A three-in-one multifunctional emergency power management system

Through the design of the three-in-one multi-function emergency power management system, the GAN model and RBF neural network are used to predict and regulate temperature, which solves the problems of insufficient reaction time and large losses in emergency power management, and achieves more accurate status monitoring and prediction , improves battery performance and safety.

CN118074315BActive Publication Date: 2024-11-26NANJING LINGDA ELECTRICAL
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202311809708.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-12-26
Publication Date
2024-11-26
Estimated Expiration
2043-12-26

AI Technical Summary

Technical Problem

The existing emergency power management system has insufficient power supply working status prediction and pre-reaction, resulting in insufficient reaction time and excessive loss, which cannot effectively avoid safety accidents caused by rising temperatures.

Method used

A three-in-one multi-function emergency power management system is designed, including the input layer, management layer and display layer. The battery temperature is monitored in real time through the monitoring module, the temperature prediction is predicted using the GAN model and the RBF neural network, and the corresponding regulation is carried out through the control unit. , including wireless control of the cooling unit, warning or power off in advance to avoid bad conditions.

Benefits of technology

It realizes more accurate status monitoring and prediction of emergency power supplies, reduces battery losses, improves battery performance and service life, promptly detects faults and hidden dangers, and enhances the safety of power management.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118074315B_ABST
    Figure CN118074315B_ABST
Patent Text Reader

Abstract

The invention discloses a three-in-one multifunctional emergency power supply management system, which relates to the field of power supply management. The technical scheme thereof includes: the management layer includes a power supply management module and a monitoring module, the monitoring module monitors the operating environment of the system, and performs data monitoring and collection during the working process of the rectiverter emergency power supply within the current cycle time, the power supply management module includes a judgment unit, a prediction unit and a control unit, the judgment unit is used to judge the temperature safety state of the rectiverter emergency power supply within the current cycle time and the next cycle time, the prediction unit expands the samples within the current cycle time through a GAN model, and predicts the temperature of the rectiverter emergency power supply within the next cycle time through an RBF neural network model, and the control unit performs corresponding work according to the output of the judgment unit, so as to realize the temperature state monitoring and safety protection of the three-in-one multifunctional emergency power supply.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of power management, and more particularly to a three-in-one multifunctional emergency power management system. Background Art

[0002] As electricity has become an indispensable part of life, under special conditions such as large-scale power outages, for some emergency power needs, emergency power supply, as an independent third emergency power supply, can help avoid the situation where emergency events are stopped due to power outages and help events or work continue to move forward.

[0003] During the use of emergency power supply, due to the urgent use of power supply, especially for some areas with extremely high electricity demand, the operating temperature of the emergency power supply will rise rapidly during the process of uninterrupted and high-efficiency power generation. Poor operating temperature of the emergency battery will not only affect the performance of the battery and accelerate the battery degradation, but also cause safety accidents. In the real-time processing of power supply, there will be less reaction time, resulting in untimely response. Therefore, in this case, advance warning of faults will increase the reaction time and effectively avoid accidents.

[0004] The invention disclosed in the publication number CN117134326A discloses a power management system that can improve the reliability of remote power supply to slave devices. The power management system includes a master device, an optoelectronic splitter, a first power supply, and a second power supply; the first power supply is connected to the master device, and the master device is connected to the optoelectronic splitter; the second power supply is connected to the optoelectronic splitter, and the optoelectronic splitter is connected to at least one slave device; the optoelectronic splitter is used to control the first power supply to supply power to the at least one slave device when it is determined that the power supply of the second power supply is abnormal. This invention is only designed for the technical problem of power supply to slave devices when the power supply is abnormal, but does not involve the specific power supply working state prediction and pre-response, so the present invention is innovatively designed for the above-mentioned problems. Summary of the invention

[0005] In view of the problems of insufficient pre-reaction time and excessive loss in the power management process in the prior art, the purpose of the present invention is to provide a three-in-one multifunctional emergency power management system to achieve safety in the power management process.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A three-in-one multifunctional emergency power management system, comprising an input layer, a management layer and a display layer:

[0008] The input layer controls the AC and DC output of the energy storage device through the adjustment module. The energy storage device is also called the Rectivoerter HE module, which is used for the energy integration of the Rectiverter emergency power supply. The adjustment module enables the energy storage device to communicate with the controller through the CAN bus network.

[0009] The management layer includes a power management module and a monitoring module. The monitoring module monitors the operating environment of the system and performs data monitoring and collection during the working process of the Rectiverter emergency power supply during the current cycle time. The power management module includes a judgment unit, a prediction unit and a control unit. The judgment unit is used to judge the temperature safety status of the Rectiverter emergency power supply during the current cycle time and the next cycle time. The prediction unit expands the samples within the current cycle time through the GAN model and predicts the temperature of the Rectiverter emergency power supply during the next cycle time through the RBF neural network model. The control unit performs corresponding work according to the output of the judgment unit;

[0010] The display layer is used for the end page display of data information and provides a system operation interface for operation and maintenance personnel.

