AI intelligent management system based on power supply module

Abstract

The invention discloses an AI intelligent management system based on a power supply module, and relates to the technical field of power supply management, and the system comprises the following steps: obtaining power supply monitoring information, carrying out the abnormality analysis of the power supply monitoring information, and obtaining power supply abnormality information, the power supply monitoring information comprising the temperature information, voltage information and current information of a power supply; the power supply monitoring information is processed and stored; building an intelligent analysis model based on the DQN model, carrying out the iterative training of the intelligent analysis model through the power monitoring information, and generating a power management strategy based on the trained intelligent analysis model; generating a power supply control instruction based on the power supply management strategy and the power supply abnormal information to control and manage the power supply; the method is used for solving the problem that the existing power supply management technology cannot perform optimal management on the operation benefit of the power supply by using the artificial intelligence technology while monitoring the temperature, voltage and current of the power supply and ensuring the normal operation of the power supply.

Description

Technical Field

[0001] The present invention relates to the technical field of power management, and in particular to an AI intelligent management system based on a power module. Background Art

[0002] Power management technology refers to the technology of controlling, monitoring and managing power supply to ensure the stable operation of equipment under different load conditions, improve energy utilization efficiency and extend the service life of equipment. The core of power management includes power distribution, power protection and power energy saving. Power protection involves taking corresponding protective measures under abnormal power conditions to prevent equipment damage; power energy saving is to reduce power supply operating costs and energy waste by intelligently controlling power supply;

[0003] In the existing improvement process for the management of power modules, more and more reliance is placed on digital analysis by computers to get rid of the original manual optimization method. However, in the existing process of using computers for digital analysis to manage power modules, the management method used is too simple. Usually, thresholds are set based on the basic parameters of the power modules, and the power modules are monitored by computers. When the operating parameters of the power modules exceed the thresholds, the power modules are shut down. For example, in the patent application with publication number CN114356064A, a power management method and system are disclosed. The scheme is to set the thresholds of the standby time of the power module and the power consumption of the power supply. By adjusting the current preset standby time, the power consumption of the power management system is within the preset range, thereby improving the performance of the equipment and reducing the power consumption. However, the scheme relies on a fixed program and lacks learning ability and adaptability, resulting in the optimization control of energy consumption being not precise and intelligent enough, and the degree of optimization is limited. In addition, the power supply operation process is not monitored, and robustness is lacking. The existing power management technology cannot monitor the temperature, voltage and current of the power supply to ensure the normal operation of the power supply while optimizing the operation efficiency of the power supply using artificial intelligence technology. Summary of the invention

[0004] The present invention aims to solve one of the technical problems in the prior art to at least a certain extent, by monitoring the temperature, voltage and current of the power supply to ensure the normal operation of the power supply, and then building an intelligent optimization model based on the DQN model, and using the intelligent optimization model to optimize the management of the power supply; so as to solve the problem that the existing power management technology is unable to monitor the temperature, voltage and current of the power supply to ensure the normal operation of the power supply while using artificial intelligence technology to optimize the management of the operating efficiency of the power supply.

[0005] To achieve the above objectives, the present application provides an AI intelligent management system based on a power module, including: a power monitoring module, a data management module, an AI intelligent module and a central control module;

[0006] The power supply monitoring module is used to obtain power supply monitoring information, and perform abnormal analysis on the power supply monitoring information to obtain power supply abnormality information. The power supply monitoring information includes temperature information, voltage information, and current information of the power supply;

[0007] The data management module includes a data processing unit and a data storage unit, the data processing unit is used to process the power monitoring information, and the data storage unit is used to store data;

[0008] The AI ​​intelligent module includes a construction unit and an analysis unit. The construction unit constructs an intelligent optimization model based on the DQN model, and constructs the state space, action space and reward function of the intelligent optimization model; the analysis unit includes an intelligent optimization model, which is used to generate a power management strategy;

[0009] The central control module includes an instruction generation unit and an instruction execution unit. The instruction generation unit generates power control instructions based on power management strategy and power abnormality information and transmits them to the instruction execution unit. The instruction execution unit controls and manages the power based on the power control instructions.

[0010] Furthermore, the power supply monitoring module includes a temperature monitoring unit, and the temperature monitoring unit is configured with a temperature monitoring strategy, and the temperature monitoring strategy includes:

[0011] The temperature sensor installed inside the power module collects the temperature information T of the power supply in real time and sends the temperature information to the data management module;

[0012] The safety temperature is set based on the power usage parameters. When the temperature monitoring unit detects that the power temperature exceeds the safety temperature, the temperature monitoring unit issues a temperature abnormality alarm, generates temperature abnormality information, and sends the temperature abnormality information to the central control module.

