Electric travel cloud management and energy efficiency optimization system

Abstract

An electric travel cloud management and energy efficiency optimization system disclosed by the present invention is composed of a terminal device, a cloud end, an execution device, a plurality of relays and a plurality of sockets, the terminal device comprises a master control user end and a common user end, the cloud end comprises a data management module and a data big model, and the data big model comprises a face comparison module, a face recognition module and a face recognition module. The execution device comprises a Wi-Fi module, a single-chip microcomputer, an alarm module, an electric energy meter module and a battery detection module, and the execution device comprises a wireless fidelity (Wi-Fi) module, a single-chip microcomputer, an alarm module, an electric energy meter module and a battery detection module. The terminal equipment, the cloud and the execution equipment communicate through an MQTT protocol, charging data of an authorized user of the terminal equipment is collected and transmitted to the cloud for analysis, an instruction is issued, the execution equipment operates, the charging efficiency is improved, energy consumption is optimized through a flexible pricing mechanism and an off-peak charging strategy, and the charging efficiency is improved. And convenient and safe charging service is provided for residents.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent data cloud processing, and in particular to a cloud-based Internet of Things charging management system. Background Art

[0002] With the continuous development of society and science and technology, the Internet of Things technology is also advancing day by day, and smart products are gradually entering every family. However, for most young people, the residential houses where they live are still traditional residential houses, and the energy management and safety monitoring are still to be tested. There is also a lack of management for the means of transportation frequently used by residents in residential houses: electric cars. The safety and charging management of these vehicles have not yet been intelligentized, which not only adds manpower management, but also makes it difficult to take measures to optimize energy. This is undoubtedly a waste of resources, which not only affects the living experience of residents, but also adds to the burden on administrators.

[0003] Most of the current cloud-based charging management equipment is used for commercial purposes, while electric bicycle charging in residential houses is more inclined to family management equipment. Commercial equipment is undoubtedly a considerable expense for residents who live in residential houses for a long time, and commercial equipment is undoubtedly inappropriate for the traditional model of charging a fixed monthly charging fee for residential houses. The interests of landlords and managers must also be considered. Commercial equipment is obviously not suitable for small venues, and the family's smart management system is not suitable for use in more than a dozen families. For such a situation, energy management for more than a dozen families is also crucial. Energy consumption optimization can not only increase the residents' life experience but also reduce the landlord's additional expenses.

[0004] The present invention discloses an electric travel cloud management and energy efficiency optimization system, in which a master control user and an ordinary user upload data in a terminal device for transmission and store the data in the cloud. A user authorized by face recognition can control the on and off of the electric bicycle socket in the terminal device, and can also prevent unauthorized users from using it. External users charge the bicycle by paying a fee. The cloud data big model can also provide users with off-peak charging services based on the data uploaded by users and the real-time remaining charging positions, and can also optimize the energy consumption of the equipment and adjust the appropriate electricity price for landlords and managers based on the energy consumption of this month. This not only greatly improves the convenience of charging electric vehicles in residential buildings, but also improves the reliability and intelligence of the equipment, and has the characteristics of user interactivity and real-time updating. Summary of the invention

[0005] In view of the defects existing in the prior art, the purpose of the present invention is to propose an electric travel cloud management and energy efficiency optimization system, which changes the cumbersome operation of traditional manual management of charging automatic bicycles, increases the convenience of residents, and provides better functional services for landlords and managers, bringing a better experience and intelligence to the living environment of the entire residential building.

[0006] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0007] An electric travel cloud management and energy efficiency optimization system comprises: a terminal device, a cloud, an execution device, and a plurality of relays and sockets. The terminal device, the cloud and the execution device all communicate via the MQTT communication protocol. The terminal device comprises a master control user terminal (a1), a common user terminal (a2) and an external user terminal (a4). The terminal device can communicate with the cloud via a user upload module (a3). The cloud comprises a data management module (b1) and a data big model (b2), wherein the data big model (b2) comprises a face comparison module (b201), a user power consumption analysis module (b202), and a user power consumption analysis module (b203). b202), intelligent peak-shifting charging module (b203), intelligent energy consumption optimization and electricity price adjustment module (b204), the execution device includes a Wi-Fi module (c1), a single-chip computer module (c2), an alarm module (c3), an electric energy meter module (c4), and a battery detection module (c5). The battery detection module (c5) includes a plurality of current sensors connected in series with the socket, and a plurality of voltage sensors connected in parallel with the socket. The execution device controls the socket through a relay and communicates with the cloud through the Wi-Fi module. The relay is controlled by the execution device and is responsible for controlling the on and off of multiple sockets.

[0008] Furthermore, the alarm module (c3) includes a temperature sensor (e1) for detecting the ambient temperature; a humidity sensor (e2), a light sensor (e3), a smoke sensor (e4), a button (e5), and a buzzer (e6); the temperature sensor (e1), the humidity sensor (e2), the light sensor (e3), and the smoke sensor (e4) are all connected to the control system.

[0009] Furthermore, the face comparison module (b201) of the data big model (b2) in the cloud includes the following steps:

[0010] Step A1: Upload facial data to the cloud server;

[0011] Step A2: Receive and test the integrity and correctness of the uploaded data, including checking the data format, image quality, resolution and integrity;

[0012] Step A3: Preprocess the face data, including image denoising, standardization and alignment operations. The specific methods include:

[0013] Denoising: Use Gaussian filtering to remove noise on the face. The model obtains the filtered image according to the formula algorithm I filtered :

[0014]

[0015] G(i,j) is the Gaussian kernel, I(i,j) is the face image

[0016] Normalization: Normalize the image pixel values ​​to the interval [0,1];

[0017] Alignment: Detect key points of the face: eyes, nose, mouth, and center the face;

[0018] Step A4: Extract the facial feature vector from the preprocessed facial image, and use the neural convolutional network deep learning model to map the filtered facial image to the feature space to obtain the feature vector F:

[0019] F=e[I filtered (x,y)]

[0020] e[I filtered (x,y)] represents the matrix vector that transforms the image into space in the model

[0021] Step A5: Input the extracted feature vector into the pre-trained deep learning model and match it with the face data in the database. Use cosine similarity and Euclidean distance to measure the similarity of the feature vectors:

[0022] Cosine similarity algorithm formula:

[0023]

[0024] Euclidean distance algorithm formula:

[0025]

[0026] [F] is the matrix of the face image in space, F 1 is the vector in the image, F 2 is the database vector, n is the dimension of the matrix;

[0027] Step A6: Determine whether the matching result exceeds a preset threshold;

[0028] Set the threshold γ = 0.8, set the threshold Threshold = 0.7,

[0029] When d(F 1 ,F 2 )≤Thershold,Similarity cos If ≥γ, the judgment result is yes, otherwise it is no;

[0030] If the result of step A6 is no, then execute step A7;

[0031] If the result of step A6 is yes, the match is successful;

[0032] Step A7: Re-upload the facial data and perform comparison again.

