Carbon flow monitoring and accounting method based on Internet of Things

Abstract

The invention discloses a carbon flow monitoring and accounting method based on the Internet of Things. The method specifically comprises the following steps: S1, collecting power flow information of a power system; s2, performing flow calculation on the power network to obtain an active power flow matrix F of the whole power network; s3, performing hierarchical iterative calculation on the power grid carbon flow density, and gradually obtaining carbon flow data of the whole power grid; s4, arranging the carbon flow data to obtain a carbon flow matrix C, and providing a final visual result; s5, collecting carbon emission data; s6, establishing a carbon flow neural network model; s7, training the carbon flow neural network model; and S8, checking the carbon emission of the whole system by using the carbon flow neural network model. According to the invention, an Internet of Things data acquisition technology and a data analysis visualization technology are mainly utilized, the flow path of the carbon flow in the whole power network is visually displayed for a power system manager, and low carbon and greenization of the whole system are helped to be realized.

Description

Technical Field

[0001] The present invention relates to the field of carbon flow of the Internet of Things, and in particular to a carbon flow monitoring and accounting method based on the Internet of Things. Background Art

[0002] Carbon flow monitoring provides a feasible implementation path for the allocation of carbon responsibility in the power system. However, there is currently no technology that can intuitively display the flow of carbon flow in the entire power network for monitoring and accounting. Therefore, there is an urgent need for a carbon flow monitoring and accounting method, which allows power system managers to intuitively display the flow path of carbon flow in the entire power network and carbon emission data including but not limited to carbon emissions from fossil energy combustion, carbon emissions from production processes, carbon emissions from electricity and water resources, and other carbon emissions, to help achieve low-carbonization and greening of the entire system. Summary of the invention

[0003] In view of the above shortcomings, the present invention provides a carbon flow monitoring and accounting method based on the Internet of Things, which specifically includes the following steps:

[0004] S1. Collect power system flow information, including node power, network loss power, and line topology;

[0005] S2. Calculate the flow of the power network to obtain the active power flow matrix F of the entire power network;

[0006] S3. Perform hierarchical iterative calculation on the carbon flow density of the power grid to gradually obtain the carbon flow data of the entire power grid;

[0007] S4. Obtain the carbon flow matrix C by sorting the carbon flow data and provide the final visualization result;

[0008] S5. Collect carbon emission data;

[0009] S6. Establish a neural network model for carbon flow;

[0010] S7. Training the carbon flow neural network model, randomly dividing the data set obtained in step S5 into a training set and a validation set at a ratio of 9:1; using the training set to train the carbon flow neural network model and verifying it with the validation set to obtain the optimal model;

[0011] S8. Use the trained carbon flow neural network model to calculate the carbon emissions of the entire system.

[0012] In particular, the active power flow matrix F in step S2 is as follows:

[0013] F k5 =P ij -T ij (1);

[0014] In the formula, F k5 represents the outflow power of the kth line, P ij represents the active power provided by node i to node j; T ij represents the power loss between node i and node j and i≠j.

[0015] In particular, the carbon flow density iteration algorithm in step S3 is as follows:

[0016] (1) Starting from the initial node, the carbon flow density of the line leading to the next level is calculated according to the known carbon emission intensity of the power flow starting node. The calculation process is as follows:

[0017] F k4 / F k5 =ρ i ' j / ρ i ' j ' (2);

[0018] Where Fk4 represents the injected power of the power flow on the kth line; Fk5 represents the outflow power of the power flow on the kth line; ρ i ' j represents the carbon flow density of the starting node of the power flow from node i to node j and ρ i ' j =ρ i ρ i represents the carbon emission intensity of node i; ρ i ' j ' represents the end node carbon flow density of the power flow from node i to node j;

[0019] (2) Detect whether the node at this level has a next-level node; if so, calculate the carbon emission intensity of the next-level node and re-execute step (1); if there is no next-level node, terminate the algorithm and output the carbon flow matrix C of the entire power network.