[0013] Furthermore, the power supply monitoring module includes a voltage monitoring unit, and the voltage monitoring unit is configured with a voltage monitoring strategy, and the voltage monitoring strategy includes:

[0014] The voltage monitoring unit is based on a subtraction operational amplifier circuit composed of an operational amplifier. The two input terminals of the subtraction operational amplifier circuit are connected to the positive and negative electrodes of the power supply respectively, and the output voltage of the subtraction operational amplifier circuit is collected. The output voltage U of the power supply is calculated by the principle formula of the subtraction operational amplifier circuit. The voltage monitoring unit obtains the output voltage U of the power supply in real time and sends the output voltage to the data management module.

[0015] The safe voltage range is set based on the voltage usage parameters. When the voltage monitoring unit detects that the output voltage U exceeds the safe voltage range, the voltage monitoring unit issues a voltage abnormality alarm, generates voltage abnormality information, and sends the voltage abnormality information to the central control module.

[0016] Furthermore, the power monitoring module includes a current monitoring unit, and the current monitoring unit is configured with a current monitoring strategy, and the current monitoring strategy includes:

[0017] The current monitoring unit is connected in series with a resistor R in the positive path of the power supply output end. i The resistor is marked as the sampling resistor; the voltage U of the sampling resistor is obtained by using the subtraction amplifier circuit. i , according to Ohm's law The current monitoring unit obtains the output current I of the power supply in real time and sends the output current to the data management module;

[0018] The safe current range is set based on the power usage parameters. When the temperature monitoring unit detects that the power output current I exceeds the safe current range, the current monitoring unit issues a current abnormality alarm, generates current abnormality information, and sends the current abnormality information to the central control module.

[0019] Furthermore, the data management module is configured with a data management strategy, which includes:

[0020] The data processing unit calculates the output power of the power supply based on the output voltage and the output current to obtain the output power P, and calculates the output power based on the output power to obtain the output power W;

[0021] The data storage unit is used to store the power monitoring information obtained by the power monitoring module and the output power P and output power W obtained by the data processing unit.

[0022] Furthermore, the construction unit is configured with a construction strategy, which includes:

[0023] Based on the DQN model, the activation function ReLU function of the DQN model is replaced by the PReLU function, the improved DQN model is marked as an intelligent optimization model, and the intelligent optimization model is output to the analysis unit.

[0024] Furthermore, the construction strategy also includes:

[0025] Set the power supply operation cost as C, and set minimizing the power supply operation cost as the intelligent management goal;

[0026] Construct state space: define the state space as S, which includes: the output power P of the power supply, the maximum output power P of the power supply m and electricity price V, the state space S is represented by S = {P, P m , V}, S t represents the state space at time t;

[0027] Construct action space: Define the action space as A, which is expressed as A = {Pi}, P i Represents the set of output powers of the power supply at different output currents;

[0028] Set reward function: Based on the intelligent management goal, set the reward function R. Through the reward function R, the intelligent optimization model learns the action of minimizing the power supply operation cost while ensuring the normal operation of the power supply.

[0029] Furthermore, the analysis unit includes an intelligent optimization model, and the intelligent optimization model is used to generate a power management strategy including:

[0030] Initialize the parameters of the intelligent optimization model, set the experience replay pool and related hyperparameters; obtain the state space s1 of the current power supply, use the ε-greedy strategy to select an action space with probability ε, marked as a1, execute the selected action space a1, the state space of the power supply changes, obtain the next state space of the power supply, marked as s2, and calculate the reward r, put the state space s1, action space a1, next state space s2 and reward r into the experience replay pool; while the intelligent optimization model power supply interacts, the intelligent optimization model uses the data in the experience replay pool for training iterations; generate a power management strategy based on the intelligent optimization model that has completed the training iteration, and output the power management strategy to the central control module.

[0031] Furthermore, the instruction generation unit is configured with an instruction generation strategy, and the instruction generation strategy includes:

[0032] The power control instructions include power adjustment instructions and power off instructions;

[0033] When the instruction generation unit receives power abnormality information, it generates a power off instruction and outputs the power off instruction to the instruction execution unit. The power abnormality information includes: temperature abnormality information, voltage abnormality information and current abnormality information;

[0034] When the instruction generation unit receives the power management policy, it generates a power adjustment instruction according to the power management policy and outputs the instruction to the instruction execution unit.

