiting speed of the i-th data at time t; D t,i Indicates the distance between the i-th medical carbon metering source and the object that sends the carbon metering request at time t;

[0054] The smaller the molecular sieve life, the smaller the round-trip delay of the medical carbon metering source, and the better the carbon metering property performance; when the molecular sieve life value is obtained from the same medical carbon metering source multiple times, it is more reliable to smooth the molecular sieve life. The smoothing method can be used, and the calculation formula is as follows:

[0055] N t,i =u×N t-1,i +e×(R t,i -N t-1,i ) (3);

[0056] Among them, R t,i Represents the currently observed instantaneous molecular sieve life value, N t-1,i The molecular sieve life represents the round-trip time between the medical carbon metering source and the carbon metering demand object of the i-th medical carbon metering source at the first moment, u=1, e=0.125;

[0057] Under the same cumulative running time, the larger the offset, the more idle the medical carbon metering source is, and the larger the remaining available cumulative running time is; on the contrary, the smaller the offset, the smaller the remaining available cumulative running time of the medical carbon metering source is. The calculation formula is:

[0058] M t,i =(1-σ)×M t-1,i +σ×(Data t,i -Data t-1,i ) (4);

[0059] Where, σ=0.5, Data t,i Indicates the time when the current capability attribute data is sent, M t-1,i represents the time offset of the i-th medical carbon metering source sending the data packet at the first moment, Data t-1,i Indicates the time when the last capability attribute data was sent. The difference between the two is the time offset.

[0060] V t,i >0 means the corresponding medical carbon metering source and the carbon metering demand object are waiting in the same direction, V t,i <0 means the corresponding medical carbon metering source and the object requesting carbon metering demand are waiting towards each other; the lower the speed, the more stable the medical carbon metering source;

[0061] D t,i represents the distance between the i-th medical carbon measurement source and the carbon measurement demand object sending the request at time t, D t,i >0 means the corresponding medical carbon measurement source is in front of the carbon measurement demand object, D t,i <0 means that the corresponding medical carbon measurement source is behind the object requiring carbon measurement; the closer the distance, the higher the stability of the medical carbon measurement source;

[0062] Use the multiple capability attribute data contained as the basis for sorting in turn to filter the medical carbon metering sources with status; 4 statuses can select one medical carbon metering source respectively, and the 4 medical carbon metering sources contain 4 groups of 16 capability attribute data in total, which are input as the attribute set of the carbon metering object to be measured.

[0063] Furthermore, in the step of sequentially using the multiple capability attribute data as a basis for sorting and filtering the medical carbon metering sources with status, it includes:

[0064] Calculate the optimal value for each state;

[0065] Calculate the execution time of action a t Query the score of medical carbon measurement source;

[0066] By adjusting the carbon metering attribute index of the attribute set of the object to be metered, the state index is normalized, the value range of the action index is adjusted, the mapping relationship between the medical carbon metering source and the state index value is generated, and the medical carbon metering source with the highest score is determined as the final query result;

[0067] The method also includes the step of setting a rich carbon metering value corresponding to a medical carbon metering source.

[0068] Furthermore, in the step of optimizing the model index and weight of the carbon metering demand object attribute set of the carbon metering demand object receiving the medical carbon metering source, the steps include:

[0069] The carbon measurement demand object attribute set of the carbon measurement demand object learns environmental samples and optimizes the carbon measurement attribute model and model weight by interacting with the real environment;

[0070] Optimize the carbon measurement demand object attribute set model;

[0071] In the steps of interacting with the real environment, learning environmental samples and optimizing the carbon measurement attribute model and model weights,

[0072] Optimize the carbon measurement attribute model and model weights, use local data and environmental status to train the carbon measurement demand object attribute set model, and transmit the weight results to a carbon emission storage end after multiple iterations. The carbon emission storage end will collect the weight results of all carbon measurement demand objects and calculate the gradient average.

[0073] According to the second embodiment of the present invention, referring to Figure 3 The present invention claims protection for a carbon emission intelligent meter, comprising:

[0074] A carbon metering attribute generation module is used to generate a carbon metering demand object attribute set, wherein the carbon metering demand object attribute set includes a historical object attribute set and a to-be-metered object attribute set;

[0075] A carbon metering attribute training module is used to take the capability attribute data of multiple existing medical carbon metering sources as a training data set to train the historical object attribute set, and after the training is completed, apply the carbon metering attribute index of the historical object attribute set to the to-be-metered object attribute set;

[0076] The medical carbon metering source query module is used to input the capability attribute data of multiple medical carbon metering sources received in real time into the attribute set of the carbon metering object to be trained during the waiting process of the carbon metering demand object, and the output result is the query result of the medical carbon metering source;

[0077] A transmission module, used for transmitting the selected medical carbon metering source control authority to the carbon metering demand object through an encrypted communication line;

[0078] The optimization module is used to optimize the model index and weight of the carbon metering demand object attribute set of the carbon metering demand object receiving the medical carbon metering source.

