New energy consumption big data analysis optimization system

Abstract

The invention discloses a new energy consumption big data analysis and optimization system, and the system comprises a block division module which completes the division of a power utilization region through collecting the information of a power transmission line of a power grid in combination with geographic data; the electricity consumption monitoring module collects electricity consumption data of each block by means of an electricity meter, counts an average value of peak and valley periods and stores the average value into a database; the new energy generating capacity prediction module is used for collecting data such as historical weather and generating equipment parameters, and predicting generating capacity and generating a curve through a processing training model; and the optimization module is used for integrating data of other modules to construct a formula and setting constraint conditions, solving by utilizing a PuLP library to realize optimal distribution of the new energy electric power in different blocks in peak and valley periods, and maximally absorbing the new energy electric power. According to the system, various technical means and algorithms are comprehensively applied, the problem of new energy power distribution is effectively solved, the energy utilization efficiency is improved, and the system has great significance in promoting efficient application of new energy in a power grid.

Description

Technical Field

[0001] The present invention relates to the technical field of big data analysis, and in particular to a new energy consumption big data analysis and optimization system. Background Art

[0002] New energy, also known as unconventional energy, refers to various forms of energy other than traditional energy (such as coal, oil, natural gas, etc.). These energy sources are usually renewable, environmentally friendly, and low-carbon, including but not limited to solar energy, wind energy, geothermal energy, biomass energy, etc.

[0003] As an advanced data processing technology, big data analysis plays an important role in the consumption of new energy. Through in-depth mining and analysis of massive amounts of new energy power generation data, power consumption data, meteorological data, etc., it is possible to accurately predict new energy power generation, monitor power demand in real time, and optimize power dispatch and allocation strategies. This data-based decision-making method can significantly improve the efficiency of new energy power consumption, reduce wind and solar power abandonment, and promote the healthy development of the new energy industry.

[0004] In the process of power dispatching and distribution, existing technologies often rely on manual experience and rules and lack intelligent optimization algorithms, resulting in unreasonable distribution of new energy electricity in power consumption areas, local surplus or shortage problems, and limiting the further improvement of new energy consumption efficiency.

[0005] To this end, we propose a big data analysis and optimization system for new energy consumption. Summary of the invention

[0006] The purpose of the present invention is to provide a new energy consumption big data analysis and optimization system to solve the above-mentioned deficiencies in the prior art.

[0007] In order to achieve the above-mentioned purpose, the present invention provides the following technical solution: a new energy consumption big data analysis and optimization system, comprising:

[0008] Block division module: This module is connected to the power grid system and is used to obtain the power transmission line information of the power grid system in real time, and divide the power consumption area into blocks according to the actual layout of the transmission line and the power supply coverage;

[0009] Power consumption monitoring module: It is connected to the power secondary equipment and the power grid system to collect power consumption data in each block in real time, collect statistics on the collected data according to the preset time interval, and calculate the average power consumption of each block during the daily peak / valley period in the time interval;

[0010] Renewable energy power generation prediction module: It is used to extract historical meteorological data, real-time operating status parameters of new energy power generation equipment, and historical power generation efficiency data, and predict the renewable energy power generation within a period of time by using big data analysis algorithms and machine learning models, and generate and store power generation prediction curves;

[0011] Optimization module: It is respectively connected to the block division module, power consumption monitoring module and new energy power generation prediction module, receives the data transmitted by each module, and based on the average power consumption of each block and the predicted value of new energy power generation, allocates the new energy power to different blocks in different time periods (peak / valley time) through optimization algorithm, so as to maximize the consumption of new energy power in the power consumption area.

[0012] As a further description of the above technical solution: the steps of constructing the block division module are as follows:

[0013] A1: Use a multi-interface data acquisition card to directly connect to the power grid system switch through an optical fiber communication link to capture the dynamic information of the power transmission lines in the power grid system in real time;

[0014] A2: Use analytical algorithm software to parse the received power grid data. Through the parameters of the transmission line, including but not limited to line impedance, conductor model, branch node location, combined with the pre-entered regional geographic information data, a density-based clustering spatial partitioning algorithm is used to divide the power consumption area into blocks based on the power supply end and branch point of the line.

[0015] As a further description of the above technical solution: the operation steps of the power consumption monitoring module are as follows:

[0016] B1: The power parameters of current, voltage, active power and reactive power are collected in real time through the power secondary equipment of the power grid system, and the data collected by the power secondary equipment are exchanged through the wireless gateway;

[0017] B2: After the wireless gateway receives the data from the electricity meter, it uploads it to the electricity consumption monitoring module via the 4G network. The electricity consumption monitoring module runs a Python-based data analysis script, uses the Numpy library to perform mathematical operations according to the daily peak and valley period division standards, and calculates the average electricity consumption of each block during the peak / valley period. Finally, the results are stored in the MySQL relational database for use by other modules.