[0033] Furthermore, the electric energy meter module (c4) in the execution device comprises the following steps:

[0034] System initialization: start the power monitoring process;

[0035] Periodic detection: Determine whether the user has reached the 30-day monitoring period. If so, reset the energy value.

[0036] Status detection: determine whether the energy meter module and Wi-Fi module are operating normally. If normal, accumulate the energy.

[0037] Threshold comparison: compare the current cycle accumulated power with the preset maximum power threshold E(max);

[0038] Data transmission: If the power exceeds the threshold, data is transmitted through the Wi-Fi module, and the general user client is notified to display the message "this month's charging usage is abnormal", and the master user client displays the abnormal information of the specific user. If the power does not exceed the threshold, the data is only transmitted to the cloud server through the Wi-Fi module;

[0039] Data storage: The power data is stored in the cloud server for subsequent query and analysis.

[0040] Furthermore, the user uploading (a3) ​​of the terminal device includes the following steps:

[0041] APP startup and user identity authentication: Start the APP and prompt the user to register or log in, check whether the network status is stable, if stable, continue, otherwise restart the APP;

[0042] User type judgment and data storage: judge whether the current user is a master user. If so, store the data in the master user database; if so, store the data in the common user database.

[0043] First-time use check: Determine whether it is the first time to use the APP. If so, upload the face and password data;

[0044] Face matching and password verification: If you are a master user, you can enter the master function after passing the verification. If you fail, you need to re-verify. If you are a common user, you can enter the common user function after passing the verification. If you fail, you need to re-verify.

[0045] Permission control: After the master user passes the verification, he can modify the ordinary user database or open the corresponding socket. After the ordinary user passes the verification, he can open the corresponding socket.

[0046] Furthermore, the master control client (a1) and the common client (a2) of the terminal device send instructions to the execution device through the cloud, including the following steps:

[0047] System initialization: Start the system and check whether each module is operating normally. If not, reinitialize;

[0048] Command reception: The system determines whether there is a command sent from the cloud. If so, it receives the command and sends it to the target terminal;

[0049] Instruction analysis:

[0050] If the instruction is to query the status of the charging position, the system will query the number of currently powered sockets a and compare it with the total number of sockets m; if there are vacant charging positions, the status data will be sent to the cloud through the Wi-Fi module, and a prompt "remaining ma charging positions" will be displayed; if there are no vacant charging positions, the status data will be sent to the cloud through the Wi-Fi module, and a prompt "no vacant charging positions" will be displayed;

[0051] If the command is to turn on the socket, the system controls the relay to make the socket powered on, and sends the power-on status to the cloud through the Wi-Fi module, and displays "the socket is powered on" on the terminal;

[0052] If the instruction is to turn off the socket, the system controls the relay to put the socket in a power-off state, and sends the power-off status to the cloud through the Wi-Fi module, and displays "The socket is powered off" on the terminal.

[0053] Furthermore, the operation process of the battery detection module (c5) in the terminal device includes the following steps:

[0054] System initialization and sensor detection: The system initializes and detects the normal operation status of the current / voltage sensor. If it is abnormal, it will be reinitialized;

[0055] Socket power status check: Check whether the socket is powered on. If not, send data to the cloud via the Wi-Fi module and display the prompt "Please connect the plug" on the terminal;

[0056] Current / voltage data collection and vehicle identification: The system collects current / voltage data and determines the specifications of the electric bicycle based on the data;

[0057] Current / voltage anomaly detection: Determine whether the current and voltage are normal based on the preset threshold. If the current and voltage are normal, send data to the cloud through the Wi-Fi module and display "Waiting for the electric bicycle to be charged well" on the terminal. If the current or voltage is abnormal, continue to determine the abnormality type;

[0058] Abnormal type judgment and data sending: If the power supply voltage Vcc exceeds the set range, data is sent to the cloud through the Wi-Fi module, and the corresponding voltage abnormality prompt (too low or too high) is displayed on the terminal. If the power supply current i exceeds the set range, data is sent to the cloud through the Wi-Fi module, and the corresponding current abnormality prompt (too low or too high) is displayed on the terminal.

[0059] Furthermore, the operation of the alarm module (c3) in the execution device includes the following steps:

[0060] System initialization and sensor detection: The system initializes and detects the normal operating status of temperature, humidity, light, and smoke sensors. If it is abnormal, it will be reinitialized; data monitoring and threshold judgment: Continuously monitor the data of each sensor and determine whether it exceeds the preset safety threshold; if the data exceeds the threshold, the alarm process is triggered; Alarm process: Trigger all relays to enter the normally open state, and start the alarm to send an alarm signal and transmit the sensor data to the cloud through the Wi-Fi module; Manual shutdown and prompt information: Determine whether there is a manual button to trigger the alarm to be turned off. If so, turn off the alarm and display the prompt message "There is a suspected major safety hazard, please resolve it as soon as possible" on the terminal device.

[0061] Furthermore, the operation of the electric energy linkage algorithm function module of the intelligent energy consumption optimization and electricity price adjustment (b204) in the cloud data big model (b2) includes the following steps:

[0062] Step G1: Collect the user's real-time electricity consumption data through the energy meter module (c4), including the power consumption P in each time period time , total power consumption per day P day ;

[0063] Step G2: Collect the dynamic electricity price data C of the current month month , market supply and demand and environmental related dataE month ;

[0064] Step G3: pre-process the collected user power consumption data and market environment data. The daily power consumption is calculated using the matrix A = [P day C month E month ]express;

[0065] Normalization of data:

[0066] Where P min and P max The minimum and maximum power consumption in the corresponding time period of this month;

[0067] Step G4: Based on the preprocessed data, optimize the energy consumption of the user's electricity consumption, with the goal of minimizing the daily electricity cost and peak load. The optimization algorithm formula is:

[0068]

[0069] min Power = max t∈T P time

[0070] Where min C and min Power are the minimum daily electricity cost and peak load, T = 24, Indicates that the accumulation is performed every hour;