[0011] Preferably, the adjustment module includes a bidirectional DC-DC converter and a DC-AC converter, the energy storage device is composed of one or more lithium-ion batteries, the AC power is input to the energy storage device, and after the DC and AC conversion of the DC-AC converter and the voltage conversion of the DC-DC converter, 32A AC and 100A DC are output.

[0012] Preferably, the display layer includes a human-computer interaction interface, and the human-computer interaction module provides identity authentication for user login and functionally adjusts relevant parameters according to specific needs. The human-computer interaction module is connected to the display device of the monitoring module to monitor and evaluate the operating status of the system.

[0013] Preferably, the monitoring module includes a Smartpack-2 Touch monitor, the Smartpack-2 Touch monitor includes an indication area and a connection area, the Smartpack-2 Touch monitor is connected to a local area network via Ethernet, and the indication area displays the working status by indicating the color of an LED light.

[0014] Preferably, the monitoring module monitors the temperature status of the battery in real time through the Smartpack-2 Touch monitor, sets the monitoring module to perform continuous periodic measurements, and acquires data of the Rectiverter emergency power supply multiple times within the current cycle time, thereby obtaining n real-time temperature values, and the judgment unit averages the real-time temperature values ​​to form a temperature representative value within the cycle time T, and judges whether the temperature representative value is within the value range to which it belongs.

[0015] If the temperature representative value is within the first value interval, a safety signal is output and the control unit does not work;

[0016] If the temperature representative value is within the second value interval, a first-level alarm is output, and the control unit controls the prediction unit to start;

[0017] If the temperature representative value is within the third value interval, a second-level alarm is output and the control unit turns off the emergency power supply of the Rectiverter.

[0018] Preferably, the prediction unit is used to predict the temperature safety state of the Rectiverter emergency power supply within the next cycle time, and the prediction unit is started when the output is a minor alarm within the current cycle time. The specific working process of the prediction unit includes the following working process:

[0019] Step S61: Obtain the real-time temperature value and measurement time point corresponding to the secondary warning within the current cycle time, and form an import set D λ ={(T1,t1),......,(T a ,t a ),......,(T n ,t n )};

[0020] Step S62: Import the imported set as a sample into the trained GAN model for sample expansion, and set the output set formed by the expansion as S θ ={(T1,t1),......,(T i ,t i ),......,(T m ,t m )}, and m>n. Where i and a represent the labels of the sub-values ​​in the output set and the import set, and m and n represent the number of sub-values ​​in the output set and the import set.

[0021] Preferably, each measurement time point in the output set is used as an input of the RBF neural network model, and the real-time temperature value corresponding to each measurement time point is used as an output, the RBF neural network model is trained, the weight vector is updated, and then in the next cycle time, n measurement time points are randomly selected and input into the trained RBF neural network model to generate the real-time temperature values ​​corresponding to the n measurement time points;

[0022] The judgment unit judges the real-time temperature values ​​corresponding to the n measurement time points within the next cycle time:

[0023] If the output is a safety signal, continue to work normally;

[0024] If the output is a first-level alarm, the control unit controls the heat dissipation unit for regulation;

[0025] If the output is a level 2 alarm, a pre-warning is given or the Rectiverter emergency power is shut down.

[0026] Preferably, the control unit includes a heat dissipation unit, which wirelessly controls a heat dissipation device, and the heat dissipation device includes a front fan. The heat dissipation unit controls the output power of the front fan to increase and decrease the airflow from front to back, and also monitors whether there are any obstacles in the air duct to remind operation and maintenance personnel to remove the obstacles.

[0027] Preferably, the management layer is applied to a microcomputer, the human-computer interaction interface includes multiple clients, the clients are connected via wireless or wired communication equipment, the communication connection follows the HTTP protocol, and an integrated information processing system and a multi-client information architecture are adopted.

[0028] Preferably, the three-in-one multifunctional emergency power supply management system is applied to a cloud processing platform, applied to the cloud, and performs distributed storage.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] 1. In the present invention, the temperature status of the Rectiverter emergency power supply is monitored in real time by the monitoring module, so that more accurate battery status information can be obtained, and reliable data support is provided for the working status analysis of the Rectiverter emergency power supply. The data transmitted by the monitoring module is processed by the judgment module, and the judgment information of the Rectiverter emergency power supply is generated. The control unit performs relevant control according to the judgment information of the judgment unit, so as to realize timely control of power supply abnormality. By collecting data such as voltage, current, temperature, etc. in the Rectiverter emergency power supply, the battery status can be detected more accurately. By analyzing these data, better protection measures are provided for the battery pack, so that the performance of the battery can be fully utilized.