[0018] As a further description of the above technical solution: the specific steps of using the new energy power generation prediction module are as follows:

[0019] C1: Build a meteorological data collection submodule to find historical meteorological data on temperature, air pressure, precipitation, and light intensity in the renewable energy power generation area. Use the data cleaning tool OpenRefine to purify the collected meteorological data, remove invalid data and redundant information, and then store it in the local Hive data warehouse and archive it in a time series format.

[0020] C2: Install IoT sensor groups for new energy power generation equipment, including but not limited to speed sensors, temperature sensors, current sensors, etc., to monitor the operating status parameters of the equipment in real time, and aggregate the data of the equipment and upload it to the new energy power generation prediction module through the MQTT protocol;

[0021] C3: After the renewable energy power generation prediction module receives the operating status parameters of renewable energy power generation equipment, it builds a model through the Python-based Scikit-learn machine learning library, inputs historical meteorological data, power generation equipment status parameters, and historical power generation efficiency data as feature vectors into the model for training, and optimizes the model parameters through cross-validation.

[0022] 6. As a further description of the above technical solution: The specific steps for achieving the maximum consumption of new energy electricity in the power consumption area are:

[0023] D1: The optimization module integrates the data from the block division module, the power consumption monitoring module and the new energy power generation prediction module, and constructs the following formula:

[0024]

[0025] Among them, S represents the number of divided blocks, X fi represents the total renewable energy power allocated to the i-th block during the peak period, X gi represents the total amount of renewable energy electricity allocated to the ith block during the off-peak period, Sfi represents the average electricity consumption of the ith block during the daily peak period, Sgi represents the average daily electricity consumption of the ith block during the off-peak period, and Y fi is a binary variable, indicating whether the i-th block allocates renewable energy power during the peak period, 1 means allocation, 0 means no allocation, Y gi It is also a binary variable, indicating whether the i-th block allocates renewable energy power during the valley period;

[0026] D2: After the formula in step D1 is constructed, the following constraints are imposed on the formula:

[0027]

[0028] Where X 总 Represents the total renewable energy power generation predicted in the entire time interval;

[0029] The relationship between the binary variable and the allocation amount is expressed by the following formula:

[0030] If Y f i=0, then X f i=0

[0031] If Y g i=0, then X g i=0

[0032] Then, we use the following formula to enter the non-negative constraint: X f i,X g i is greater than or equal to 0, and Y fi and Y gi The following constraints need to be satisfied: fi , Y gi ∈{0,1}.

[0033] As a further description of the above technical solution: the model used in step C3 includes but is not limited to a random forest regression model and a neural network model.

[0034] In the above technical solution, the new energy consumption big data analysis and optimization system provided by the present invention has the following beneficial effects:

[0035] Through the coordinated work of the block division module, electricity consumption monitoring module, new energy power generation prediction module and optimization module, the system can accurately predict the new energy power generation, and optimize the allocation of new energy power according to the electricity demand and new energy power generation situation of each block, thereby improving the consumption efficiency of new energy power and reducing waste. DETAILED DESCRIPTION

[0036] The embodiment of the present invention provides a technical solution: a new energy consumption big data analysis and optimization system, comprising:

[0037] Block division module: This module is connected to the power grid system and is used to obtain the power transmission line information of the power grid system in real time, and divide the power consumption area into blocks according to the actual layout of the transmission line and the power supply coverage;

[0038] It should be noted that the steps for building the block division module are as follows:

[0039] A1: Use a multi-interface data acquisition card to directly connect to the power grid system switch through an optical fiber communication link to capture the dynamic information of the power transmission lines in the power grid system in real time;

[0040] A2: Use analytical algorithm software to analyze the received power grid data, and then combine the parameters of the transmission line, including but not limited to line impedance, conductor model, branch node location, and pre-entered regional geographic information data, and use the density clustering-based spatial partitioning algorithm to divide the power consumption area into blocks based on the power supply end and branch point of the line;

[0041] Another way to divide the blocks can be used here. Since the power grid planning often has a paper version of the planning map as a record, you can choose to use high-resolution image acquisition equipment, such as a professional-grade high-definition scanner or a high-precision camera carried by a drone, to collect images of the paper or electronic planning map of the power grid construction to obtain a clear power grid line layout map;

[0042] If it is a paper drawing, it is necessary to digitize the drawing information through optical character recognition (OCR) technology combined with graphic recognition algorithm; if it is an electronic drawing, the format conversion and data analysis are directly performed;