[0035] Furthermore, the instruction execution unit is configured with an instruction execution strategy, and the instruction execution strategy includes:

[0036] When the instruction execution unit receives a power-off instruction, it shuts down the power; when the instruction execution unit receives a power-adjustment instruction, it adjusts the power accordingly based on the power-adjustment instruction, and outputs the state space of the adjusted power to the AI ​​intelligent module.

[0037] Beneficial effects of the present invention: the present invention first obtains power supply monitoring information, performs abnormal analysis on the power supply monitoring information to obtain power supply abnormality information, the power supply monitoring information includes temperature information, voltage information and current information of the power supply; then processes and stores the power supply monitoring information; then builds an intelligent analysis model based on the DQN model, iteratively trains the intelligent analysis model using the power supply monitoring information, and then generates a power supply management strategy based on the trained intelligent analysis model; finally, generates a power supply control instruction based on the power supply management strategy and the power supply abnormality information to control and manage the power supply; the present invention monitors the temperature, current and voltage of the power supply, and generates a power supply management strategy using the intelligent analysis model, which can realize the optimization management of the operation efficiency of the power supply by artificial intelligence technology while ensuring the normal operation of the power supply;

[0038] The DQN model of the present invention constructs an intelligent analysis model. The DQN model combines deep learning and reinforcement learning technologies, has self-supervised learning capabilities and continuous improvement capabilities, can make effective decisions in complex environments, and ensures optimization efficiency and accuracy in the process of optimizing the operating benefits of a power supply. The temperature, current and voltage of the power supply are monitored to understand the operating status of the power supply in real time, ensure the normal operation of the power supply, and improve safety and robustness. Replacing the activation function ReLU function of the DQN model with the PReLU function can alleviate the mean shift problem and the neuron death problem, and the PReLU function has a faster convergence speed, which can improve the model accuracy, reduce the training time of the model, and thus improve the processing efficiency of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is a functional block diagram of the system of the present invention;

[0040] Figure 2 is a flow chart of the method steps of the present invention;

[0041] Figure 3 It is a structural diagram of the subtraction operation amplifier circuit of the present invention;

[0042] Figure 4 This is a structural diagram of the intelligent optimization model of the present invention. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] Example 1, please refer to Figure 1As shown, in the first aspect, the present application provides a system for generating an electrical stimulation signal, including a power monitoring module, a data management module, an AI intelligent module and a central control module;

[0045] The power supply monitoring module is used to obtain power supply monitoring information, and perform abnormal analysis on the power supply monitoring information to obtain power supply abnormality information. The power supply monitoring information includes temperature information, voltage information, and current information of the power supply;

[0046] The power supply monitoring module includes a temperature monitoring unit, which is configured with a temperature monitoring strategy. The temperature monitoring strategy includes: collecting temperature information T of the power supply in real time through a temperature sensor installed inside the power supply module, and sending the temperature information to the data management module;

[0047] A safe temperature is set based on the power usage parameters. When the temperature monitoring unit detects that the power temperature exceeds the safe temperature, the temperature monitoring unit issues a temperature abnormality alarm, generates temperature abnormality information, and sends the temperature abnormality information to the central control module;

[0048] The safety temperature must be set according to the actual power parameters. For example, the operating temperature range of lithium batteries is -20°C to 60°C. When the power source is a lithium battery, the safety temperature can be set to [-20°C, 60°C]. When the power temperature exceeds 60°C or is lower than -20°C, it is detected that the power temperature exceeds the safety temperature.

[0049] The power supply monitoring module includes a voltage monitoring unit, which is configured with a voltage monitoring strategy. The voltage monitoring strategy includes: the voltage monitoring unit is based on a subtraction operational amplifier circuit composed of an operational amplifier, see Figure 3 As shown, the two input terminals of the subtraction operational amplifier circuit are respectively connected to the positive and negative electrodes of the power supply, the output terminal voltage of the subtraction operational amplifier circuit is collected, and the output voltage U of the power supply is calculated by the principle formula of the subtraction operational amplifier circuit; the voltage monitoring unit obtains the output voltage U of the power supply in real time, and sends the output voltage to the data management module; in this embodiment, the principle formula of the subtraction operational amplifier circuit is: When R1=R2 and R f =R3, the principle formula is: When R f =R1, then: μ0 = μ i2 -μ i1 , and the voltage amplification factor is In the above formula, R1 is the negative input resistance of the subtraction amplifier circuit; R2 is the positive input resistance of the subtraction amplifier circuit; R3 is the ground resistance of the subtraction amplifier circuit; R f is the feedback resistor of the subtraction amplifier circuit; μ0 is the output voltage of the subtraction amplifier circuit; μ i1 is the negative voltage of the subtraction amplifier circuit; μi2 A is the positive voltage of the subtraction amplifier circuit; μf It is the voltage amplification factor; just connect the two input ends of the subtraction amplifier circuit to the positive and negative ends of the single battery, and connect the output end to the analog-to-digital conversion pin of the microcontroller, and the single battery voltage can be monitored and collected; this design not only has low power consumption, high accuracy, and real-time monitoring, but also can directly transmit the output voltage to the monitoring equipment, which greatly facilitates the data transmission and processing of the system and expands the voltage detection range;