[0020] In particular, the calculation method of carbon emission intensity of the next level node is as follows:

[0021]

[0022] In the formula, ρ j Represents the carbon emission intensity of node j.

[0023] In particular, the first column of the carbon flow matrix C in step S4 is the node number i;

[0024] The second column of the carbon flow matrix C in step S4 is the power injected into the grid by node i, and the calculation process is as follows:

[0025]

[0026] The third column of the carbon flow matrix C in step S4 is the carbon emission intensity of node i:

[0027] C i3 =ρ i (5);

[0028] The fourth column of the carbon flow matrix C in step S4 is the carbon flow time density of node i (in kg CO 2 / s):

[0029] C i4 =C i3 *C i2 (6).

[0030] In particular, step S6 establishes a carbon flow neural network model as follows:

[0031] The data matching error function is:

[0032]

[0033] In the formula, NN(C i ) represents the carbon flow neural network model for the carbon flow matrix C i The carbon emission prediction of the whole system at y i Indicates the true value of carbon emissions of the entire system at this time;

[0034] The physical governing equation is:

[0035]

[0036] in, represents the gradient operation, Represents Laplace operation;

[0037] The loss function caused by the physical control equation is:

[0038]

[0039] In particular, the calculation process of the carbon flow neural network model in step S8 is as follows:

[0040] Input the carbon flow matrix in a sliding window into the trained carbon flow neural network model to obtain the predicted value sequence Pre(t) of the carbon emissions of the entire system;

[0041] Compare with the currently reported carbon emission sequence Rel(t) and calculate the error ε. The calculation method is as follows:

[0042]

[0043] In the formula, n represents the total number of sample points in the sliding window;

[0044] The mean, variance, and range of the prediction errors of n sample points in the sliding window are as follows:

[0045]

[0046] The error threshold is calculated as follows:

[0047]

[0048] In the formula, ε up and ε down are the upper and lower thresholds of the error ε respectively; α is the coefficient of the standard deviation; β is the coefficient of the range.

[0049] In particular, the coefficient α of the standard deviation and the coefficient β of the range are determined as follows:

[0050]

[0051] Generally speaking, the greater the volatility of the forecast sequence, the less reliable the forecast result, and the threshold range for judging the error should be expanded accordingly. Considering the above principles, the method for determining the coefficient α of the standard deviation and the coefficient β of the range is confirmed by formula (13).

[0052] When the error between the carbon emission data estimated by the carbon flow neural network model and the accounting result exceeds a threshold, the accounting result is considered to have failed; when the error between the carbon emission data estimated by the carbon flow neural network model and the accounting result is within a threshold range, the accounting result is considered to have succeeded.

[0053] Compared with the prior art, the beneficial effects of the present invention are: utilizing Internet of Things data collection technology and data analysis visualization technology to intuitively display the flow path of carbon flow in the entire power network for power system managers, helping to achieve low-carbonization and greening of the entire system. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for describing the embodiments are briefly introduced below.

[0055] Figure 1 This is a flow chart of a carbon flow monitoring and accounting method based on the Internet of Things according to an embodiment of the present invention;

[0056] Figure 2 A carbon flow path visualization result diagram of an IEEE14-node system of a carbon flow monitoring and accounting method based on the Internet of Things according to an embodiment of the present invention;

[0057] Figure 3This is a structural diagram of a carbon flow neural network model of a carbon flow monitoring and accounting method based on the Internet of Things according to an embodiment of the present invention. DETAILED DESCRIPTION

[0058] The technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all of the embodiments. Based on the embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0059] In the description of the present invention, it should be noted that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside", etc. indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction. Therefore, they cannot be understood as limitations on the present invention.

[0060] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal connection of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. In addition, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0061] The present invention provides a carbon flow monitoring and accounting method based on the Internet of Things, such as Figure 1 As shown, the specific steps include:

[0062] S1. Collect power system flow information, including node power, network loss power, and line topology; and obtain the active power flow of the system through data processing.

