Flywheel energy storage monitoring system

Abstract

The invention relates to a flywheel energy storage monitoring system, the system comprises a sensing detection unit, a signal processing unit, a PLC control unit and a converter which are integrated in the same control cabinet, the sensing detection unit is connected with the PLC control unit through the signal processing unit, the converter is connected with the PLC control unit through a control cable, and the PLC control unit is connected with the converter through the control cable. And the PLC control unit is used for transmitting control and feedback signals and acquiring, processing and controlling flywheel operation parameters, and the PLC control unit is internally provided with a long-short-term memory neural network algorithm based on a racoon dog algae algorithm to predict an abnormal operation state value of the flywheel energy storage motor and outputs data information of the abnormal operation state value of the flywheel energy storage motor. The motor health state monitoring accuracy can be improved, the problem of frequency adjustment lag of a traditional energy storage system is solved, and the frequency stability of a power system is remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of flywheel energy storage technology, and in particular to a flywheel energy storage monitoring system. Background Technology

[0002] Flywheel energy storage systems are physical energy storage devices that store electrical energy in the form of kinetic energy by driving a flywheel rotor to high speed using power electronic devices. They possess outstanding advantages such as high power density, long cycle life, fast response speed, and no environmental pollution, and have become one of the core equipment in fields such as power quality regulation and grid frequency support. However, as flywheel systems develop towards higher speeds and greater power densities, their operating states become increasingly complex, posing extreme challenges to system reliability, safety, and lifespan prediction. Existing monitoring solutions typically have the following defects and shortcomings: Existing monitoring technologies cannot simultaneously collect multiple key parameters during motor operation, such as temperature, vibration, noise, and vacuum level. They can only monitor individual parameters in isolation, such as focusing only on vibration, and cannot establish a comprehensive overview of the motor's health status. This makes it difficult for maintenance personnel to fully and accurately grasp the actual operating condition of the motor and to promptly detect potential faults.

[0003] Traditional monitoring systems typically employ simple threshold alarm methods, lacking the ability to analyze and predict parameter change trends. For example, for parameters such as bearing temperature, they cannot provide early warnings based on the rate of increase; they only issue alarm signals when the temperature exceeds a preset threshold. By this time, the fault may have already progressed to a more serious stage, making it impossible to detect and take action in the early stages, thus increasing the risk of equipment failure and maintenance costs.

[0004] Traditional flywheel energy storage monitoring systems do not support remote monitoring via mobile devices, preventing maintenance personnel from accessing equipment operating status information anytime, anywhere. Furthermore, alarm messages typically only indicate abnormal parameters without providing specific troubleshooting suggestions, significantly complicating troubleshooting efforts and increasing both time and complexity. Summary of the Invention

[0005] In view of the shortcomings of the prior art, the present invention provides a flywheel energy storage monitoring system, which not only improves the accuracy of monitoring the health status of the motor, but also solves the problem of frequency regulation lag in traditional energy storage systems, and significantly improves the frequency stability of the power system.

[0006] To achieve the above and other related objectives, the present invention provides the following technical solution: A flywheel energy storage monitoring system includes a sensing and detection unit, a signal processing unit, a PLC control unit, and a converter integrated in the same control cabinet. The sensing and detection unit and the PLC control unit are connected via the signal processing unit. The sensing and detection unit includes a temperature sensor, a vibration sensor, a noise sensor, and a vacuum sensor. The sensing and detection unit transmits the collected analog signals to the signal processing unit. After preprocessing by the signal processing unit, the signals are transmitted to the PLC control unit in digital form via an industrial communication bus. The converter is connected to the PLC control unit via a control cable for transmitting control and feedback signals, collecting, processing, and controlling the flywheel operating parameters. The PLC control unit has a built-in long short-term memory neural network algorithm based on the Algae Algorithm to predict the abnormal operating state values ​​of the flywheel energy storage motor and outputs the data information of the abnormal operating state values ​​of the flywheel energy storage motor.