[0050] A safe voltage range is set based on the voltage usage parameters. When the voltage monitoring unit detects that the output voltage U exceeds the safe voltage range, the voltage monitoring unit issues a voltage abnormality alarm, generates voltage abnormality information, and sends the voltage abnormality information to the central control module;

[0051] The power supply monitoring module includes a current monitoring unit, and the current monitoring unit is configured with a current monitoring strategy including: the current monitoring unit is connected in series with a resistor R in the positive path of the power supply output end. i The resistor is marked as the sampling resistor; the voltage U of the sampling resistor is obtained by using the subtraction amplifier circuit. i , according to Ohm's law The output current I is calculated; the current monitoring unit obtains the output current I of the power supply in real time and sends the output current to the data management module; the current monitoring strategy can improve the current detection range and accuracy by amplifying the voltage multiple, and it can be realized by simply connecting a small resistor in series with the system charging and discharging branch and performing voltage detection on it. The circuit is simple and easy to implement. Under normal circumstances, a resistor with a smaller resistance value is selected as the sampling resistor, generally at the milliohm level;

[0052] A safe current range is set based on the power usage parameters. When the temperature monitoring unit detects that the power output current I exceeds the safe current range, the current monitoring unit issues a current abnormality alarm, generates current abnormality information, and sends the current abnormality information to the central control module;

[0053] In the specific implementation process, through temperature monitoring, abnormal reactions of power supply under extreme temperature conditions can be discovered and handled in real time, thus avoiding dangers;

[0054] The voltage monitoring method of the power supply must have the characteristics of high accuracy, high sensitivity and real-time monitoring. The method of detecting voltage using a subtraction operational amplifier circuit composed of an integrated operational amplifier can not only detect the battery terminal voltage in real time with high accuracy, but also expand or reduce the voltage detection range according to actual conditions. In addition, this solution has low circuit cost and power consumption, which meets the requirements for voltage detection.

[0055] When the load is short-circuited or the discharge current is too large, the life of the power supply may be shortened or it may be scrapped prematurely, and it is easy to generate sparks, combustion or even explosion. Therefore, the current must be monitored in real time during the use of the power supply. Traditional monitoring and protection mechanisms mostly use a fuse mechanism, that is, when the current is too large, the current fuse connected in series with the power supply will be directly blown. This method can effectively avoid damage caused by excessive current, but after the fuse blows, the user needs to replace the fuse or directly replace the protection module, which is very inconvenient and wasteful. According to Ohm's law, the design of obtaining the battery current by detecting the voltage of the series resistor avoids the disadvantage of one-time consumption of the traditional fuse mechanism.

[0056] The data management module includes a data processing unit and a data storage unit, the data processing unit is used to process the power monitoring information, and the data storage unit is used to store data;

[0057] The data management module is configured with a data management strategy, which includes: the data processing unit calculates the output power of the power supply based on the output voltage and output current to obtain the output power P, calculates the output power based on the output power to obtain the output power W; the formula for calculating the output power is Where W represents the output power, P represents the output power, and t1-t2 represents the time range;

[0058] The data storage unit is used to store the power monitoring information obtained by the power monitoring module and the output power P and output power W processed by the data processing unit;

[0059] In the specific implementation process, by accurately calculating the output power of the power supply, it is possible to effectively identify and reduce energy waste phenomena, such as excessive power distortion, thereby improving overall energy efficiency. The power supply strategy can be adjusted in combination with the load characteristics to ensure the best match between the power supply and the load under different load conditions, thereby improving energy utilization. It can also prevent equipment damage and power failure caused by overload and ensure stable operation of the power supply.