[0031] 2. In the present invention, by using the GAN model to expand the sample library, the data analysis error caused by the small sample size is reduced, and more basis is provided for the data processing in the RBF neural network prediction process. By training the RBF neural network with data in the current cycle time, updating the weight vector, and then predicting the real-time temperature value in the next cycle time, data support is provided for the work of the judgment unit and the control unit, and the working state of the Rectiverter emergency power supply in the next cycle time is pre-analyzed to achieve advanced control, reduce the loss of the battery in a bad state, and timely discover possible faults and hidden dangers, effectively control the state of the battery, and improve the performance and service life of the battery. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 The present invention proposes a structural schematic diagram of a three-in-one multifunctional emergency power management system;

[0033] Figure 2 It is a schematic diagram of the process in the present invention;

[0034] Figure 3 is a schematic diagram of the Rectivoerter HE module of the present invention;

[0035] Figure 4 It is a circuit diagram of the emergency power supply in the present invention. DETAILED DESCRIPTION

[0036] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0037] Embodiment 1

[0038] Reference Figure 1 , Figure 2 and Figure 3 Embodiment 1 further illustrates a three-in-one multifunctional emergency power management system proposed by the present invention.

[0039] A three-in-one multifunctional emergency power management system, comprising an input layer, a management layer and a display layer:

[0040] The input layer controls the AC and DC output of the energy storage device through the adjustment module. The energy storage device is also called the Rectivoerter HE module, which is used for the energy integration of the Rectiverter emergency power supply. The adjustment module enables the energy storage device to communicate with the controller through the CAN bus network. The adjustment module includes a bidirectional DC-DC converter and a DC-AC inverter (i.e., PWM inverter). The energy storage device is composed of one or more lithium-ion batteries. The AC input of the mains is given to the Rectivoerter HE module, which is converted from DC to AC by the DC-AC inverter and converted by voltage by the DC-DC converter to output 32A AC and 100A DC.

[0041] The Rectiverter emergency power supply uses a resonant topology to make efficiency the industry leader, with modular design and ultra-compact size. The Rectiverter HE module stands out with an efficiency of 96%, and is exquisitely designed and powerful. It integrates the inverter module, adjustment module, and static switch into one, reducing the number of components by 40%, improving reliability, extending life, and reducing working failures. The plug-and-play connection method reduces installation costs and saves installation time.

[0042] Rectiverter emergency power supply is powered by AC power supply and can output current to AC load, DC load and battery pack at the same time. When AC power is cut off, the power core can automatically invert the DC power stored in the battery pack to power AC load. It is an intelligent power supply product that uses built-in technology to transfer AC and DC power to provide high-quality, stable and uninterrupted DC and AC power.

[0043] The Rectiverter HE adjustment module complies with CE and UL requirements, meets NEBS certification, and can be installed and used worldwide.

[0044] The management layer includes a power management module and a monitoring module. The monitoring module monitors the operating environment of the system and performs data monitoring and collection during the working process of the Rectiverter emergency power supply in the current cycle time. The power management module includes a judgment unit, a prediction unit and a control unit. The judgment unit is used to judge the temperature safety status of the Rectiverter emergency power supply in the current cycle time and the next cycle time. The prediction unit expands the samples in the current cycle time through the GAN model, and predicts the temperature of the Rectiverter emergency power supply in the next cycle time through the RBF neural network model. The control unit performs corresponding work according to the output of the judgment unit.

[0045] The display layer is used for the end page display of data information and provides a system operation interface for operation and maintenance personnel. The display layer includes a human-computer interaction interface. The human-computer interaction module provides identity authentication for user login and makes functional adjustments to relevant parameters according to specific needs. The human-computer interaction module connects to the display device of the monitoring module to monitor and evaluate the operating status of the system.

[0046] The user authenticates his identity through the human-machine interface. The filter will intercept each manager's login request, and the authentication gateway will check the user's Ukey digital certificate with the identity information stored in the database to determine whether he has login permission. Only managers with login permission are allowed to log in to the system. After logging in to the system, managers have the permission to access the Smartpack-2 Touch monitor. The human-machine interface can be a computer or other smart devices that meet the functions. Users can use permissions to project the menu of the Smartpack-2 Touch monitor on the human-machine interface, and remotely call and read the monitoring data in the Smartpack-2 Touch monitor through the human-machine interface.

[0047] Users can also view the Smartpack-2 Touch monitor separately through its LCD screen. The Smartpack-2 Touch monitor's LCD screen is independent of the human-computer interaction interface. However, when querying the Smartpack-2 Touch monitor through the LCD screen, users also need to log in or perform remote authorization through the human-computer interaction interface.