[0071] Step G5: Adjust the electricity price according to the user's electricity consumption, market environment and electricity price changes. The algorithm formula is as follows:

[0072]

[0073] a and b are electricity price sensitive parameters, which can be adjusted according to different regions. It indicates the recommended monthly electricity price for residential buildings;

[0074] Step G6: extracting feature data for model training according to the requirements of the optimization model;

[0075] Step G7: Monitor the energy consumption of the user in real time, generate energy consumption feedback data, and feed the monitoring data back to the optimization model;

[0076] Step G8: Based on the collected feature data and user energy consumption monitoring data, a large model is trained to optimize the energy consumption strategy so that the energy consumption model can be adjusted according to the dynamic electricity price and the model prediction accuracy is improved by minimizing the loss function:

[0077]

[0078] Where P time.pred and P time.actual Indicates the power consumption predicted by the model and the actual power consumption in the current time period;

[0079] Step G9: Adjust model training parameters or strategies based on user feedback data;

[0080] Step G10: The large model provides an optimized energy consumption solution and electricity price adjustment suggestions

[0081] Step G11: Execute the operation suggestions given by the optimization model, adjust the user's energy consumption mode or electricity price, and finally the electricity consumption allocation algorithm formula is:

[0082]

[0083] The threshold θ can be set according to different regions. Prices above 20% of θ are considered high prices, and prices below 20% of θ are considered low prices. and the value of θ.

[0084] Furthermore, the operation of the algorithm function module of the intelligent peak-shifting charging (b203) in the cloud data big model (b2) includes the following steps:

[0085] Step H1: Collect the user's charging related data from the cloud database through the single chip microcomputer, including the power-on and end time of each charging position and the current usage of the charging position E t (indicates the number of charging positions used in time period t);

[0086] Step H2: Collect the charging time data of users after get off work in the cloud database, including the charging time data T of different electric bicycles of different users off ;

[0087] Step H3: Preprocess the collected data, including cleaning, normalization or other data processing operations. Data normalization: Where E min and E max is the minimum and maximum number of used charging positions in the corresponding time period t;

[0088] Step H4: According to the requirements of the optimization model, extract the feature data required for model training, including the user's historical charging period usage U t and average charging time T charge ;

[0089] Step H5: Real-time monitoring of the number of charging positions in use E t And the current charging position occupancy is fed back to the model. The detection algorithm formula is: (N is the total number of charging positions, H t To detect occupancy in real time);

[0090] Step H6: Perform large model training based on historical data and real-time data to optimize the charging strategy to ensure reasonable allocation during peak hours. The allocation is based on the Lagrange multiplier algorithm formula:

[0091]

[0092] Where T wait,i It is the waiting time of the i-th user, P t It is expressed as the average power in a certain period of time, PeakLoad represents the maximum power in a certain period of time, and ε is a constant between the allowable excess load;

[0093] Step H7: Use the large model to give prediction opinions and optimization suggestions for user charging behavior, and use the real-time charging power P transmitted by the energy meter module (c4) charge and the average charging time T used by users off The remaining charging demand Q remaining Calculate the remaining charging time T remaining , the algorithm formula is as follows:

[0094]

[0095] Step H8: Calculate the remaining charging time according to the user;

[0096] Step H9: Execute the operation suggestions given by the large model and optimize the charging behavior.

[0097] Compared with the prior art, the advantages of the present invention are:

[0098] The electric travel cloud management and energy efficiency optimization system of the present invention and the designed face / password recognition system are particularly suitable for charging management in residential houses, which saves human resources to manage the use of electricity by outsiders, protects the interests of landlords, prevents outside electric bicycles from occupying the electric bicycle charging places in residential houses, and improves the living experience of ordinary residents.

[0099] The electric travel cloud management and energy efficiency optimization system of the present invention is designed with an electric energy meter module for terminal equipment, which can detect the electricity consumption of ordinary users. If it exceeds a preset value, it will prompt the main control user that the electricity consumption is abnormal, which can prevent the tenants of the residential building from frequently lending charging positions to outsiders. For residential buildings that charge a fixed charging fee, it can effectively protect the interests of the landlord.

[0100] An electric travel cloud management and energy efficiency optimization system of the present invention is designed with a battery detection module for the execution device, which determines the specifications of the electric bicycle according to the power, current and voltage measured by the power supply circuit, and then gives whether the battery is abnormal according to the situation, and can remind the residents of the battery condition, thereby increasing the residents' living experience and strengthening the safety of residential buildings.

[0101] An electric travel cloud management and energy efficiency optimization system of the present invention has a peak-shifting charging function performed by a designed cloud data big model. By collecting data on users' commuting time, it calculates the remaining charging time of electric bicycles and the real-time remaining charging positions. It can reasonably arrange the time according to the different homecoming times of different tenants of residential houses, and intelligently manage unnecessary congestion and charging queues. This function not only saves the time cost of residents, but also reduces the management cost of landlords, and increases the intelligence of residential houses.

[0102] An electric travel cloud management and energy efficiency optimization system of the present invention has the functions of intelligent energy consumption optimization and electricity price adjustment performed by the designed cloud data big model. The charging energy consumption of all sockets is obtained through the user data collected by the electric energy meter module. According to the scheduling optimization algorithm of the big model and the electricity consumption in different time periods, unnecessary electricity consumption is reduced and electricity resources are arranged reasonably. This is not only an energy-saving function, but also improves the interests of the landlord. The big model can also give the landlord advice on whether to adjust the electricity price based on the electricity consumption of the month, the market electricity price, and environmental factors, to prevent the electricity price from being too high or too low, thereby improving the living experience of the residents and helping to solve the landlord's worries about pricing. BRIEF DESCRIPTION OF THE DRAWINGS

[0103] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0104] Figure 1 It is an overall framework diagram of an electric mobility cloud management and energy efficiency optimization system of the present invention;

[0105] Figure 2 It is a schematic diagram of the communication method between the terminal device, the cloud and the execution device of the present invention;

[0106] Figure 3 It is a schematic diagram of the internal hardware of the alarm module of the present invention;

[0107] Figure 4 It is a face comparison flow chart of the cloud big data model of the present invention;

[0108] Figure 5 It is a flow chart of the operation of the electric energy meter module in the execution device of the present invention;

[0109] Figure 6 It is a flow chart of the terminal device user of the present invention uploading to the cloud database and controlling the power on and off of the socket;

[0110] Figure 7 It is a flow chart of the cloud sending instructions to the execution device for operation of the present invention;

[0111] Figure 8 It is a flow chart of the operation of the battery detection module in the execution device of the present invention;

[0112] Fig. 9 It is a flowchart of the operation of the alarm module in the execution device of the present invention;

[0113] Fig.10 It is an operation flow chart of the intelligent energy consumption optimization and electricity price adjustment functions in the cloud data big model of the present invention;

[0114] Fig.11 It is an operation flow chart of the intelligent peak-shifting charging function in the cloud data big model of the present invention;

[0115] Fig.12 It is a schematic diagram of the internal hardware of the battery detection module in the execution device of the present invention;

[0116] Fig.13 It is a schematic diagram of the intelligent electric bicycle charging management system of the present invention; DETAILED DESCRIPTION

[0117] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present application. Figures 1 to 13 Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0118] The configuration is performed according to actual use requirements and is not specifically limited in this application. The embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings.