[0060] The AI ​​intelligent module includes a construction unit and an analysis unit. The construction unit constructs an intelligent optimization model based on the DQN model, and constructs the state space, action space and reward function of the intelligent optimization model; the analysis unit includes an intelligent optimization model, and the intelligent optimization model is used to generate a power management strategy;

[0061] The construction unit is configured with a construction strategy, which includes: improving the DQN model, replacing the activation function ReLU function of the DQN model with the PReLU function, marking the improved DQN model as an intelligent optimization model, and outputting the intelligent optimization model to the analysis unit; replacing the ReLU function with the PReLU function can optimize and alleviate the mean shift problem and neuron death problem in the use process of the intelligent optimization model, and the PReLU function has a faster convergence speed, which can improve the model accuracy, reduce the model training time, and thus improve the model processing efficiency;

[0062] The construction strategy also includes: setting the power supply operation cost as C, and setting minimizing the power supply operation cost as the intelligent management goal;

[0063] In this embodiment, the operating cost C of minimizing the power supply is expressed as Where V(t) represents the electricity price at time t, P(t) represents the output power of the power supply at time t, and T represents the duration of the power supply outputting at P(t).

[0064] Construct state space: define the state space as S, which includes: the output power P of the power supply, the maximum output power P of the power supply m and electricity price V, the state space S is represented by S = {P, P m , V}, S t represents the state space at time t;

[0065] Construct action space: Define the action space as A, which is expressed as A = {P i}, P i Represents the set of output powers of the power supply at different output currents;

[0066] Setting reward function: Based on the intelligent management goal, a reward function R is set. Through the reward function R, the intelligent optimization model learns the action of minimizing the power supply operation cost under the premise of ensuring the normal operation of the power supply. In this embodiment, the reward function R is expressed as, R = (V(t) - V min )*P(t)*T, where V(t) represents the electricity price at time t, V min represents the lowest electricity price in the T time period, P(t) represents the output power of the power supply at time t, and T represents the duration of the power supply outputting at P(t);

[0067] The analysis unit includes an intelligent optimization model, and the intelligent optimization model is used to generate a power management strategy including: initializing the parameters of the intelligent optimization model, setting the experience replay pool and related hyperparameters; in this embodiment, the learning rate in the hyperparameters is set to 0.0008, the discount factor is set to 0.90, and the batch size is set to 31; the target network parameter update step is set to 300;

[0068] Get the state space s1 of the current power source, use the ε-greedy strategy to select an action space with probability ε, marked as a1, execute the selected action space a1, the state space of the power source changes, get the next state space of the power source, marked as s2, and calculate the reward r, put the state space s1, action space a1, next state space s2 and reward r into the experience replay pool; while the intelligent optimization model power source interacts, the intelligent optimization model uses the data in the experience replay pool for training iteration; in this embodiment, specifically, please refer to Figure 4 As shown in the figure, the intelligent optimization model includes the current network and the target network. The current network inputs the current state s1 and action a1, and outputs an estimated Q value. The Q value represents the expected total return that can be obtained from the current state to the future state after taking a specific action in a given state. The target network inputs the state s2 at the next moment, calculates the reward r, puts the state space s1, action space a1, next state space s2 and reward r into the experience replay pool, and outputs the maximum Q value as the target Q value. The main network and the target network are fed back through the DQN error function to train the main network, and the network parameters of the main network are copied to the target network at regular intervals.

[0069] Generate a power management strategy based on the intelligent optimization model that has completed training iterations, and output the power management strategy to the central control module;

[0070] In the specific implementation process, the DQN model combines deep learning and reinforcement learning technologies, has self-supervised learning capabilities and continuous improvement capabilities, and can make effective decisions in complex environments. An intelligent optimization model is built based on the DQN model, which ensures optimization efficiency and accuracy in the process of optimizing the operating benefits of the power supply.

[0071] The probability ε in the ε-greedy strategy can be set by yourself. Based on the ε-greedy strategy, the optimal action is explored and utilized. In the exploration stage, the algorithm selects the action with the highest Q value with a probability of 1-ε, and selects other actions with a probability of ε; in the initial stage of training, ε can be set to a larger value so that the algorithm can explore as many actions as possible; during the training process, the exploration rate ε gradually decreases with the increase in the number of iterations, and the algorithm gives priority to the action with the largest Q value, which has better convergence; after the training stage, the algorithm directly selects the action with the highest Q value. The Q value represents the expected total return from the current state to the future state after taking a specific action under a given state;

[0072] The central control module includes an instruction generation unit and an instruction execution unit. The instruction generation unit generates power control instructions based on power management strategy and power abnormality information and transmits them to the instruction execution unit. The instruction execution unit controls and manages the power based on the power control instructions.