[0048] The monitoring module includes a Smartpack-2 Touch monitor. The Smartpack-2 Touch monitor includes two functional areas: an indication area (LCD display, LED indicator) and a connection area (port). The connection area includes multiple connection ports. The Smartpack-2 Touch monitor is connected to the local area network via Ethernet, and the default IP is 192.168.10.20. The fixed IP address is debugged through another Ethernet port, which is used to access the fixed IP address of the controller 10.10.01, and the three-pin terminal of the third-party device is connected through the CAN line port. The indication area displays the working status through the LED light indication color, and the LED light indication is displayed through color, including the following:

[0049] □ "Power" (green light) indicates whether the power supply is working normally: if the green light is on, it means the power supply is working normally and outputs safety information. No adjustment is required and the next cycle of work can be continued. If it is not on, it means the monitor's working power supply is faulty or there is no AC power supply. The operation and maintenance personnel need to be reminded to check whether the AC power and AC cable are abnormal. This prompt information can be displayed on the human-computer interaction interface.

[0050] □ “Alarm” (red light) indicates whether the system is in alarm state: If the red light is on, it means the system is in alarm state (major alarm). If the red light is on all the time, it means the Rectiverter emergency power supply may have the following situations:

[0051] (1) The Rectiverter emergency power supply (or energy storage device) is shut down due to grid undervoltage, internal overtemperature or output overvoltage protection, reminding the operation and maintenance personnel to check whether the mains power is too low or obstacles are blocking the air duct. The Rectiverter emergency power supply can also be pulled out, and then pushed back into the power supply frame after a period of time and restarted. If the fault still exists, replace the Rectiverter emergency power supply with a new one;

[0052] (2) The auxiliary power supply in the Rectiverter emergency power supply module (energy storage device) fails or the CAN line fails, reminding the operation and maintenance personnel to replace the new Rectiverter emergency power supply;

[0053] (3) Fan failure (single or double fan failure), reminding the operation and maintenance personnel to check whether the fan is working properly;

[0054] (4) The output voltage of the Rectiverter emergency power supply module (energy storage device) is too low, reminding the operation and maintenance personnel to check the module output voltage.

[0055] □ "Warning" (yellow light) indicates whether the system is in an alarm state: If the yellow light is on, it means the system is in an alarm state (minor alarm). If the yellow light is on all the time, the Rectiverter emergency power supply may have the following situations:

[0056] (1) The power module is in a derated output state (reduced power output) due to internal overtemperature, grid undervoltage or fan failure. The operation and maintenance personnel are reminded to check whether the fan is normal, whether the grid voltage is too low, or remove obstacles blocking the air duct.

[0057] (2) The system starts remote current limiting charging to remind the operation and maintenance personnel to check whether the module charges the battery with too much current;

[0058] (3) The AC input voltage of the Rectiverter emergency power supply module (energy storage device) is out of range, reminding the operation and maintenance personnel to check the AC voltage;

[0059] (4) The Rectiverter emergency power supply is used alone, or the communication with Smartpack-2 is interrupted, reminding the operation and maintenance personnel to check whether the communication cable is loose;

[0060] The flashing red light indicates that the Rectiverter emergency power supply may be in AC overvoltage protection state, reminding the operation and maintenance personnel to check whether the mains voltage exceeds 300V;

[0061] The red and yellow lights are off, indicating that the Rectiverter emergency power supply is in normal working condition.

[0062] The above fault information and reminder information can be displayed on the human-computer interaction interface.

[0063] The LCD display is a very important part of the user interface of the power system. Smartpack2 Touch has a 4.4-inch responsive touch screen to run the web interface, that is, it has the same functions as a PC browser on the front display. Therefore, the indication area in the monitoring module is connected to the human-computer interaction interface of the display layer. The display layer has the highest authority to operate the monitoring module and can be remotely controlled through the wireless network. Only authorized users in the monitoring module can log in and have functional access and operation.

[0064] The monitoring module monitors the temperature status of the battery in real time through the Smartpack-2 Touch monitor. The monitoring module is set to perform continuous periodic measurements and acquire data from the Rectiverter emergency power supply multiple times within the current cycle time, thereby obtaining n real-time temperature values. The judgment unit averages the real-time temperature values ​​to form a temperature representative value within the cycle time T, and judges the value range to which the temperature representative value belongs.

[0065] If the temperature representative value is within the first value interval, a safety signal is output and the control unit does not work; if the temperature representative value is within the second value interval, a first-level alarm is output and the control unit controls the prediction unit to start;

[0066] If the temperature representative value is within the third value interval, a second-level alarm is output and the control unit turns off the emergency power supply of the Rectiverter.

[0067] The monitoring module can independently monitor each Rectiverter emergency power supply in the energy storage device, and this embodiment only performs relevant processing and regulation on one Rectiverter emergency power supply, while it is still applicable to multiple Rectiverter emergency power supplies in the corresponding energy storage device.