[0079] Furthermore, the historical object attribute set and the to-be-measured object attribute set of the carbon measurement demand object attribute set adopt DOMP neural carbon measurement attributes; wherein the step of training the historical object attribute set includes:

[0080] Collecting medical carbon metering sources and extracting capability attribute data through the collector of the historical object attribute set, dividing the data into a training sample set and a test sample set as training data of the historical object attribute set;

[0081] Generate the historical object attribute set; the historical object attribute set includes two carbon measurement path samples OMP, and synchronously train the two carbon measurement path samples OMP;

[0082] Input the training sample set into two carbon metering path sample OMPs for training, optimize the index of the first carbon metering path sample OMP, find the action with the maximum Q value in the first carbon metering path sample OMP, and take the second carbon metering path sample OMP to calculate the Q value that meets the requirements;

[0083] Generate a residual function of the historical object attribute set, which represents the difference between the Q value of the current carbon measurement demand and the target Q value. When the residual function is minimized, decide that the training of the historical object attribute set is completed, and input the test sample set into the trained historical object attribute set to calibrate the accuracy of the training.

[0084] Furthermore, after the training of the historical object attribute set is completed, the carbon metering attribute index of the historical object attribute set is applied to the object attribute set to be carbon metered; after the object attribute set to be carbon metered queries the medical carbon metering source, it generates experience information and sends it to the historical object attribute set for training iteration of the historical object attribute set.

[0085] Furthermore, the step of receiving capability attribute data of multiple medical carbon metering sources in real time includes:

[0086] The carbon metering demand object sends a detection interest packet to find potential medical carbon metering sources and capability attribute data;

[0087] After receiving the medical carbon measurement source, it is filtered through anomaly denoising to obtain a medical carbon measurement source set as the input of the attribute set of the object to be carbon measured.

[0088] Furthermore, before performing the query of the medical carbon metering source, a step of triggering the query of the medical carbon metering source is included; the reasons for triggering the query of the medical carbon metering source are determined to include at least: the carbon metering demand object queries the medical carbon metering source for the carbon metering demand object when initializing; or the connected carbon metering transmission channel deteriorates in quality or is disconnected, and a new medical carbon metering source is queried for the carbon metering demand object;

[0089] When deciding whether the quality of the connected carbon metering transmission channel deteriorates or the connection is interrupted, the interruption probability of the current carbon metering demand object connecting to the medical carbon metering source is calculated, and the current carbon metering demand priority is used as the random seed, and the hit replacement probability is calculated using a random function, which is expressed as:

[0090] ;

[0091] Generate a random number below 200 through the rand() function. Mod100 means taking the remainder after dividing by 100. The formula is:

[0092] ;

[0093] Among them, H represents the medical carbon measurement source replacement certification mark, Timestamp represents the current timestamp, Indicates that the average interruption probability is less than or equal to the current interruption probability if It means the probability of executing the replacement of medical carbon metering source, p is the current interruption probability; P max It represents the maximum interruption probability, that is, when the interruption probability of the carbon metering transmission channel is greater than the maximum interruption probability, the probability calculation method will no longer be used for replacement identification, and the medical carbon metering source will be replaced directly.

[0094] The specific implementation methods of the invention are described in detail above, but they are only examples, and the present application is not limited to the specific implementation methods described above. For those skilled in the art, any equivalent modification or substitution of the invention is also within the scope of the present application, and therefore, the equalization, modification, and improvement made without departing from the spirit and principle of the present application should be included in the scope of the present application.