[0073] The instruction generation unit is configured with an instruction generation strategy, and the instruction generation strategy includes: the power control instruction includes a power adjustment instruction and a power shutdown instruction;

[0074] When the instruction generation unit receives power abnormality information, it generates a power off instruction and outputs the power off instruction to the instruction execution unit. The power abnormality information includes: temperature abnormality information, voltage abnormality information and current abnormality information;

[0075] When the instruction generation unit receives the power management policy, it generates a power adjustment instruction according to the power management policy and outputs it to the instruction execution unit;

[0076] The instruction execution unit is configured with an instruction execution strategy, and the instruction execution strategy includes: when the instruction execution unit receives a power-off instruction, executing a power-off operation on the power supply;

[0077] When the instruction execution unit receives the power adjustment instruction, it adjusts the power supply accordingly based on the power adjustment instruction, and outputs the state space of the adjusted power supply to the AI ​​intelligent module;

[0078] In the specific implementation process, shutting down the power supply in time when receiving power abnormality information can prevent hardware damage caused by power fluctuations and reduce long-term potential damage. Continuous power fluctuations have long-term negative effects on electrical equipment, and the accumulated damage may cause equipment failure at some point in the future. Shutting down the power supply in time can reduce the probability of such potential damage and prevent electrical accidents, including electric shock or fire.

[0079] Example 2, please refer to Figure 2 As shown, in the second aspect, the present application provides a power module AI intelligent management method, comprising the following steps:

[0080] Step S1, obtaining power monitoring information, performing abnormal analysis on the power monitoring information to obtain power abnormality information, the power monitoring information including temperature information, voltage information and current information of the power supply; Step S1 includes the following sub-steps:

[0081] Step S101, collecting temperature information T of the power supply in real time through a temperature sensor installed inside the power supply module, and sending the temperature information to a data management module;

[0082] Step S102, setting a safe temperature based on power usage parameters;

[0083] Step S103, when the temperature monitoring unit detects that the power supply temperature exceeds the safe temperature, the temperature monitoring unit issues a temperature abnormality alarm, generates temperature abnormality information, and sends the temperature abnormality information to the central control module;

[0084] Step S104, based on the subtraction operational amplifier circuit formed by the operational amplifier, connecting two input terminals of the subtraction operational amplifier circuit to the positive electrode and the negative electrode of the power supply respectively;

[0085] Step S105, collecting the output voltage of the subtraction operational amplifier circuit, calculating the output voltage U of the power supply by the principle formula of the subtraction operational amplifier circuit, and sending the output voltage U to the data management module;

[0086] Step S106, setting a safe voltage range based on voltage usage parameters;

[0087] Step S107, when the voltage monitoring unit detects that the output voltage U exceeds the safe voltage range, the voltage monitoring unit issues a voltage abnormality alarm, generates voltage abnormality information, and sends the voltage abnormality information to the central control module;

[0088] Step S108: Connect a resistor R in series to the positive path of the power output terminal. i The resistor is marked as sampling resistor;

[0089] Step S109, using a subtraction amplifier circuit to obtain the voltage U of the sampling resistor i , according to Ohm's law The current monitoring unit obtains the output current I of the power supply in real time and sends the output current to the data management module;

[0090] Step S110, setting a safe current range based on power usage parameters;

[0091] Step S111, when the temperature monitoring unit detects that the magnitude of the power supply output current I exceeds the safe current range, the current monitoring unit issues a current abnormality alarm, generates current abnormality information, and sends the current abnormality information to the central control module.

[0092] Step S2, processing and storing the power monitoring information; Step S2 includes the following sub-steps:

[0093] Step S201, calculating the output power of the power supply based on the output voltage and the output current to obtain the output power P;

[0094] Step S202, calculating the output power based on the output power to obtain the output power W;

[0095] Step S203: storing the power monitoring information acquired by the power monitoring module and the output power P and output power W obtained by the data processing unit.