[0068] Assuming the real number η1<η2<η3<η4, the first value interval is [η1, η2], the second value interval is [η2, η3], and the third value interval is [η2, η3].

[0069] In this embodiment, the temperature status of the Rectiverter emergency power supply is monitored in real time by the monitoring module, so that more accurate battery status information can be obtained, and reliable data support is provided for the working status analysis of the Rectiverter emergency power supply. The data transmitted by the monitoring module is processed by the judgment module, and the judgment information of the Rectiverter emergency power supply is generated. The control unit performs relevant control according to the judgment information of the judgment unit to realize timely control of power supply abnormality. By collecting data such as voltage, current, and temperature in the Rectiverter emergency power supply, the battery status can be detected more accurately. By analyzing these data, better protection measures are provided for the battery pack, so that the battery performance can be fully utilized.

[0070] Embodiment 2

[0071] Reference Figure 1 and Figure 2 , Example 2 further illustrates a three-in-one multifunctional emergency power management system proposed by the present invention.

[0072] A three-in-one multifunctional emergency power management system, wherein the prediction unit is used to predict the temperature safety state of the Rectiverter emergency power supply within the next cycle time, and the prediction unit is started when the output is a minor alarm within the current cycle time. The specific working process of the prediction unit includes the following working process:

[0073] Step S61: Obtain the real-time temperature value and measurement time point corresponding to the secondary warning within the current cycle time, and form an import set D λ ={(T1,t1),......,(T a ,t a ),......,(T n ,t n )};

[0074] Step S62: Import the imported set as a sample into the trained GAN model for sample expansion, and set the output set formed by the expansion as S θ ={(T1,t1),......,(T i ,t i ),......,(T m ,t m )}, and m>n. Where i and a represent the labels of the sub-values ​​in the output set and the import set, and m and n represent the number of sub-values ​​in the output set and the import set.

[0075] Use the GAN model to expand the sample library. The training process of the GAN model output reference samples is as follows:

[0076] The framework of the GAN model is composed of a generator G (Generator) and a discriminator D (Generator). The entire training process is a confrontation between the two: given p data (x), p data (x) represents the real input data, and we hope to learn the generator G and the discriminator D so that p g =p data (x), p g Indicates the generated pseudo data, so that the discriminator D cannot correctly distinguish the two;

[0077] Generator G can be any neural network structure, and the input is random noise z~p z (z), random noise can be Gaussian distribution, uniform distribution, etc. The training goal is to make the output G(z) have a higher probability in the subsequent discrimination stage; the discriminator D can be any neural network structure, and the input is x or G(z); the training goal is to accurately distinguish the two, giving x a greater probability and G(z) a smaller probability (binary classification). The two networks are trained alternately, and the discriminator D is optimized for multiple rounds, while the generator G is only optimized for one round, and the capabilities are improved simultaneously.

[0078] GAN training is a solution to a minmax problem, and the corresponding objective function is as follows:

[0079]

[0080] In the formula, p data (x) is the real sample distribution, which represents the sample set sampled from the real training sample. In the present invention, it represents the import set obtained within the current cycle time. The import set is composed of the real-time temperature value set and the measurement time point set. That is, each single sub-value in the import set includes a real-time temperature value and the measurement time point corresponding to the real-time temperature value, which can be expressed as:

[0081] Real-time temperature value set X = {T1,......,T a ,......,T n};

[0082] The measurement time point set H = {t1,......,t a ,......,t n};

[0083] Import Set D λ ={(T1,t1),......,(Ta ,t a ),......,(T n ,t n )};

[0084] represents a real sample randomly sampled from the real sample distribution, that is, a real-time temperature value and measurement time point randomly sampled from the imported set; p z (z) is the random noise distribution; represents a noise sample randomly sampled from a random noise distribution, G(x) represents a neural network function with a real sample as a variable, G(z) represents a neural network function with a noise sample z as a variable, and D(G(z)) represents a neural network function with G(z) as a variable;

[0085] The training goal of the GAN model is to maximize the objective function. Through iterative training, the parameters of the generator and the discriminator are continuously adjusted, and finally the GAN model is obtained.

[0086] In the present invention, the real-time temperature values ​​measured by the Rectiverter emergency power supply of the Smartpack-2 Touch monitor within the current cycle time are aggregated to form a real-time temperature value set, the sample temperature values ​​and measurement time points are extracted from the import set, and selected and input into the trained GAN model, and the global optimal solution of the objective function of GAN is obtained to generate w groups of pseudo data, and the import set and the w groups of pseudo data are merged to form a new data set, that is, this new data set is the output set, and the output set is used as the output and input of the temperature prediction model to obtain the change weight of the temperature prediction model.