Claims

1. A carbon emission intelligent meter, characterized in that: include: Carbon metering attribute generation module: used to generate a carbon metering demand object attribute set, which includes a historical object attribute set and a to-be-metered object attribute set; The carbon metering attribute training module is used to take the capability attribute data of multiple existing medical carbon metering sources as the training data set to train the historical object attribute set. After the training is completed, the carbon metering attribute indicators of the historical object attribute set are applied to the carbon metering object attribute set; The medical carbon metering source query module is used to input the capability attribute data of multiple medical carbon metering sources received in real time into the attribute set of the trained carbon metering object during the waiting process of the carbon metering demand object, and the output result is the query result of the medical carbon metering source; A transmission module, used for transmitting the selected medical carbon metering source control authority to the carbon metering demand object through an encrypted communication line; An optimization module, used for optimizing the model index and weight of the carbon metering demand object attribute set of the carbon metering demand object receiving the medical carbon metering source; The steps for receiving capability attribute data from multiple healthcare carbon metering sources in real time include: After receiving the medical carbon measurement source, it is filtered through abnormal denoising to obtain a medical carbon measurement source set as the input of the attribute set of the carbon measurement object to be measured; including: The acquisition module receives the capability attribute data s of the i-th medical carbon metering source at time t t,i =(N t,i ,M t,i ,V t,i ,D t,i ), where N t,i M represents the round-trip time between the medical carbon measurement source and the carbon measurement demand object at time t of the i-th medical carbon measurement source; t,i represents the time offset of the data packet sent by the ith medical carbon metering source at time t; V t,i represents the waiting speed of the i-th data at time t; D t,i Indicates the distance between the i-th medical carbon measurement source and the object that sends the carbon measurement request at time t; When the round-trip time values ​​from the same medical carbon measurement source are obtained multiple times, it is more reliable to smooth the round-trip time. Through the smoothing method, the calculation formula is as follows: N t,i =u×N t-1,i +e×(R t,i -N t-1,i ); Among them, R t,i represents the currently observed instantaneous round-trip time value, u = 1, e = 0.125; Under the same cumulative running time, the larger the offset, the more idle the medical carbon metering source is, and the larger the remaining available cumulative running time is; on the contrary, the smaller the offset, the smaller the remaining available cumulative running time of the medical carbon metering source is. The calculation formula is: M t,i =(1-σ)×M t-1,i +σ×(Data t,i -Data t-1,i ); Where, σ=0.5, Data t,i Indicates the time when the current capability attribute data is sent. t-1,i Indicates the time when the last capability attribute data was sent. The difference between the two is the time offset. Use the multiple capability attribute data contained in sequence as the basis for sorting and determine the medical carbon measurement source with the highest score; each of the four states can select a medical carbon measurement source, with a total of 16 capability attribute data as the input of the attribute set of the carbon measurement object to be measured; Calculate the optimal value for each state; Calculate the execution time of action a t Query the score of medical carbon measurement source; By adjusting the carbon metering attribute indicators of the attribute set of the object to be carbon metered, the state indicators are normalized, the value range of the action indicators is adjusted, the mapping relationship between the medical carbon metering source and the state indicator value is generated, and the medical carbon metering source with the highest score is determined as the final query result.

2. The carbon emission intelligent meter according to claim 1, characterized in that: Before performing a medical carbon metering source query, including steps for triggering a medical carbon metering source query; The reasons for triggering the query of medical carbon metering sources include at least: the carbon metering demand object queries the medical carbon metering source for the carbon metering demand object when initializing; or the carbon metering transmission channel connected deteriorates or is disconnected when the carbon metering demand object queries a new medical carbon metering source; When deciding whether the quality of the connected carbon metering transmission channel deteriorates or the connection is interrupted, the interruption probability of the current carbon metering demand object connecting to the medical carbon metering source is calculated, and the current carbon metering demand priority is used as the random seed, and the hit replacement probability is calculated using a random function, which is expressed as: （1）； Generate a random number below 200 using the rand() function. The formula is: （2）； Among them, H represents the medical carbon measurement source replacement certification mark, Timestamp represents the current timestamp, Indicates that the average interruption probability is less than or equal to the current interruption probability if It means the probability of executing the replacement of medical carbon metering source, p is the current interruption probability; P max It represents the maximum interruption probability, that is, when the interruption probability of the carbon metering transmission channel is greater than the maximum interruption probability, the probability calculation method will no longer be used for replacement identification, and the medical carbon metering source will be replaced directly. Mod100 means taking the remainder after dividing by 100.

3. The carbon emission intelligent meter according to claim 1, characterized in that: The step of optimizing the model index and weight of the carbon metering demand object attribute set of the carbon metering demand object receiving the medical carbon metering source includes the steps of: The carbon measurement demand object attribute set of the carbon measurement demand object learns environmental samples and optimizes the carbon measurement attribute model and model weight by interacting with the real environment; Optimize the carbon measurement demand object attribute set model; In the steps of interacting with the real environment, learning environmental samples and optimizing the carbon measurement attribute model and model weights, Optimize the carbon measurement attribute model and model weights, use local data and environmental status to train the carbon measurement demand object attribute set model, and transmit the weight results to a carbon emission storage end after multiple iterations. The carbon emission storage end will collect the weight results of all carbon measurement demand objects and calculate the gradient average.

Images