[0043] Use computer-aided design software, such as CAD, to import the parsed power grid planning map data and extract key information of the transmission line, including line direction, tower coordinates, substation location, etc., and then combine it with the regional geographic boundary data in the geographic information system (GIS) software, and use the rule-based spatial division algorithm to accurately divide the power consumption area into blocks based on factors such as line power supply area, branch conditions, and geographical barriers; both block division methods can be used;

[0044] Power consumption monitoring module: It is connected to the power secondary equipment and the power grid system to collect power consumption data in each block in real time, collect statistics on the collected data according to the preset time interval, and calculate the average power consumption of each block during the daily peak / valley period in the time interval;

[0045] The operation steps of the power consumption monitoring module are as follows:

[0046] B1: The power parameters of current, voltage, active power and reactive power are collected in real time through the power secondary equipment of the power grid system, and the data collected by the power secondary equipment are exchanged through the wireless gateway;

[0047] B2: After the wireless gateway receives the data from the electricity meter, it uploads it to the electricity consumption monitoring module via the 4G network. The electricity consumption monitoring module runs a Python-based data analysis script, uses the Numpy library to perform mathematical operations according to the daily peak and valley time division standards, calculates the average electricity consumption of each block during the peak / valley period, and finally stores the results in the MySQL relational database for use by other modules;

[0048] It should be noted that when exchanging data through the wireless gateway, data compression and encryption transmission technology is used to compress the collected power consumption data to about 70% of the original volume according to a lossless compression algorithm, such as the LZ77 algorithm, and then encrypt and transmit it, which not only improves transmission efficiency, but also ensures data security and prevents theft or tampering during transmission;

[0049] In addition, for the Python-based data analysis script running in the power consumption monitoring module, in addition to calculating the average power consumption during peak / valley periods, it can also use statistical methods (such as the box plot method) to identify and eliminate abnormal data points that exceed the normal power consumption range (mean ± 3 times the standard deviation) to prevent abnormal values ​​from interfering with the average value calculation;

[0050] After the results are stored in the MySQL relational database, the database table structure design follows the third normal form to reduce data redundancy, improve storage efficiency, establish indexes for the stored average power consumption and related data, optimize query performance, and ensure query response efficiency when other modules call data.

[0051] Renewable energy power generation prediction module: It is used to extract historical meteorological data, real-time operating status parameters of new energy power generation equipment, and historical power generation efficiency data, and predict the renewable energy power generation within a period of time by using big data analysis algorithms and machine learning models, and generate and store power generation prediction curves;

[0052] The specific steps for using the new energy power generation prediction module are as follows:

[0053] C1: Build a meteorological data collection submodule to find historical meteorological data on temperature, air pressure, precipitation, and light intensity in the renewable energy power generation area. Use the data cleaning tool OpenRefine to purify the collected meteorological data, remove invalid data and redundant information, and then store it in the local Hive data warehouse and archive it in a time series format.

[0054] C2: Install IoT sensor groups for new energy power generation equipment, including but not limited to speed sensors, temperature sensors, current sensors, etc., to monitor the operating status parameters of the equipment in real time, and aggregate the data of the equipment and upload it to the new energy power generation prediction module through the MQTT protocol;

[0055] C3: After the new energy power generation prediction module receives the operating status parameters of the new energy power generation equipment, it builds a model through the Python-based Scikit-learn machine learning library, inputs historical meteorological data, power generation equipment status parameters, and historical power generation efficiency data as feature vectors for model training, and optimizes model parameters through cross-validation;

[0056] C4: Use the trained model to predict the renewable energy power generation within a time interval (e.g., one month), and generate a power generation prediction curve within the time interval;

[0057] The model used in step C3 includes but is not limited to a random forest regression model and a neural network model;

[0058] For IoT sensor groups installed on new energy power generation equipment, the layout should be optimized according to the equipment type and operating characteristics. For example, stress and vibration sensors should be reasonably distributed at the root of wind turbine blades, hubs and key parts of the nacelle to accurately capture changes in the mechanical properties of the equipment; in the solar photovoltaic panel array, light and temperature sensors should be arranged crosswise in rows and columns to ensure comprehensive monitoring of the power generation unit status. In addition, the sensor should have a built-in self-diagnosis program to automatically detect the working status of the sensor at regular intervals, such as two hours. Once a fault is found, the operation and maintenance personnel will be notified in reverse through the MQTT protocol to ensure the continuity of data collection;