[0007] Furthermore, the prediction of abnormal operating state values ​​of the flywheel energy storage motor based on the raccoon algae algorithm-optimized long short-term memory neural network algorithm includes: L1. Real-time acquisition of temperature data of flywheel energy storage motor based on temperature sensor, real-time acquisition of vibration frequency data of flywheel energy storage motor based on vibration sensor, real-time acquisition of operating noise data of flywheel energy storage motor based on noise sensor, and real-time acquisition of cavity vacuum state parameters of flywheel energy storage motor based on vacuum sensor. L2. Input the data information of temperature, vibration frequency, operating noise and cavity vacuum state of the flywheel energy storage motor into the long short-term memory neural network model to initialize the weight parameters of the input gate, forget gate and output gate of the model, and obtain the data information of the weight parameters of the input gate, forget gate and output gate of the initialized model; L3. Based on the weight parameters of the input gate, forget gate, and output gate of the initialized model, the *Alternaria tectoris* population is initialized, the parameters of individual populations are determined, and the data information of the initialized *Alternaria tectoris* population is obtained. L4. Based on the data information of the initialized *Rhizophora stylosa* population, establish the fitness value function W_R for individual *Rhizophora stylosa* individuals in the population. , Where, x i The parameters of the i-th individual in the initialized raccoon algae population are given by α, β and σ, which are any constant parameters between 0 and 1, and n is the sample size. The fitness values ​​of the individuals in the raccoon algae population are calculated to obtain the fitness value data of the individuals in the raccoon algae population. Individuals with fitness values ​​within the preset range (p1, p2) are selected to update the raccoon algae population. L5. Based on the updated *Rhinoceros characinus* population, establish the objective function QF. , Among them, y k To optimize the parameters of the kth individual in the updated *Rhizophora stylosa* population, the output gate, forget gate, and output gate weight parameters of the model corresponding to the individual are optimized to obtain the optimized long short-term memory neural network model.

[0008] Furthermore, the method of predicting abnormal operating state values ​​of flywheel energy storage motors using the raccoon algae algorithm-optimized long short-term memory neural network algorithm also includes: L6. Based on the optimized long short-term memory neural network model, an optimized long short-term memory neural network model is obtained. The temperature, vibration frequency, operating noise and cavity vacuum state data of the flywheel energy storage motor are input, the abnormal operating state value of the flywheel energy storage motor is predicted, and the abnormal operating state value data of the flywheel energy storage motor is output.

[0009] Furthermore, the constraints on the constant parameters α, β, and σ are as follows: .

[0010] Furthermore, based on the data information of the abnormal operating status value of the flywheel energy storage motor, a preset safety threshold is set. If the abnormal operating status value of the flywheel energy storage motor is greater than the safety threshold, the flywheel energy storage motor is in an abnormal operating state and needs to be maintained and repaired in a timely manner. If the abnormal operating status value of the flywheel energy storage motor is less than the safety threshold, the motor is operating normally.

[0011] Furthermore, the sensing and detection unit includes sensors deployed on the flywheel body and motor. Each sensor is physically connected to the analog input module channel of the signal processing unit hardware interface through shielded cables. The signal processing unit has a built-in signal conditioning circuit that filters, amplifies, and converts the weak electrical signals output by the sensors to analog-to-digital conversion. Then, it establishes data interaction with the PLC control unit through the industrial Ethernet bus to realize the real-time acquisition and transmission of flywheel operating parameters.

[0012] Furthermore, the sensing and detection unit, signal processing unit, and control unit adopt a distributed monitoring architecture; the sensors in the sensing and detection unit are equipped with standard aviation plug interfaces for plug-and-play connection; the electrical connections and mechanical structures of each unit in the distributed monitoring architecture comply with the electromagnetic compatibility requirements of industrial control equipment specified in the IEC 61131-2 standard.

[0013] Furthermore, the temperature sensor is a PT100 platinum resistance temperature sensor, which is embedded in the stator winding of the flywheel motor, the bearing housing, and the wall of the vacuum chamber. The measurement range is -50℃ to 200℃, the accuracy is ±0.1℃, and the measurement error introduced by the lead resistance is eliminated by a three-wire connection method.

[0014] Furthermore, the vibration sensor is an integrated piezoelectric vibration sensor, installed on the flywheel base and motor end cover, with bidirectional measurement capabilities in both vertical and horizontal directions, used to collect abnormal bearing vibration and mechanical resonance signals.