[0063] Taking the IEEE14-node system as an example, the system includes two generator nodes. Node 1 is a coal-fired generator set with a carbon emission intensity of 0.855 kgCO2 / (kWh), and node 2 is a natural gas generator set with a carbon emission intensity of 0.382 kgCO2 / (kW-h).

[0064] The system has a total of 14 nodes and 21 lines. The node power information is stored in the matrix P. The element P in the matrix P is ijIt represents the active power provided by node i to node j, in MW. The network loss power information and line topology information are stored in the matrix T. The element T in the matrix T ij If it is 0, it means there is no path between node i and node j. ij If it is non-zero, it means there is a path between node i and node j. Its value represents the network loss of the line, in MW.

[0065] S2. Calculate the flow of the power network to obtain the active power flow matrix F of the entire power network.

[0066] The active power flow of the power network is obtained by data processing, and the processing method is as follows:

[0067] For the upper triangular matrix T′ of matrix T, traverse all its upper triangular elements in the order of rows first and columns later. If T′ ij =1 and i≠j, then add a row to the active power flow matrix F. The first column of the row is the row number, the second column is the number i (indicating the starting node number of the power flow), the third column is the number j (indicating the ending node number of the power flow), and the fourth column is the injection power of the power flow on the line (i.e., P ij ), the fifth column is the outflow power of the power flow on the line, and the calculation process is as follows:

[0068] F k5 =P ij -T ij (1);

[0069] In the formula, F k5 Represents the outflow power of the kth line in the power network.

[0070] Since the system has a total of 21 lines, the F matrix in this case is a 21*5 matrix. The active power flow matrix data of the IEEE14-bus system is shown in Table 1.

[0071] Table 1 Active power flow matrix data table of IEEE14 bus system

[0072]

[0073]

[0074] S3. Perform hierarchical iterative calculation on the carbon flow density of the power grid to gradually obtain the carbon flow data of the entire power grid; the iterative algorithm flow of the carbon flow density is as follows:

[0075] (1) Starting from the initial node, the carbon flow density of the line leading to the next level is calculated according to the known carbon emission intensity of the power flow starting node. The calculation process is as follows:

[0076] Fk4 / F k5 =ρ i ' j / ρ i ' j ' (2);

[0077] In the formula, F k4 represents the injected power of the power flow on the kth line; F k5 represents the outflow power of the power flow on the kth line; ρ i ' j represents the carbon flow density of the starting node of the power flow from node i to node j and ρ i ' j =ρ i ρ i represents the carbon emission intensity of node i; ρ i ' j ' represents the terminal node carbon flow density of the power flow from node i to node j.

[0078] (2) Check whether the node at this level has a next-level node. If so, calculate the carbon emission intensity of the next-level node and re-execute step (1); if there is no next-level node, terminate the algorithm and output the carbon flow density matrix of the entire power network. The calculation method of the carbon emission intensity of the next-level node is as follows:

[0079]

[0080] In the formula, ρ j Represents the carbon emission intensity of node j.

[0081] S4. Obtain the carbon flow matrix C by sorting the carbon flow data and provide the final visualization result. As shown in Table 2, the carbon flow matrix C in this case is a 14-row matrix, and each row stores the carbon flow information of a corresponding node.

[0082] The first column of the carbon flow matrix C is the node number i.

[0083] The second column of the carbon flow matrix C is the power injected into the grid by node i. The calculation process is as follows:

[0084]

[0085] The third column of the carbon flow matrix C is the carbon emission intensity of node i:

[0086] C i3 =ρ i (5);

[0087] The fourth column of the carbon flow matrix C is the carbon flow time density of node i (in kg CO 2 / s):

[0088] C i4 =C i3 *C i2 (6);

[0089] Visualizing the carbon flow matrix C can intuitively show the dynamic changes of carbon flow.

[0090] Table 2 Carbon flow matrix data table of IEEE14 node system

[0091]

[0092] S5. Collect carbon emission data, including but not limited to real-time data y of total carbon emissions including carbon emissions from fossil energy combustion, carbon emissions from production processes, carbon emissions from electricity and water resources, and other carbon emissions, as well as the corresponding carbon flow matrix C in the same time period.