[0119] like Figure 1 , Figure 2 , Figure 3 , Fig.12As shown, an electric travel cloud management and energy efficiency optimization system includes: a terminal device, a cloud, an execution device, and a plurality of relays and sockets. The terminal device, the cloud and the execution device all communicate through the MQTT communication protocol. The terminal device includes a main control user end (a1), a common user end (a2) and an external user end (a4). The terminal device can communicate with the cloud through a user upload module (a3); the cloud includes a data management module (b1) and a data big model (b2), wherein the data big model (b2) includes a face comparison module (b201) for user identity authentication, a user power consumption analysis module (b202) for analyzing the user's power consumption behavior, and an intelligent peak-shifting charging module (b203) for The charging time is intelligently arranged outside the peak power period, and the intelligent energy consumption optimization and electricity price adjustment module (b204) is used to dynamically adjust the charging plan according to the real-time electricity price and user demand. The execution device includes a Wi-Fi module (c1), a single-chip computer module (c2), an alarm module (c3), an electric energy meter module (c4), and a battery detection module (c5). The battery detection module (c5) includes a plurality of current sensors connected in series with the socket, and a plurality of voltage sensors connected in parallel with the socket. The execution device controls the socket through a relay, collects electric energy data and battery status at the same time, and communicates with the cloud through the Wi-Fi module; the relay is controlled by the execution device and is responsible for controlling the on and off of multiple sockets, so as to manage the charging of the electric bicycle.

[0120] like Figure 2 As shown, the main control user module (a1) in the terminal device is connected to the MQTT server (d3) through the MQTT client (d1) for sending control instructions; the ordinary user module (a2) is connected to the MQTT server (d3) through the MQTT client (d2) for user interaction with the system; the MQTT server (d3) in the cloud is used to manage and coordinate message publishing and subscription between the terminal device and the execution device; the Wi-Fi module (c1) in the execution device is connected to the MQTT server (d3) through the MQTT client (d4) for receiving instructions from the cloud and managing and controlling the charging process of the execution device.

[0121] like Figure 3As shown, the alarm module (c3) includes a temperature sensor (e1) for detecting the ambient temperature; a humidity sensor (e2) for detecting the ambient humidity; a light sensor (e3) for detecting the ambient light intensity; a smoke sensor (e4) for detecting the smoke concentration in the environment; a button (e5) for turning off the alarm; and a buzzer (e6) for sounding an alarm to remind the user. The temperature sensor (e1), the humidity sensor (e2), the light sensor (e3) and the smoke sensor (e4) are all connected to the control system. When any sensor detects an abnormality, the control system sounds an alarm through the buzzer (e6). The button (e5) can be used to manually turn off the alarm function of the buzzer (e6) when an abnormality is detected.

[0122] like Figure 4 As shown, the face comparison module (b201) of the data big model (b2) in the cloud includes the following steps:

[0123] Step A1: Upload facial data to the cloud server;

[0124] Step A2: Receive and test the integrity and correctness of the uploaded data, including checking the data format, image quality, resolution and integrity;

[0125] Step A3: Preprocess the face data, including image denoising, standardization and alignment operations. The specific methods include:

[0126] Denoising: Use Gaussian filtering to remove noise on the face. The model obtains the filtered image according to the formula algorithm I filtered :

[0127]

[0128] G(I,j) is the Gaussian kernel, I(i,j) is the face image

[0129] Normalization: Normalize the image pixel values ​​to the interval [0,1];

[0130] Alignment: Detect key points of the face: eyes, nose, mouth, and center the face;

[0131] Step A4: Extract the facial feature vector from the preprocessed facial image, and use the neural convolutional network deep learning model to map the filtered facial image to the feature space to obtain the feature vector F:

[0132] F=e[I filtered (x,y)]

[0133] e[I filtered (x,y)] represents the matrix vector that transforms the image into space in the model

[0134] Step A5: Input the extracted feature vector into the pre-trained deep learning model and match it with the face data in the database. Use cosine similarity and Euclidean distance to measure the similarity of the feature vectors:

[0135] Cosine similarity algorithm formula:

[0136]

[0137] Euclidean distance algorithm formula:

[0138]

[0139] [F] is the matrix of the face image in space, F 1 is the vector in the image, F 2 is the database vector, n is the dimension of the matrix;

[0140] Step A6: Determine whether the matching result exceeds a preset threshold;

[0141] Set the threshold γ = 0.8, set the threshold Threshold = 0.7,

[0142] When d(F 1 ,F 2 )≤Thershold,Similarity cos If ≥γ, the judgment result is yes, otherwise it is no;

[0143] If the result of step A6 is no, then execute step A7;

[0144] If the result of step A6 is yes, the match is successful;

[0145] Step A7: Re-upload the facial data and compare again;

[0146] The cloud server (d3) is used to store the face database and accurately compare the facial features through the deep learning data model (b2). The face preprocessing step (A3) further includes adjusting the image size, lighting compensation and rotation correction to optimize the extraction of facial features.