[0015] Furthermore, the noise sensor is a long rod type noise sensor, deployed inside the flywheel control cabinet, and combined with a spectrum analysis algorithm to achieve characteristic identification of mechanical abnormal noise; the vacuum sensor is a resistance vacuum gauge, integrated and installed at the exhaust port of the vacuum chamber, and has a temperature compensation function, used to accurately monitor the vacuum state of the chamber.

[0016] The present invention has the following positive effects: 1. This invention achieves comprehensive monitoring of all parameters of a flywheel motor stator, including stator temperature, bearing vibration, mechanical noise, and vacuum chamber status, by deploying a PT100 platinum resistance temperature sensor, a piezoelectric vibration sensor, a long rod-type noise sensor, and a capacitive thin-film vacuum gauge. It employs a long short-term memory neural network algorithm optimized based on the algae-based algorithm, which can accurately detect potential faults such as bearing anomalies, resonance risks, and vacuum level drops, solving the problems of single monitoring parameters and delayed anomaly identification in traditional methods.

[0017] 2. This invention simplifies system wiring and reduces costs by integrating multiple signal processing functions into a single signal processing unit. Its multi-level isolation design significantly improves the system's anti-interference capability and reliability, ensuring the accuracy and stability of signal transmission, making it highly suitable for complex industrial environments.

[0018] 3. This invention communicates with the converter in real time via an RS485 module, integrating flywheel parameters with the converter's DC-side voltage, current, and other status information to achieve coordinated monitoring of the "flywheel operating status - energy conversion process." The converter rapidly executes charging and discharging actions based on PLC control unit instructions, providing millisecond-level response to power system frequency fluctuations. This solves the problem of lag in frequency regulation in traditional energy storage systems and significantly improves power system frequency stability.

[0019] 4. This invention employs a tiered response design in its early warning execution unit: Level 1 early warning uses local audible and visual indicators and a buzzer to promptly report minor anomalies; Level 2 early warning prioritizes pushing fault codes and real-time data to the host computer. If the host computer fails to respond within a timeout period, electromagnetic braking is directly triggered, forming a closed loop of "active early warning - automatic protection." This mechanism effectively prevents fault escalation, reduces the risk of equipment damage, and ensures the long-term safe and stable operation of the flywheel energy storage system. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the system framework of the present invention; Figure 2This is a flowchart illustrating the optimization of the long short-term memory neural network algorithm based on the raccoon algae algorithm of the present invention. Figure 3 This is a flowchart of the PLC control unit and early warning logic of the present invention; Figure 4 This is a flowchart of the converter charging and discharging control process of the present invention. Detailed Implementation

[0021] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0022] Example 1: As Figure 1 As shown, a flywheel energy storage monitoring system includes a sensing and detection unit, a signal processing unit, a PLC control unit, and a converter integrated in the same control cabinet. The sensing and detection unit and the PLC control unit are connected through the signal processing unit. The sensing and detection unit includes a temperature sensor, a vibration sensor, a noise sensor, and a vacuum sensor. The sensing and detection unit transmits the collected analog signals to the signal processing unit. After preprocessing by the signal processing unit, the signals are transmitted to the PLC control unit in digital form via an industrial communication bus. The converter is connected to the PLC control unit via a control cable for transmitting control and feedback signals, collecting, processing, and controlling the flywheel operating parameters. The PLC control unit has a built-in long short-term memory neural network algorithm based on the Algae Algorithm to predict the abnormal operating state values ​​of the flywheel energy storage motor and outputs the data information of the abnormal operating state values ​​of the flywheel energy storage motor.