[0093] S6. Establish a carbon flow neural network model; the data matching error function is:

[0094]

[0095] In the formula, NN(C i ) represents the carbon flow neural network model for the carbon flow matrix C i The carbon emission prediction of the whole system at y i Indicates the true value of carbon emissions of the entire system at this time.

[0096] The physical governing equation is:

[0097]

[0098] in, represents the gradient operation, Represents the Laplace operation.

[0099] The loss function caused by the physical control equation is:

[0100]

[0101] S7. Train the carbon flow neural network model, and randomly divide the data set obtained in step S5 into a training set and a validation set at a ratio of 9:1. Use the training set to train the carbon flow neural network model and validate it with the validation set to obtain the optimal model.

[0102] As shown in Table 3, the prediction accuracy of the carbon flow neural network model is demonstrated.

[0103] Table 3 Prediction accuracy of carbon flow neural network model

[0104]

[0105] S8. Use the trained carbon flow neural network model to calculate the carbon emissions of the entire system. The calculation steps are as follows: Input the carbon flow matrix in a sliding window into the trained optimal physical information neural network model to obtain the carbon emission prediction value sequence Pre(t) for the current entire system. Generally, a five-day sliding window is used, and the carbon flow matrix of the system to be calculated is obtained every hour in a sliding window. In this case, a total of 120 sample points can be obtained.

[0106] Compare with the currently reported carbon emission sequence Rel(t) and calculate the error ε. The calculation method is as follows:

[0107]

[0108] In the formula, n represents the total number of sample points in the sliding window.

[0109] The mean, variance, and range of the prediction errors of n sample points in the sliding window are as follows:

[0110]

[0111] The error threshold is calculated as follows:

[0112]

[0113] In the formula, ε up and ε down are the upper and lower thresholds of the error ε respectively; α is the coefficient of the standard deviation; β is the coefficient of the range.

[0114] Generally speaking, the greater the volatility of the forecast sequence, the less reliable the forecast result, and the threshold range for judging the error should be expanded accordingly. Taking the above principles into account, the method for determining α and β is as follows:

[0115]

[0116] When the error between the carbon emission data estimated by the carbon flow neural network model and the accounting result exceeds a threshold, the accounting result is considered to have failed; when the error between the carbon emission data estimated by the carbon flow neural network model and the accounting result is within a threshold range, the accounting result is considered to have succeeded.

[0117] The carbon flow path visualization result of the IEEE 14-node system according to the embodiment of the present invention is shown in the figure Figure 2 As shown in the figure, the established carbon flow neural network model structure is as follows Figure 3 shown.

[0118] This embodiment proposes a carbon flow monitoring and visualization method based on the Internet of Things. The method uses the Internet of Things data acquisition technology and data analysis visualization technology, and mainly includes four steps: data collection, power flow calculation, carbon flow density iterative calculation, and carbon flow path visualization. First, the power system flow information is collected, including node power, network loss power, and line topology. Secondly, the flow is calculated for the power network to obtain the active power flow matrix F of the entire power network. Then, the carbon flow density of the power grid is calculated hierarchically and iteratively, and the carbon flow matrix C of the entire power network is gradually obtained, and the final visualization result is provided. Finally, a carbon flow neural network model is established to calculate the total carbon emissions of the entire system. This method will intuitively display the flow path of carbon flow in the entire power network for power system managers, including but not limited to carbon emissions from fossil energy combustion, carbon emissions from production processes, carbon emissions from electricity and water resources, and other carbon emissions, and help achieve the low-carbonization and greening of the entire system.