[0059] When constructing the model, when the input feature dimension is high, such as more than 20 meteorological and equipment status features, and the data scale is large, such as the power generation data of each time period accumulated for several years, a multi-layer perceptron architecture with 3 hidden layers is adopted, and the number of neurons in the hidden layer is set according to a decreasing rule to gradually extract the deep features of the data. The hidden layer uniformly uses the ReLU function, which can effectively solve the gradient disappearance problem and accelerate the convergence speed of model training; and because the prediction of power generation is a continuous value regression problem, the linear function can directly output the predicted power generation value, so the output layer uses a linear function;

[0060] In addition to the combination of L1 and L2 regularization, the early stopping method is also introduced. During the training process, the data is divided into training set, validation set and test set. After each training epoch (the whole data is traversed once), the model performance is evaluated on the validation set, and the mean square error is monitored. When the mean square error is no longer decreasing or even increases after 5 consecutive epochs, the training is stopped and the current optimal model parameters are saved. At the same time, the Dropout technology is used to randomly disconnect the connections between hidden layer neurons with a certain probability (such as 0.2-0.3) during the training process, forcing the neurons to learn more representative features, further improving the robustness of the model and preventing overfitting;

[0061] Here, the prediction of the power generation of new energy power is carried out by environmental factors, which belongs to the existing technology and can also be carried out by other existing methods, which will not be described in detail here;

[0062] Optimization module: It is respectively connected to the block division module, the power consumption monitoring module, and the new energy power generation prediction module, receives the data transmitted by each module, and allocates the new energy power to different blocks at different time periods (peak / valley time) through the optimization algorithm based on the average power consumption of each block and the predicted value of new energy power generation, so as to realize the maximum consumption of new energy power in the power consumption area;

[0063] The specific steps to achieve maximum consumption of new energy electricity in the power consumption area are:

[0064] D1: The optimization module integrates the data from the block division module, the power consumption monitoring module and the new energy power generation prediction module, and constructs the following formula:

[0065]

[0066] Among them, S represents the number of divided blocks, X fi represents the total renewable energy power allocated to the i-th block during the peak period, X gi represents the total amount of renewable energy electricity allocated to the ith block during the off-peak period, Sfi represents the average electricity consumption of the ith block during the daily peak period, Sgi represents the average daily electricity consumption of the ith block during the off-peak period, and Y fi is a binary variable, indicating whether the i-th block allocates renewable energy power during the peak period, 1 means allocation, 0 means no allocation, Y gi It is also a binary variable, indicating whether the i-th block allocates renewable energy power during the valley period;

[0067] D2: After the formula in step D1 is constructed, the following constraints are imposed on the formula:

[0068]

[0069] Where X 总 Represents the total renewable energy power generation predicted in the entire time interval;

[0070] The relationship between the binary variable and the allocation amount is expressed by the following formula:

[0071] If Y f i=0, then X f i=0

[0072] If Y g i=0, then X g i=0

[0073] Then, we use the following formula to enter the non-negative constraint: X f i,X g i is greater than or equal to 0, and Y fi and Y giThe following constraints need to be satisfied: fi , Y gi ∈{0,1};

[0074] It should be noted here that the steps of model building are: first

[0075] Create a new linear programming model in the PuLP library and define two types of decision variables. One is a continuous variable representing the allocation of renewable energy power in peak and valley periods of each block. Its lower limit is 0 to ensure that the allocated power is non-negative. The other is a binary variable, which is used to mark whether the corresponding block participates in the allocation of renewable energy power in the peak and valley periods. The value is limited to 0 or 1. According to the optimization goal, the objective function expression is constructed using the grammatical rules supported by the library. At the same time, the total power generation constraints and the constraints associated with the binary variables and the allocated power are converted into standard constraint settings within the model.

[0076] The built-in CBC solver of the practical PuLP library is used to start the solution process. A reasonable upper limit on the solution time is set in advance according to the problem scale (mainly referring to the number of blocks) to prevent the solution process from falling into a meaningless long loop. After the solution is completed, the optimal value of each decision variable is accurately extracted from the returned result set. These values ​​​​clarify the precise allocation plan of new energy electricity during peak and valley periods in each block.