[0147] like Figure 5 As shown, the electric energy meter module (c4) in the execution device comprises the following steps:

[0148] Step B1: Initialize the system and start the power monitoring process;

[0149] Step B2: Check whether the user has reached the 30-day monitoring period;

[0150] Step B3: When the 30-day cycle is reached, the electric energy value is reset;

[0151] Step B4: Determine whether the energy meter module and the Wi-Fi module are operating normally. If they are operating abnormally, reinitialize the system;

[0152] Step B5: Accumulate the user's power usage value in the current cycle;

[0153] Step B6: Compare the accumulated power with a preset maximum power threshold E(max) to determine whether it exceeds the threshold;

[0154] Step B7: If the threshold is not exceeded, continue to monitor the power and transmit data via the Wi-Fi module;

[0155] Step B8: If the threshold is exceeded, the abnormal power data is transmitted to the cloud server via the Wi-Fi module;

[0156] Step B9: The ordinary user client displays a prompt message "this month's charging usage is abnormal";

[0157] Step B10: The master control user client displays abnormal information of the specific user, such as "XXX user's charging usage this month is abnormal";

[0158] Step B11: storing relevant power data to a cloud server;

[0159] The step of accumulating power (B5) accumulates all power consumption during the user's charging process to enable more accurate energy monitoring. The abnormal detection result is transmitted to the cloud server in real time via the Wi-Fi module (c1), and abnormal prompt information is displayed on the ordinary user terminal (a2) and the master user terminal (a1). The step of storing data in the cloud server (B11) is used for subsequent user power consumption analysis (b202).

[0160] like Figure 6 As shown, the user uploading (a3) ​​of the terminal device includes the following steps:

[0161] Step C1: Start the APP and prompt the user to register or log in;

[0162] Step C2: Check whether the current network status is stable. If it is stable, proceed to the next step;

[0163] Step C3: Determine whether the current user is the master user;

[0164] Step C4: If it is a master user, the data is stored in the master user database;

[0165] Step C5: If it is a common user, the data is stored in the common user database;

[0166] Step C6: If the user is the primary user, check whether it is the first time to use the APP;

[0167] Step C7: If the user is a normal user, check whether it is the first time to use the APP;

[0168] Step C8: If the user is the primary user and it is the first time to use the system, upload the face and password data;

[0169] Step C9: If the user is a common user and it is the first time to use the device, upload the face and password data;

[0170] Step C10: Perform face comparison on the master user to determine whether the comparison passes. If so, execute step C16;

[0171] Step C11: Perform face comparison on ordinary users to determine whether the comparison passes. If so, execute step C17;

[0172] Step C12: Enter the password of the master user;

[0173] Step C13: Enter the password of the ordinary user;

[0174] Step C14: Perform password authentication on the master user to determine whether the authentication is passed. If passed, execute step C16;

[0175] Step C15: Perform password authentication on the master control user to determine whether the authentication is passed. If passed, execute step C17;

[0176] Step C16: The master user can modify the database of ordinary users or open the corresponding socket;

[0177] Step C17: Ordinary users can turn on the corresponding sockets;

[0178] The system stores data in the master user database or the ordinary user database according to the user type, the master user terminal (a1) and the ordinary user terminal (a2), to ensure the authority management of different users. When the system fails in face matching and password authentication at the same time (steps C10, C11, C14, C15), the socket cannot be used until the verification is passed.

[0179] like Figure 7 As shown, the common user terminal (a2) and the master user terminal (a1) of the terminal device send instructions to the execution device through the cloud, including the following steps:

[0180] Step D1: system initialization, start the system;

[0181] Step D2: Check whether each module in the system is operating normally. If not, perform troubleshooting. If normal, continue;

[0182] Step D3: Check whether there is a command sent from the cloud, if there is no command, wait, if there is, continue;

[0183] Step D4: receiving cloud instructions and sending the instructions to the target terminal;

[0184] Step D5: determining whether the received instruction is to open the socket, close the socket, and query whether there are remaining charging positions;

[0185] Step D6: Query the number of currently powered sockets a;

[0186] Step D7: Compare the number of powered sockets a with the total number of sockets m to determine whether m is greater than a;

[0187] Step D8: If it is a socket-opening instruction, the microcontroller controls the relay to be in a normally closed state, and the socket is in a powered-on state;

[0188] Step D9: If it is a socket closing instruction, the microcontroller controls the relay to be in a normally open state, and the socket is in a power-off state;

[0189] Step D10: When the condition m>a is not met, the status data is sent to the cloud via the Wi-Fi module;

[0190] Step D11: When the condition m>a is met, the status data is sent to the cloud via the Wi-Fi module;

[0191] Step D12: When the socket is powered on normally, the system sends the power-on status data to the cloud via the Wi-Fi module;

[0192] Step D13: When the socket is powered off, the system sends the power-off status data to the cloud via the Wi-Fi module;

[0193] Step D14: the system outputs a prompt message "No vacant charging position" on the terminal device;

[0194] Step D15: The system outputs a prompt message "ma charging positions remaining" on the terminal device;

[0195] Step D16: the system outputs a prompt message "the socket is powered on" on the terminal device;

[0196] Step D17: the system outputs a prompt message "the socket is powered on" on the terminal device;

[0197] The system dynamically controls the power-on status of the sockets by comparing the current number a of powered-on sockets with the total number m of sockets, thereby ensuring reasonable allocation of power resources.

[0198] like Figure 8 As shown, the operation process of the battery detection module (c5) in the terminal device includes the following steps:

[0199] Step E1: System initialization, start the system;

[0200] Step E2: Check whether the current / voltage sensor works normally. If not, perform fault processing. If normal, continue;

[0201] Step E3: Detect whether the socket is powered on;

[0202] Step E4: If there is no power, the data is sent to the cloud via the Wi-Fi module;

[0203] Step E5: "Please connect the plug" is displayed on the terminal device;

[0204] Step E6: The system collects data from the current / voltage sensor;

[0205] Step E7: judging what specification the electric bicycle belongs to according to the collected current / voltage data;

[0206] Step E8: judging whether the current current and voltage are abnormal according to a preset current or voltage threshold;

[0207] Step E9: If the current and voltage are normal, the data is sent to the cloud via the Wi-Fi module;

[0208] Step E10: Displaying “waiting for the electric bicycle to be in good charging state” on the terminal device;

[0209] Step E11: If the voltage is abnormal, determine whether the current supply voltage Vcc of the detection circuit is less than the minimum threshold a or greater than the maximum threshold b;

[0210] Step E12: If the current is abnormal, determine whether the current supply current i of the current detection circuit is less than the minimum threshold c or greater than the maximum threshold d;

[0211] Step E13: If the current voltage Vcc is less than the minimum threshold a, data is sent to the cloud via the Wi-Fi module;

[0212] Step E14: If the current voltage Vcc is greater than the maximum threshold b, data is sent to the cloud via the Wi-Fi module;

[0213] Step E15: When the current i is less than the threshold c, the data is sent to the cloud via the Wi-Fi module;

[0214] Step E16: When the current i is greater than the threshold d, the data is sent to the cloud via the Wi-Fi module;

[0215] Step E17: "Your battery charging voltage is too low. It is recommended that you replace the battery / charger" is displayed on the terminal device;

[0216] Step E18: "Your battery charging voltage is too high. It is recommended that you replace the battery / charger" is displayed on the terminal device;

[0217] Step E19: "Your battery charging current is too low. It is recommended that you replace the battery / charger" is displayed on the terminal device;

[0218] Step E20: "The charging current of your battery is too high. It is recommended that you replace the battery / charger" is displayed on the terminal device;

[0219] The detection and feedback of the current and voltage are determined by preset thresholds a, b (voltage thresholds) and c, d (current thresholds) to ensure that the charging state of the electric bicycle is within a safe range and provide prompt information according to abnormal conditions.