[0087] Each measurement time point in the output set is used as the input of the RBF neural network model, and the real-time temperature value corresponding to each measurement time point is used as the output. The RBF neural network model is trained and the weight vector is updated. Then, in the next cycle time, n measurement time points are randomly selected and input into the trained RBF neural network model to generate the real-time temperature values ​​corresponding to the n measurement time points.

[0088] The judgment unit judges the real-time temperature values ​​corresponding to the n measurement time points within the next cycle time:

[0089] If the output is a safety signal, continue to work normally;

[0090] If the output is a first-level alarm, the control unit controls the heat dissipation unit for regulation;

[0091] If the output is a level 2 alarm, a pre-warning is given or the Rectiverter emergency power is shut down.

[0092] In the RBF neural network, the input layer receives input data and passes it to the hidden layer; the hidden layer is composed of a series of radial basis functions, which calculate the output based on the distance between the input data and their respective center points; each basis function corresponds to a center point and a deviation parameter, and by adjusting these parameters, the basis function can adapt to the characteristics of the input data; the output layer calculates the final prediction result based on the output of the hidden layer; usually, the output layer uses a linear activation function to obtain continuous prediction values.

[0093] The present invention adopts RBF neural network to predict the real-time temperature value within the next cycle time, which solves the generalization problem of traditional BP neural network prediction evaluation. RBF neural network adopts random approximation method to obtain better prediction results:

[0094] Assume {t1,......,t a ,......,t n} T is the input vector, {b1,......,b a ,......,b n} T is the center point of the network radial basis function, Gn is the Gaussian basis function;

[0095]

[0096] And 1≤a≤n, b is the center vector input to the a-th hidden layer, δ is the basis width parameter of the a-th node, and is a number greater than zero;

[0097] So the network output is: As the weight vector, n measurement time points are randomly selected in the next cycle time and imported into the trained RBF neural network to generate the predicted real-time temperature values ​​corresponding to the n randomly selected measurement time points.

[0098] Because the current cycle time is different from the next cycle time, the n measurement time points selected in the current cycle time are also different from the n measurement time points selected in the next cycle time. For example, 5 measurement time points are selected within 1 minute of the current cycle time, and the measurement time points correspond to the charging time in the current cycle time respectively. Assuming that the charging time at the beginning of the current cycle time is 120 minutes, then assuming that the 5 measurement time points are 120.15, 120.2, 120.3, 120.7 and 120.9, the random 5 measurement time points in the next cycle time are 121.1, 121.2, 121.5, 121.8 and 120.9. Therefore, the measurement time points selected in each cycle time are different, so the real-time temperature values ​​predicted in the next cycle time are also different and have reference value.

[0099] The control unit includes a heat dissipation unit, which wirelessly controls the heat dissipation device, and the heat dissipation device includes a front fan. The heat dissipation unit controls the output power of the front fan to increase and decrease the airflow from front to back. It also includes monitoring whether there are any obstacles in the air duct and reminding the operation and maintenance personnel to remove the obstacles.

[0100] The front-mounted high-efficiency, long-life fan ensures airflow from front to back. The radiator is directly mounted on the inner wall of the aluminum alloy frame and the independent front-to-back straight-through air duct design allows the module to adapt to a wide range of working environments, making system solutions easier and more convenient to implement.

[0101] In this embodiment, by using the GAN model to expand the sample library, the data analysis error caused by the small sample size is reduced, and more basis is provided for the data processing in the RBF neural network prediction process. By training the RBF neural network with data in the current cycle time, updating the weight vector, and then predicting the real-time temperature value in the next cycle time, data support is provided for the work of the judgment unit and the control unit, and the working state of the Rectiverter emergency power supply in the next cycle time is pre-analyzed to achieve advanced control, reduce the loss of the battery in a bad state, and timely discover possible faults and hidden dangers, effectively control the state of the battery, and improve the performance and service life of the battery.

[0102] Embodiment 3

[0103] Reference Figure 1 , Figure 3 and Figure 4 , Example 2 further illustrates a three-in-one multifunctional emergency power management system proposed by the present invention.

[0104] A three-in-one multifunctional emergency power supply management system, wherein the management layer is applied to a microcomputer, the human-computer interaction interface includes multiple clients, the clients are connected via wireless or wired communication instruments, the communication connection follows the HTTP protocol, and an integrated information processing system and a multi-client information architecture are adopted. The three-in-one multifunctional emergency power supply management system is applied to a cloud processing platform, to the cloud, and performs distributed storage.

[0105] The three-in-one multifunctional emergency power management system includes the following steps:

[0106] Step S1: The input layer controls the AC and DC output of the energy storage device through the adjustment module;

[0107] Step S2: The monitoring module monitors the operating environment of the system and performs data monitoring and collection during the working process of the Rectiverter emergency power supply during the current cycle time.