[0096] Step S3, building an intelligent analysis model based on the DQN model, and iteratively training the intelligent analysis model using power monitoring information; Step S3 includes the following sub-steps:

[0097] Step S301, based on the DQN model, the activation function ReLU function of the DQN model is replaced with the PReLU function;

[0098] Step S302, marking the improved DQN model as an intelligent optimization model, and outputting the intelligent optimization model to the analysis unit;

[0099] Step S303, setting the power supply operation cost to C, and setting minimizing the power supply operation cost as the intelligent management target;

[0100] Step S304, constructing a state space: defining the state space as S, the state space S includes: the output power P of the power supply, the maximum output power P of the power supply m and electricity price V, the state space S is represented by S = {P, P m , V}, S t represents the state space at time t;

[0101] Step S305, constructing an action space: defining the action space as A, which is represented by A={P i}, P i Represents the set of output powers of the power supply at different output currents;

[0102] Step S306, setting a reward function: based on the intelligent management target, setting a reward function R, through which the intelligent optimization model learns the action of minimizing the power supply operation cost under the premise of ensuring the normal operation of the power supply;

[0103] Step S307, initializing the parameters of the intelligent optimization model, setting the experience replay pool and related hyperparameters;

[0104] Step S308, take the state space s1 of the current power supply, use the ε-greedy strategy to select an action space with probability ε, mark it as a1, execute the selected action space a1, the state space of the power supply changes, obtain the next state space of the power supply, mark it as s2, and calculate the reward r, put the state space s1, action space a1, next state space s2 and reward r into the experience replay pool; while the intelligent optimization model interacts with the power supply, the intelligent optimization model uses the data in the experience replay pool for training iteration;

[0105] Step S309, generating a power management strategy based on the intelligent optimization model that has completed training iterations, and outputting the power management strategy to the central control module.

[0106] Step S4, generating a power management strategy based on the trained intelligent analysis model, and generating a power control instruction based on the power management strategy and power abnormality information to control and manage the power supply; Step S4 includes the following sub-steps:

[0107] Step S401, when the instruction generation unit receives power abnormality information, it generates a power off instruction and outputs the power off instruction to the instruction execution unit, the power abnormality information includes: temperature abnormality information, voltage abnormality information and current abnormality information;

[0108] Step S402, when the instruction generation unit receives the power management policy, it generates a power adjustment instruction according to the power management policy and outputs it to the instruction execution unit;

[0109] Step S403, when the instruction execution unit receives the power off instruction, the power off operation is performed on the power supply;

[0110] Step S404, when the instruction execution unit receives the power adjustment instruction, the power is adjusted accordingly based on the power adjustment instruction, and the state space of the adjusted power is output to the AI ​​intelligent module.

[0111] Through the description of the above implementation methods, the embodiments of the present invention can be provided as methods, systems or computer program products. Based on such an understanding, the above technical solutions can be essentially or partly contributed to the prior art in the form of software products, which can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and include several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0112] In the embodiments provided in the present application, it should be understood that the disclosed system or method can be implemented in other ways. The embodiments described above are merely schematic. For example, the division of modules or units is only a logical function division. There may be other division methods in actual implementation. For example, multiple modules or units can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of systems, modules and units can be electrical, mechanical or other forms.

[0113] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. Based on the AI ​​intelligent management system of power module, it is characterized by: include: Power monitoring module, data management module, AI intelligent module and central control module; The power supply monitoring module is used to obtain power supply monitoring information, and perform abnormal analysis on the power supply monitoring information to obtain power supply abnormality information, where the power supply monitoring information includes temperature information, voltage information, and current information of the power supply; The data management module includes a data processing unit and a data storage unit, the data processing unit is used to process the power monitoring information, and the data storage unit is used to store data; The AI ​​intelligent module includes a construction unit and an analysis unit. The construction unit constructs an intelligent optimization model based on the DQN model, and constructs the state space, action space and reward function of the intelligent optimization model; the analysis unit includes an intelligent optimization model, and the intelligent optimization model is used to generate a power management strategy; The central control module includes an instruction generation unit and an instruction execution unit. The instruction generation unit generates a power control instruction based on the power management strategy and power abnormality information and transmits it to the instruction execution unit. The instruction execution unit controls and manages the power based on the power control instruction.

2. The AI ​​intelligent management system based on power module according to claim 1 is characterized in that: The power monitoring module includes a temperature monitoring unit, and the temperature monitoring unit is configured with a temperature monitoring strategy, which includes: The temperature sensor installed inside the power module collects the temperature information T of the power supply in real time and sends the temperature information to the data management module; The safety temperature is set based on the power usage parameters. When the temperature monitoring unit detects that the power temperature exceeds the safety temperature, the temperature monitoring unit issues a temperature abnormality alarm, generates temperature abnormality information, and sends the temperature abnormality information to the central control module.