[0220] like Fig. 9 As shown, the operation of the alarm module (c3) in the execution device includes the following steps:

[0221] Step F1: System initialization, start the system;

[0222] Step F2: Check whether the temperature, humidity, light and smoke sensors are operating normally. If not, reinitialize the system. If normal, continue;

[0223] Step F3: Determine whether the data detected by the temperature, humidity, light, and smoke sensors exceed the preset thresholds. If not, continue monitoring. If so, perform the following steps;

[0224] Step F4: trigger all relays to enter the normally open state;

[0225] Step F5: Activate the alarm and send out an alarm signal;

[0226] Step F6: Transmit sensor data to the cloud via the Wi-Fi module;

[0227] Step F7: Determine whether there is a manual key press to trigger the buzzer to be turned off;

[0228] Step F8: A prompt message "major safety hazard suspected, please resolve as soon as possible" is displayed on the terminal device;

[0229] Step F9: When it is detected that the manual key is triggered, the buzzer is turned off if it is triggered, and the buzzer continues to be triggered if it is not triggered;

[0230] The system can monitor the changes in temperature, humidity, light and smoke parameters in real time, and trigger corresponding electrical equipment to enter a linkage state according to sensor feedback data.

[0231] like Fig.10 As shown, the operation of the electric energy linkage algorithm function module of the intelligent energy consumption optimization and electricity price adjustment (b204) in the cloud data big model (b2) includes the following steps:

[0232] Step G1: Collect the user's real-time electricity consumption data through the energy meter module (c4), including the power consumption P in each time period time , total power consumption per day P day ;

[0233] Step G2: Collect the dynamic electricity price data C of the current month month , market supply and demand and environmental related dataE month ;

[0234] Step G3: pre-process the collected user power consumption data and market environment data. The daily power consumption is calculated using the matrix A = [P day C month E month ]express;

[0235] Normalization of data:

[0236] Where P min and P max The minimum and maximum power consumption in the corresponding time period of this month;

[0237] Step G4: Based on the preprocessed data, optimize the energy consumption of the user's electricity consumption, with the goal of minimizing the daily electricity cost and peak load. The optimization algorithm formula is:

[0238]

[0239] min Power = max t∈T P time

[0240] Where min C and min Power are the minimum daily electricity cost and peak load, T = 24, Indicates that the accumulation is performed every hour;

[0241] Step G5: Adjust the electricity price according to the user's electricity consumption, market environment and electricity price changes. The algorithm formula is as follows:

[0242]

[0243] a and b are electricity price sensitive parameters, which can be adjusted according to different regions. It indicates the recommended monthly electricity price for residential buildings;

[0244] Step G6: extracting feature data for model training according to the requirements of the optimization model;

[0245] Step G7: Monitor the energy consumption of the user in real time, generate energy consumption feedback data, and feed the monitoring data back to the optimization model;

[0246] Step G8: Based on the collected feature data and user energy consumption monitoring data, a large model is trained to optimize the energy consumption strategy so that the energy consumption model can be adjusted according to the dynamic electricity price and the model prediction accuracy is improved by minimizing the loss function:

[0247]

[0248] Where P time.pred and P time.actual Represents the power consumption predicted by the model and the actual power consumption in the current time period

[0249] Step G9: Adjust model training parameters or strategies based on user feedback data;

[0250] Step G10: The large model provides an optimized energy consumption solution and electricity price adjustment suggestions

[0251] Step G11: Execute the operation suggestions given by the optimization model, adjust the user's energy consumption mode or electricity price, and finally the electricity consumption allocation algorithm formula is:

[0252]

[0253] The threshold θ can be set according to different regions. Prices above 20% of θ are considered high prices, and prices below 20% of θ are considered low prices. and the numerical value of θ;

[0254] The system can dynamically adjust electricity prices and energy consumption optimization plans based on different markets, environmental data and user electricity consumption behaviors.

[0255] like Fig.11 As shown, the operation of the algorithm function module of the intelligent peak-shifting charging (b203) in the cloud data big model (b2) includes the following steps:

[0256] Step H1: Collect the user's charging related data from the cloud database through the single chip microcomputer, including the power-on and end time of each charging position and the current usage of the charging position E t (indicates the number of charging positions used in time period t);

[0257] Step H2: Collect the charging time data of users after get off work in the cloud database, including the charging time data T of different electric bicycles of different users off ;

[0258] Step H3: Preprocess the collected data, including cleaning, normalization or other data processing operations. Data normalization: Where E min and E max is the minimum and maximum number of used charging positions in the corresponding time period t;

[0259] Step H4: According to the requirements of the optimization model, extract the feature data required for model training, including the user's historical charging period usage U t and average charging time T charge ;

[0260] Step H5: Real-time monitoring of the number of charging positions in use E t And the current charging position occupancy is fed back to the model. The detection algorithm formula is: (N is the total number of charging positions, H t To detect occupancy in real time);

[0261] Step H6: Perform large model training based on historical data and real-time data to optimize the charging strategy to ensure reasonable allocation during peak hours. The allocation is based on the Lagrange multiplier algorithm formula:

[0262]

[0263] Where T wait,i It is the waiting time of the i-th user, P t It is expressed as the average power in a certain period of time, PeakLoad represents the maximum power in a certain period of time, and ε is a constant between the allowable excess load;

[0264] Step H7: Use the large model to give prediction opinions and optimization suggestions for user charging behavior, and use the real-time charging power P transmitted by the energy meter module (c4) charge and the average charging time T used by users off The remaining charging demand Q remaining Calculate the remaining charging time T remaining , the algorithm formula is as follows:

[0265]

[0266] Step H8: Calculate the remaining charging time according to the user;

[0267] Step H9: Execute the operation suggestions given by the large model, optimize the charging behavior, ensure charging during off-peak hours, and reduce power consumption during peak hours;

[0268] The system can dynamically adjust the charging time according to the user's off-get off work time and the real-time monitoring of the charging position, thereby optimizing the user's charging and time costs.