[0108] Step S3: The judging unit judges the temperature safety state of the Rectiverter emergency power supply within the cycle time, and the control unit performs corresponding work according to the output of the judging unit;

[0109] Step S4: The prediction unit expands the samples in the current cycle time through the GAN model, and predicts the temperature of the Rectiverter emergency power supply in the next cycle time through the RBF neural network model;

[0110] Step S5: Repeat step S3.

[0111] During emergency power supply operation

[0112] Select the AC input power of the emergency power supply (the AC switch is recommended to use more than 40A, and the two AC inputs should use the same source and the same phase), and connect it to the energy storage device through an AC cable. The energy storage device is connected to the battery pack through a cable, and the DC load output is performed through the DC connection cable, and the AC load output is performed through the AC connection cable. Control the AC input switch, power on the emergency power supply module, confirm that the DC output and AC output of the emergency power supply are normal, close the temporary battery switch (at this time, the emergency power supply will charge the temporary battery pack), and then the emergency power supply is put into operation.

[0113] In the circuit diagram, including the connection of the line, PDU means power distribution unit. In the process of connecting the AC cable and the ground wire, the AC input circuit breaker of the emergency power supply system is a 2P (63A) molded case switch, and the AC input is a single-phase AC input. Each Rectiverter emergency power supply back panel of the energy storage equipment is equipped with an AC fast plug socket; the AC input of this system adopts single-phase AC mains power supply, and the AC upper circuit breaker should be ≥63A; one end of the ground wire is connected to the ground bar, and the other end is connected to the grounding column on the back panel of the emergency power supply for inspection and confirmation with a multimeter. The order of AC wiring: first connect the ground wire, then connect the neutral wire, and finally connect the live wire.

[0114] In the process of connecting the battery cable and the DC load cable, the DC load output circuit breaker of the Rectiverter emergency power supply is a 2P (100A) molded case DC dedicated switch, and the back panel of the Rectiverter emergency power supply is equipped with 2 load output DC fast plug sockets (200A). The battery output circuit breaker of the emergency power supply is a 2P (100A) molded case DC dedicated switch, and the back panel of the emergency power supply is equipped with 2 battery output DC fast plug sockets (200A). Both ends of the DC load connection cable are fast plug-ins, and one end of the battery connection cable is a fast plug-in and the other end is a wire nose.

[0115] During the AC load cable connection process, the AC output of the Rectiverter emergency power supply is composed of 2 PDUs and two AC output cables (which can be connected to the AC distribution unit for use). The four-way AC output circuit breaker uses 4 2P (32A) micro-break AC switches. The back panel of the Rectiverter emergency power supply is equipped with 2 output AC flow rate plug sockets (32A) and 2 8-bit 10A new national standard five-hole PDUs. One end of the AC output cable is a speed plug-in and the other end is a wire end. The order of AC wiring: first connect the ground wire, then connect the neutral wire, and finally connect the live wire. Among them, SC represents the switch, L represents the live wire or positive pole, N represents the neutral wire or negative pole, and PE represents the ground wire.

[0116] In addition, according to the implementation of the present application, the process described in the figure of a three-in-one multifunctional emergency power management system can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, which stores machine-readable instructions, and the machine-readable instructions can be executed by a processor to execute instructions corresponding to the method steps provided in the present application. Of course, the architecture shown in the figure of a three-in-one multifunctional emergency power management system is only exemplary. When implementing different devices, adaptive selection or adjustment can be performed according to actual needs.

[0117] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and thresholds in the formula are set by technicians in this field according to actual conditions.

[0118] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technical users in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A three-in-one multifunctional emergency power management system, characterized in that: The three-in-one multifunctional emergency power management system includes an input layer, a management layer and a display layer: The input layer controls the AC and DC output of the energy storage device through the adjustment module, which is used for the energy integration of the Rectiverter emergency power supply. The adjustment module enables the energy storage device to communicate with the controller through the CAN bus network. The management layer includes a power management module and a monitoring module. The monitoring module monitors the operating environment of the system and performs data monitoring and collection during the working process of the Rectiverter emergency power supply during the current cycle time. The power management module includes a judgment unit, a prediction unit and a control unit. The judgment unit is used to judge the temperature safety status of the Rectiverter emergency power supply during the current cycle time and the next cycle time. The prediction unit expands the samples within the current cycle time through the GAN model and predicts the temperature of the Rectiverter emergency power supply during the next cycle time through the RBF neural network model. The control unit performs corresponding work according to the output of the judgment unit; The training process of the GAN model output reference sample is as follows: The framework of the GAN model is constructed. The framework consists of a generator G and a discriminator D. The objective function of GAN is as follows: In the formula, p data (x) is the real sample distribution, which means the sample set sampled from the real training samples. represents a real sample randomly sampled from the real sample distribution, p z (z) is the random noise distribution; represents a noise sample randomly sampled from a random noise distribution, D(x) represents the neural network function with the real sample as the variable, G(z) represents the neural network function with the noise sample z as the variable, and D(G(z)) represents the neural network function with G(z) as the variable; The training goal of the GAN model is to maximize the objective function. Through iterative training, the parameters of the generator and the discriminator are continuously adjusted, and finally the GAN model is obtained. The sample temperature values ​​and measurement time points extracted from the import set are input into the trained GAN model to generate w sets of pseudo data. The import set and the w sets of pseudo data are merged to form a new data set, which is the output set. Each measurement time point in the output set is used as the input of the RBF neural network model, and the real-time temperature value corresponding to each measurement time point is used as the output. The RBF neural network model is trained and the weight vector is updated. Then, in the next cycle time, n measurement time points are randomly selected and input into the trained RBF neural network model to generate the real-time temperature values ​​corresponding to the n measurement time points. The judgment unit judges the real-time temperature values ​​corresponding to the n measurement time points within the next cycle time: If the output is a safety signal, continue to work normally; If the output is a first-level alarm, the control unit controls the heat dissipation unit for regulation; If the output is a level 2 alarm, a pre-warning is given or the Rectiverter emergency power supply is shut down; The display layer is used for the end page display of data information and provides a system operation interface for operation and maintenance personnel.