[0077] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that, for those skilled in the art, the described embodiments can be modified in various ways without departing from the spirit and scope of the present invention. Therefore, the above description is illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. New energy consumption big data analysis and optimization system, characterized by: include: Block division module: This module is connected to the power grid system and is used to obtain the power transmission line information of the power grid system in real time, and divide the power consumption area into blocks according to the actual layout of the transmission line and the power supply coverage; Power consumption monitoring module: It is connected to the power secondary equipment and the power grid system to collect power consumption data in each block in real time, collect statistics on the collected data according to the preset time interval, and calculate the average power consumption of each block during the daily peak / valley period in the time interval; Renewable energy power generation prediction module: It is used to extract historical meteorological data, real-time operating status parameters of new energy power generation equipment, and historical power generation efficiency data, and predict the renewable energy power generation within a period of time by using big data analysis algorithms and machine learning models, and generate and store power generation prediction curves; Optimization module: It is respectively connected to the block division module, power consumption monitoring module and new energy power generation prediction module, receives the data transmitted by each module, and based on the average power consumption of each block and the predicted value of new energy power generation, allocates the new energy power to different blocks in different time periods (peak / valley time) through optimization algorithm, so as to maximize the consumption of new energy power in the power consumption area.

2. The new energy consumption big data analysis and optimization system according to claim 1 is characterized in that: The steps of constructing the block division module are as follows: A1: Use a multi-interface data acquisition card to directly connect to the power grid system switch through an optical fiber communication link to capture the dynamic information of the power transmission lines in the power grid system in real time; A2: Use analytical algorithm software to parse the received power grid data. Through the parameters of the transmission line, including but not limited to line impedance, conductor model, branch node location, combined with the pre-entered regional geographic information data, a density-based clustering spatial partitioning algorithm is used to divide the power consumption area into blocks based on the power supply end and branch point of the line.

3. The new energy consumption big data analysis and optimization system according to claim 2 is characterized in that: The operation steps of the power consumption monitoring module are as follows: B1: The power parameters of current, voltage, active power and reactive power are collected in real time through the power secondary equipment of the power grid system, and the data collected by the power secondary equipment are exchanged through the wireless gateway; B2: After the wireless gateway receives the data from the electricity meter, it uploads it to the electricity consumption monitoring module via the 4G network. The electricity consumption monitoring module runs a Python-based data analysis script, uses the Numpy library to perform mathematical operations according to the daily peak and valley period division standards, and calculates the average electricity consumption of each block during the peak / valley period. Finally, the results are stored in the MySQL relational database for use by other modules.

4. The new energy consumption big data analysis and optimization system according to claim 3 is characterized in that: The specific steps of using the new energy power generation prediction module are as follows: C1: Build a meteorological data collection submodule to find historical meteorological data on temperature, air pressure, precipitation, and light intensity in the renewable energy power generation area. Use the data cleaning tool OpenRefine to purify the collected meteorological data, remove invalid data and redundant information, and then store it in the local Hive data warehouse and archive it in a time series format. C2: Install IoT sensor groups for new energy power generation equipment, including but not limited to speed sensors, temperature sensors, current sensors, etc., to monitor the operating status parameters of the equipment in real time, and aggregate the data of the equipment and upload it to the new energy power generation prediction module through the MQTT protocol; C3: After the new energy power generation prediction module receives the operating status parameters of the new energy power generation equipment, it builds a model through the Python-based Scikit-learn machine learning library, inputs historical meteorological data, power generation equipment status parameters, and historical power generation efficiency data as feature vectors for model training, and optimizes model parameters through cross-validation; C4: Use the trained model to predict the renewable energy power generation within a time interval (e.g., one month), and generate a power generation prediction curve within the time interval.

5. The new energy consumption big data analysis and optimization system according to claim 4 is characterized in that: The specific steps for achieving maximum consumption of new energy power in the power consumption area are: D1: The optimization module integrates the data from the block division module, the power consumption monitoring module and the new energy power generation prediction module, and constructs the following formula: Among them, S represents the number of divided blocks, X fi represents the total renewable energy power allocated to the i-th block during the peak period, X gi represents the total amount of renewable energy electricity allocated to the ith block during the off-peak period, Sfi represents the average electricity consumption of the ith block during the daily peak period, Sgi represents the average daily electricity consumption of the ith block during the off-peak period, and Y fi is a binary variable, indicating whether the i-th block allocates renewable energy power during the peak period, 1 means allocation, 0 means no allocation, Y gi It is also a binary variable, indicating whether the i-th block allocates renewable energy power during the valley period; D2: After the formula in step D1 is constructed, the following constraints are imposed on the formula: Where X 总 Represents the total renewable energy power generation predicted in the entire time interval; The relationship between the binary variable and the allocation amount is expressed by the following formula: If Y f i=0, then X f i=0 If Y g i=0, then X g i=0 Then the non-negative constraint is performed by the following formula: X f i,X g i is greater than or equal to 0, and Y fi and Y gi The following constraints need to be satisfied: fi , Y gi ∈{0,1}.

6. The new energy consumption big data analysis and optimization system according to claim 4 is characterized in that: The model used in step C3 includes but is not limited to a random forest regression model and a neural network model.

Images