[0269] The above description is only the preferred implementation of the present application, so that those skilled in the art can understand or implement the present application. For various modifications to these embodiments, those skilled in the art can make several improvements or modifications without departing from the principles of the present invention, and these improvements or modifications should also be regarded as the protection scope of the present invention. Therefore, the present application will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features of the present application.

Claims

1. An electric mobility cloud management and energy efficiency optimization system, comprising: The invention relates to a terminal device, a cloud, an execution device, a plurality of relays and sockets, wherein the terminal device, the cloud and the execution device communicate via the MQTT communication protocol, the terminal device comprises a master user terminal (a1), a common user terminal (a2) and an external user terminal (a4), and the terminal device can communicate with the cloud via a user upload module (a3); the cloud comprises a data management module (b1) and a data big model (b2), wherein the data big model (b2) comprises a face comparison module (b201), a user power consumption analysis module (b202), an intelligent The electric power meter can be equipped with a peak-shifting charging module (b203), an intelligent energy consumption optimization and electricity price adjustment module (b204), and the execution device includes a Wi-Fi module (c1), a single-chip computer module (c2), an alarm module (c3), an electric energy meter module (c4), and a battery detection module (c5). The battery detection module (c5) includes a plurality of current sensors connected in series with the socket, and a plurality of voltage sensors connected in parallel with the socket. The execution device controls the socket through a relay and communicates with the cloud through the Wi-Fi module. The relay is controlled by the execution device and is responsible for controlling the on and off of multiple sockets.

2. The electric mobility cloud management and energy efficiency optimization system according to claim 1, characterized in that: The alarm module (c3) comprises a temperature sensor (e1) for detecting the ambient temperature; a humidity sensor (e2), a light sensor (e3), a smoke sensor (e4), a button (e5), and a buzzer (e6); the temperature sensor (e1), the humidity sensor (e2), the light sensor (e3), and the smoke sensor (e4) are all connected to a control system.

3. The electric mobility cloud management and energy efficiency optimization system according to claim 1, characterized in that: The face comparison module (b201) of the data big model (b2) in the cloud comprises the following steps: Step A1: Upload facial data to the cloud server; Step A2: Receive and test the integrity and correctness of the uploaded data, including checking the data format, image quality, resolution and integrity; Step A3: Preprocess the face data, including image denoising, standardization and alignment operations. The specific methods include: Denoising: Use Gaussian filtering to remove noise on the face. The model obtains the filtered image according to the formula algorithm I filtered : G(i, j) is the Gaussian kernel, and I(i, j) is the normalization of the face image: the image pixel values ​​are normalized to the interval [0, 1]; Alignment: Detect key points of the face: eyes, nose, mouth, and center the face; Step A4: Extract the facial feature vector from the preprocessed facial image, and use the neural convolutional network deep learning model to map the filtered facial image to the feature space to obtain the feature vector F: F=e[I filtered (x,y)] e[I filtered (x, y)] represents the matrix vector that transforms the image into space in the model Step A5: Input the extracted feature vector into the pre-trained deep learning model and match it with the face data in the database. Use cosine similarity and Euclidean distance to measure the similarity of the feature vectors: Cosine similarity algorithm formula: Euclidean distance algorithm formula: [F] is the matrix of the face image in space, F1 is the vector in the image, F2 is the database vector, n is the dimension of the matrix; Step A6: Determine whether the matching result exceeds a preset threshold; Set the threshold γ = 0.8, set the threshold Threshold = 0.7, When d(F1, F2)≤Thershold, Similarity cos If ≥γ, the judgment result is yes, otherwise it is no; If the result of step A6 is no, then execute step A7; If the result of step A6 is yes, the match is successful; Step A7: Re-upload the facial data and perform comparison again.

4. The electric mobility cloud management and energy efficiency optimization system according to claim 1, comprising an electric energy meter module (c4) and a Wi-Fi module (c1), characterized in that: The system performs the following steps: System initialization: start the power monitoring process; Periodic detection: Determine whether the user has reached the 30-day monitoring period. If so, reset the energy value. Status detection: determine whether the energy meter module and Wi-Fi module are operating normally. If normal, accumulate the energy. Threshold comparison: compare the current cycle accumulated power with the preset maximum power threshold E(max); Data transmission: If the power exceeds the threshold, data is transmitted through the Wi-Fi module, and the general user client is notified to display the message "this month's charging usage is abnormal". The master user client displays the abnormal information of the specific user. If the power does not exceed the threshold, data is only transmitted to the cloud server through the Wi-Fi module; Data storage: The power data is stored in the cloud server for subsequent query and analysis.

5. The electric mobility cloud management and energy efficiency optimization system according to claim 1, characterized in that: The user upload module (a3) ​​of the terminal device comprises the following steps: APP startup and user identity authentication: Start the APP and prompt the user to register or log in, check whether the network status is stable, if stable, continue, otherwise restart the APP; User type judgment and data storage: judge whether the current user is a master user. If so, store the data in the master user database; if so, store the data in the common user database. First-time use check: Determine whether it is the first time to use the APP. If so, upload the face and password data; Face matching and password verification: If you are a master user, you can enter the master function after passing the verification. If you fail, you need to re-verify. If you are a common user, you can enter the common user function after passing the verification. If you fail, you need to re-verify. Permission control: After the master user passes the verification, he can modify the ordinary user database or open the corresponding socket. After the ordinary user passes the verification, he can open the corresponding socket.

6. An electric mobility cloud management and energy efficiency optimization system as claimed in claim 1, wherein the master control user terminal (a1) and the common user terminal (a2) send instructions to the execution device through the cloud, characterized in that: The system comprises the following steps: System initialization: Start the system and check whether each module is operating normally. If not, reinitialize; Command reception: The system determines whether there is a command sent from the cloud. If so, it receives the command and sends it to the target terminal; Instruction analysis: If the instruction is to query the status of the charging position, the system will query the number of currently powered sockets a and compare it with the total number of sockets m; if there are vacant charging positions, the status data will be sent to the cloud through the Wi-Fi module, and a prompt will be given that "ma charging positions are remaining"; if there are no vacant charging positions, the status data will be sent to the cloud through the Wi-Fi module, and a prompt will be given that "no vacant charging positions are available"; If the command is to turn on the socket, the system controls the relay to power on the socket, and sends the power-on status to the cloud via the Wi-Fi module, while displaying "the socket is powered on" on the terminal; If the instruction is to turn off the socket, the system controls the relay to put the socket in a power-off state, and sends the power-off status to the cloud through the Wi-Fi module, and displays "the socket is powered off" on the terminal.