2. A three-in-one multifunctional emergency power management system according to claim 1, characterized in that: The adjustment module includes a bidirectional DC-DC converter and a DC-AC converter. The energy storage device is also called a Rectivoerter HE module. The energy storage device is composed of one or more lithium-ion batteries. The AC power is input to the Rectivoerter HE module, and after the DC-AC converter converts the DC and AC, and the DC-DC converter converts the voltage, it outputs 32A AC and 100A DC.

3. A three-in-one multifunctional emergency power management system according to claim 1, characterized in that: The display layer includes a human-computer interaction interface. The human-computer interaction module provides identity authentication for user login and functionally adjusts relevant parameters according to specific needs. The human-computer interaction module is connected to the display device of the monitoring module to monitor and evaluate the operating status of the system.

4. A three-in-one multifunctional emergency power management system according to claim 1, characterized in that: The monitoring module includes a Smartpack-2Touch monitor, which includes an indication area and a connection area. The Smartpack-2Touch monitor is connected to a local area network via Ethernet, and the indication area displays a working status through an LED light indicating a color.

5. A three-in-one multifunctional emergency power management system according to claim 4, characterized in that: The monitoring module monitors the temperature status of the battery in real time through the Smartpack-2Touch monitor, sets the monitoring module to perform continuous periodic measurements, and acquires data of the Rectiverter emergency power supply multiple times within the current cycle time, thereby obtaining n real-time temperature values. The judgment unit averages the real-time temperature values ​​to form a temperature representative value within the cycle time T, and judges the value range to which the temperature representative value belongs. If the temperature representative value is within the first value interval, a safety signal is output and the control unit does not work; If the temperature representative value is within the second value interval, a first-level alarm is output, and the control unit controls the prediction unit to start; If the temperature representative value is within the third value interval, a second-level alarm is output and the control unit turns off the emergency power supply of the Rectiverter.

6. A three-in-one multifunctional emergency power management system according to claim 4, characterized in that: The prediction unit is used to predict the temperature safety state of the Rectiverter emergency power supply within the next cycle time, and the prediction unit is started when the output is a minor alarm within the current cycle time. The specific working process of the prediction unit includes the following working process: Step S61: Obtain the real-time temperature value and measurement time point corresponding to the secondary warning within the current cycle time, and form an import set D λ ={(T1,t1),……,(T a ,t a ),……,(T n ,t n )}; Step S62: Import the imported set as a sample into the trained GAN model for sample expansion, and set the output set formed by the expansion as S θ ={(T1,t1),……,(T i ,t i ),……,(T m ,t m )}, and m>n, where i and a represent the labels of the sub-values ​​in the output set and the import set, and m and n represent the number of sub-values ​​in the output set and the import set.

7. A three-in-one multifunctional emergency power management system according to claim 6, characterized in that: The control unit includes a heat dissipation unit, which wirelessly controls a heat dissipation device, and the heat dissipation device includes a front fan. The heat dissipation unit controls the output power of the front fan to increase and decrease the airflow from front to back, and also monitors whether there are any obstacles in the air duct, and reminds the operation and maintenance personnel to remove the obstacles.

8. A three-in-one multifunctional emergency power management system according to claim 7, characterized in that: The management layer is applied to a microcomputer, and the human-computer interaction interface includes multiple clients, which are connected via wireless or wired communication instruments. The communication connection follows the HTTP protocol and adopts an integrated information processing system and a multi-client information architecture.

9. A three-in-one multifunctional emergency power management system according to claim 8, characterized in that: The three-in-one multifunctional emergency power supply management system is applied to a cloud processing platform, applied to the cloud, and performs distributed storage.