[0119] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A carbon flow monitoring and accounting method based on the Internet of Things, specifically comprising the following steps: S1. Collect power system flow information, including node power, network loss power, and line topology; S2. Calculate the flow of the power network to obtain the active power flow matrix F of the entire power network; S3. Perform hierarchical iterative calculation on the carbon flow density of the power grid to gradually obtain the carbon flow data of the entire power grid; S4. Obtain the carbon flow matrix C by sorting the carbon flow data and provide the final visualization result; S5. Collect carbon emission data; S6. Establish a neural network model for carbon flow; S7. Training the carbon flow neural network model, randomly dividing the data set obtained in step S5 into a training set and a validation set at a ratio of 9:1; using the training set to train the carbon flow neural network model and verifying it with the validation set to obtain the optimal model; S8. Use the trained carbon flow neural network model to calculate the carbon emissions of the entire system.

2. According to the carbon flow monitoring and accounting method based on the Internet of Things according to claim 1, the active power flow matrix F in step S2 is formulated as follows: F k5 =P ij -T ij (1); In the formula, F k5 represents the outflow power of the kth line, P ij represents the active power provided by node i to node j; T ij represents the power loss between node i and node j and i≠j.

3. According to the method for carbon flow monitoring and accounting based on the Internet of Things in claim 1, the carbon flow density iteration algorithm in step S3 is as follows: (1) Starting from the initial node, the carbon flow density of the line leading to the next level is calculated according to the known carbon emission intensity of the power flow starting node. The calculation process is as follows: F k4 / F k5 =p' ij / r” ij (2); Where, Fk4 represents the injected power of the power flow on the kth line; Fk5 represents the outflow power of the power flow on the kth line; ρ' ij represents the starting node carbon flow density of the power flow from node i to node j and ρ' ij =ρ i ρ i represents the carbon emission intensity of node i; ρ" ij The carbon flow density at the end node represents the power flow from node i to node j; (2) Detect whether the node at this level has a next-level node; if so, calculate the carbon emission intensity of the next-level node and re-execute step (1); if there is no next-level node, terminate the algorithm and output the carbon flow matrix C of the entire power network.

4. According to the carbon flow monitoring and accounting method based on the Internet of Things according to claim 3, the carbon emission intensity calculation method of the next level node is as follows: In the formula, ρ j Represents the carbon emission intensity of node j.

5. According to the method for carbon flow monitoring and accounting based on the Internet of Things in claim 1, the first column of the carbon flow matrix C in step S4 is the node number i; The second column of the carbon flow matrix C in step S4 is the power injected into the grid by node i, and the calculation process is as follows: The third column of the carbon flow matrix C in step S4 is the carbon emission intensity of node i: C i3 =ρ i (5); The fourth column of the carbon flow matrix C in step S4 is the carbon flow time density of node i (in kg CO 2 / s): C i4 =C i3 *C i2 (6)。 6. According to the method for carbon flow monitoring and accounting based on the Internet of Things in claim 1, step S6 establishes a carbon flow neural network model as shown below: The data matching error function is: In the formula, NN(C i ) represents the carbon flow neural network model for the carbon flow matrix C i The carbon emission prediction of the whole system at y i Indicates the true value of carbon emissions of the entire system at this time; The physical governing equation is: in, represents the gradient operation, Represents Laplace operation; The loss function caused by the physical control equation is:

7. According to the method for carbon flow monitoring and accounting based on the Internet of Things in claim 1, the carbon flow neural network model accounting process in step S8 is as follows: Input the carbon flow matrix in a sliding window into the trained carbon flow neural network model to obtain the predicted value sequence Pre(t) of the carbon emissions of the entire system; Compare with the currently reported carbon emission sequence Rel(t) and calculate the error ε. The calculation method is as follows: In the formula, n represents the total number of sample points in the sliding window; The mean, variance, and range of the prediction errors of n sample points in the sliding window are as follows: The error threshold is calculated as follows: In the formula, ε up and ε down are the upper and lower thresholds of the error ε respectively; α is the coefficient of the standard deviation; β is the coefficient of the range.

8. According to the carbon flow monitoring and accounting method based on the Internet of Things according to claim 7, the method for determining the coefficient α of the standard deviation and the coefficient β of the range is as follows:

Images