7. An electric mobility cloud management and energy efficiency optimization system as claimed in claim 1, wherein the operation process of the battery detection module (c5) in the terminal device is characterized in that: The following steps are involved: System initialization and sensor detection: The system initializes and detects the normal operation status of the current / voltage sensor. If it is abnormal, it will be reinitialized; Socket power status check: Check whether the socket is powered on. If not, send data to the cloud via the Wi-Fi module and display the prompt "Please connect the plug" on the terminal; Current / voltage data collection and vehicle identification: The system collects current / voltage data and determines the specifications of the electric bicycle based on the data; Current / voltage anomaly detection: Determine whether the current and voltage are normal based on the preset threshold. If the current and voltage are normal, send data to the cloud through the Wi-Fi module and display "Waiting for the electric bicycle to be charged well" on the terminal. If the current or voltage is abnormal, continue to determine the abnormality type; Abnormal type judgment and data sending: If the power supply voltage Vcc exceeds the set range, data is sent to the cloud through the Wi-Fi module, and the corresponding voltage abnormality prompt (too low or too high) is displayed on the terminal. If the power supply current i exceeds the set range, data is sent to the cloud through the Wi-Fi module, and the corresponding current abnormality prompt (too low or too high) is displayed on the terminal.

8. The electric mobility cloud management and energy efficiency optimization system according to claim 1, wherein: The operation process of the alarm module (c3) in the execution device is characterized by comprising the following steps: System initialization and sensor detection: The system initializes and detects the normal operating status of temperature, humidity, light, and smoke sensors. If it is abnormal, it will be reinitialized; data monitoring and threshold judgment: Continuously monitor the data of each sensor and determine whether it exceeds the preset safety threshold. If the data exceeds the threshold, the alarm process will be triggered; Alarm process: Trigger all relays to enter the normally open state, and start the alarm to send out an alarm signal and transmit the sensor data to the cloud through the Wi-Fi module; Manual shutdown and prompt information: Determine whether there is a manual button to trigger the alarm to be turned off. If so, turn off the alarm and display the prompt message "There is a suspected major safety hazard, please resolve it as soon as possible" on the terminal device.

9. The electric mobility cloud management and energy efficiency optimization system according to claim 1, characterized in that: The operation of the electric energy linkage algorithm function module of the intelligent energy consumption optimization and electricity price adjustment (b204) in the cloud data big model (b2) includes the following steps: Step G1: Collect the user's real-time electricity consumption data through the energy meter module (c4), including the power consumption P in each time period time , total power consumption per day P day ; Step G2: Collect the dynamic electricity price data C of the current month month , market supply and demand and environmental related dataE month ; Step G3: pre-process the collected user power consumption data and market environment data. The daily power consumption is calculated using the matrix A = [P day C month E month ]express; Normalization of data: Where P min and P max The minimum and maximum power consumption in the corresponding time period of this month; Step G4: Based on the preprocessed data, optimize the energy consumption of the user's electricity consumption, with the goal of minimizing the daily electricity cost and peak load. The optimization algorithm formula is: min Power=max t∈T P time Where min C and min Power are the minimum daily electricity cost and peak load, T = 24, Indicates that the accumulation is performed every hour; Step G5: Adjust the electricity price according to the user's electricity consumption, market environment and electricity price changes. The algorithm formula is as follows: a and b are electricity price sensitive parameters, which can be adjusted according to different regions. It indicates the recommended monthly electricity price for residential buildings; Step G6: extracting feature data for model training according to the requirements of the optimization model; Step G7: Monitor the energy consumption of the user in real time, generate energy consumption feedback data, and feed the monitoring data back to the optimization model; Step G8: Based on the collected feature data and user energy consumption monitoring data, a large model is trained to optimize the energy consumption strategy so that the energy consumption model can be adjusted according to the dynamic electricity price and the model prediction accuracy is improved by minimizing the loss function: Where P time.pred and P time.actual Indicates the power consumption predicted by the model and the actual power consumption in the current time period; Step G9: Adjust model training parameters or strategies based on user feedback data; Step G10: The large model provides an optimized energy consumption solution and electricity price adjustment suggestions Step G11: Execute the operation suggestions given by the optimization model, adjust the user's energy consumption mode or electricity price, and finally the electricity consumption allocation algorithm formula is: The threshold θ can be set according to different regions. Prices above 20% of θ are considered high prices, and prices below 20% of θ are considered low prices. and the value of θ.

10. The electric mobility cloud management and energy efficiency optimization system according to claim 1, characterized in that: The operation of the algorithm function module of the intelligent peak-shifting charging (b203) in the cloud data big model (b2) includes the following steps: Step H1: Collect the user's charging related data from the cloud database through the single chip microcomputer, including the power-on and end time of each charging position and the current usage of the charging position E t (indicates the number of charging positions used in time period t); Step H2: Collect the charging time data of users after get off work in the cloud database, including the charging time data T of different electric bicycles of different users off ; Step H3: Preprocess the collected data, including cleaning, normalization or other data processing operations. Data normalization: Where E min and E max is the minimum and maximum number of used charging positions in the corresponding time period t; Step H4: According to the requirements of the optimization model, extract the feature data required for model training, including the user's historical charging period usage U t and average charging time T charge ; Step H5: Real-time monitoring of the number of charging positions in use E t And the current charging position occupancy is fed back to the model. The detection algorithm formula is: (N is the total number of charging positions, H t is the real-time detection occupancy rate); Step H6: Perform large model training based on historical data and real-time data to optimize the charging strategy to ensure reasonable allocation during peak hours. The allocation is based on the Lagrange multiplier algorithm formula: Where T wait，i It is the waiting time of the i-th user, P t It is expressed as the average power in a certain period of time, PeakLoad represents the maximum power in a certain period of time, and ε is a constant between the allowable excess load; Step H7: Use the large model to give prediction opinions and optimization suggestions for user charging behavior, and use the real-time charging power P transmitted by the energy meter module (c4) charge and the average charging time T used by users off The remaining charging demand Q remaining Calculate the remaining charging time T remaining , the algorithm formula is as follows: Step H8: Calculate the remaining charging time according to the user; Step H9: Execute the operation suggestions given by the large model and optimize the charging behavior.

Images