3. The AI ​​intelligent management system based on power module according to claim 2 is characterized in that: The power supply monitoring module includes a voltage monitoring unit, which is configured with a voltage monitoring strategy. The voltage monitoring strategy includes: The voltage monitoring unit is based on a subtraction operational amplifier circuit composed of an operational amplifier. The two input terminals of the subtraction operational amplifier circuit are connected to the positive and negative electrodes of the power supply respectively, and the output voltage of the subtraction operational amplifier circuit is collected. The output voltage U of the power supply is calculated by the principle formula of the subtraction operational amplifier circuit. The voltage monitoring unit obtains the output voltage U of the power supply in real time and sends the output voltage to the data management module. The safe voltage range is set based on the voltage usage parameters. When the voltage monitoring unit detects that the output voltage U exceeds the safe voltage range, the voltage monitoring unit issues a voltage abnormality alarm, generates voltage abnormality information, and sends the voltage abnormality information to the central control module.

4. The AI ​​intelligent management system based on power module according to claim 3 is characterized in that: The power monitoring module includes a current monitoring unit, and the current monitoring unit is configured with a current monitoring strategy. The current monitoring strategy includes: The current monitoring unit is connected in series with a resistor R in the positive path of the power supply output end. i The resistor is marked as the sampling resistor; the voltage U of the sampling resistor is obtained by using the subtraction amplifier circuit. i , according to Ohm's law The current monitoring unit obtains the output current I of the power supply in real time and sends the output current to the data management module; The safe current range is set based on the power usage parameters. When the temperature monitoring unit detects that the power output current I exceeds the safe current range, the current monitoring unit issues a current abnormality alarm, generates current abnormality information, and sends the current abnormality information to the central control module.

5. The AI ​​intelligent management system based on power module according to claim 4 is characterized in that: The data management module is configured with data management strategies, which include: The data processing unit calculates the output power of the power supply based on the output voltage and the output current to obtain the output power P, and calculates the output power based on the output power to obtain the output power W; The data storage unit is used to store the power monitoring information obtained by the power monitoring module and the output power P and output power W obtained by the data processing unit.

6. The AI ​​intelligent management system based on power module according to claim 5 is characterized in that: The build unit is configured with a build strategy, which includes: Based on the DQN model, the activation function ReLU function of the DQN model is replaced by the PReLU function, the improved DQN model is marked as an intelligent optimization model, and the intelligent optimization model is output to the analysis unit.

7. The AI ​​intelligent management system based on power module according to claim 6 is characterized in that: The build strategy also includes: Set the power supply operation cost as C, and set minimizing the power supply operation cost as the intelligent management goal; Construct state space: define the state space as S, which includes: the output power P of the power supply, the maximum output power P of the power supply m and electricity price V, the state space S is represented by S = {P, P m , V}, S t represents the state space at time t; Construct action space: Define the action space as A, which is expressed as A = {P i }, P i Represents the set of output powers of the power supply at different output currents; Set reward function: Based on the intelligent management goal, set the reward function R. Through the reward function R, the intelligent optimization model learns the action of minimizing the power supply operation cost while ensuring the normal operation of the power supply.

8. The AI ​​intelligent management system based on power module according to claim 7 is characterized in that: The analysis unit includes an intelligent optimization model, which is used to generate power management strategies including: Initialize the parameters of the intelligent optimization model, set the experience replay pool and related hyperparameters; obtain the state space s1 of the current power supply, use the ε-greedy strategy to select an action space with probability ε, marked as a1, execute the selected action space a1, the state space of the power supply changes, obtain the next state space of the power supply, marked as s2, and calculate the reward r, put the state space s1, action space a1, next state space s2 and reward r into the experience replay pool; while the intelligent optimization model power supply interacts, the intelligent optimization model uses the data in the experience replay pool for training iterations; generate a power management strategy based on the intelligent optimization model that has completed the training iteration, and output the power management strategy to the central control module.

9. The AI ​​intelligent management system based on power module according to claim 8 is characterized in that: The instruction generation unit is configured with an instruction generation strategy, which includes: The power control instructions include power adjustment instructions and power off instructions; When the instruction generation unit receives power abnormality information, it generates a power off instruction and outputs the power off instruction to the instruction execution unit. The power abnormality information includes: temperature abnormality information, voltage abnormality information and current abnormality information; When the instruction generation unit receives the power management policy, it generates a power adjustment instruction according to the power management policy and outputs the instruction to the instruction execution unit.

10. The AI ​​intelligent management system based on power module according to claim 9 is characterized in that: The instruction execution unit is configured with an instruction execution strategy, which includes: When the instruction execution unit receives a power-off instruction, it shuts down the power; when the instruction execution unit receives a power-adjustment instruction, it adjusts the power accordingly based on the power-adjustment instruction, and outputs the state space of the adjusted power to the AI ​​intelligent module.

Images