[0023] In this embodiment, as Figure 2 As shown, the prediction of abnormal operating state values ​​of the flywheel energy storage motor based on the raccoon algae algorithm-optimized long short-term memory neural network algorithm includes: L1. Real-time acquisition of temperature data of flywheel energy storage motor based on temperature sensor, real-time acquisition of vibration frequency data of flywheel energy storage motor based on vibration sensor, real-time acquisition of operating noise data of flywheel energy storage motor based on noise sensor, and real-time acquisition of cavity vacuum state parameters of flywheel energy storage motor based on vacuum sensor. L2. Input the data information of temperature, vibration frequency, operating noise and cavity vacuum state of the flywheel energy storage motor into the long short-term memory neural network model to initialize the weight parameters of the input gate, forget gate and output gate of the model, and obtain the data information of the weight parameters of the input gate, forget gate and output gate of the initialized model; L3. Based on the weight parameters of the input gate, forget gate, and output gate of the initialized model, the *Alternaria tectoris* population is initialized, the parameters of individual populations are determined, and the data information of the initialized *Alternaria tectoris* population is obtained. L4. Based on the data information of the initialized *Rhizophora stylosa* population, establish the fitness value function W_R for individual *Rhizophora stylosa* individuals in the population. , Where, x i The parameters of the i-th individual in the initialized raccoon algae population are given by α, β and σ, which are any constant parameters between 0 and 1, and n is the sample size. The fitness values ​​of the individuals in the raccoon algae population are calculated to obtain the fitness value data of the individuals in the raccoon algae population. Individuals with fitness values ​​within the preset range (p1, p2) are selected to update the raccoon algae population. L5. Based on the updated *Rhinoceros characinus* population, establish the objective function QF. , Among them, y k To optimize the parameters of the kth individual in the updated *Rhizophora stylosa* population, the output gate, forget gate, and output gate weight parameters of the model corresponding to the individual are optimized to obtain the optimized long short-term memory neural network model.

[0024] In this embodiment, the prediction of abnormal operating state values ​​of the flywheel energy storage motor based on the raccoon algae algorithm-optimized long short-term memory neural network algorithm further includes: L6. Based on the optimized long short-term memory neural network model, an optimized long short-term memory neural network model is obtained. The temperature, vibration frequency, operating noise and cavity vacuum state data of the flywheel energy storage motor are input, the abnormal operating state value of the flywheel energy storage motor is predicted, and the abnormal operating state value data of the flywheel energy storage motor is output.

[0025] In this embodiment, the constraints on the constant parameters α, β, and σ are as follows: .

[0026] In this embodiment, based on the data information of the abnormal operating status value of the flywheel energy storage motor, a preset safety threshold is set. If the abnormal operating status value of the flywheel energy storage motor is greater than the safety threshold, the flywheel energy storage motor is in an abnormal operating state and needs to be maintained and repaired in a timely manner. If the abnormal operating status value of the flywheel energy storage motor is less than the safety threshold, the motor is operating normally.

[0027] In this embodiment, the sensing and detection unit includes sensors deployed on the flywheel body and motor. Each sensor is physically connected to the analog input module channel of the signal processing unit hardware interface through shielded cables. The signal processing unit has a built-in signal conditioning circuit that filters, amplifies, and converts the weak electrical signals output by the sensors to analog-to-digital conversion. Then, it establishes data interaction with the PLC control unit through the industrial Ethernet bus to realize the real-time acquisition and transmission of flywheel operating parameters.

[0028] In this embodiment, the sensing and detection unit, signal processing unit, and control unit adopt a distributed monitoring architecture; the sensors in the sensing and detection unit are equipped with standard aviation plug interfaces for plug-and-play connection; the electrical connections and mechanical structures of each unit in the distributed monitoring architecture comply with the electromagnetic compatibility requirements of industrial control equipment specified in the IEC 61131-2 standard.

[0029] In this embodiment, the temperature sensor is a PT100 platinum resistance temperature sensor, which is embedded in the stator winding of the flywheel motor, the bearing housing, and the wall of the vacuum chamber. The measurement range is -50℃ to 200℃, the accuracy is ±0.1℃, and a three-wire connection method is used to eliminate the measurement error introduced by the lead resistance.

[0030] In this embodiment, the vibration sensor is an integrated piezoelectric vibration sensor, which is installed on the flywheel base and the motor end cover. It has bidirectional measurement capabilities in both vertical and horizontal directions and is used to collect abnormal vibration and mechanical resonance signals of the bearing.

[0031] In this embodiment, the noise sensor is a long rod type noise sensor, which is deployed inside the flywheel control cabinet and combined with a spectrum analysis algorithm to identify the characteristics of mechanical abnormal noise; the vacuum sensor is a resistance vacuum gauge, which is integrated into the exhaust port of the vacuum chamber and has a temperature compensation function to accurately monitor the vacuum status of the chamber.

[0032] Example 2: Based on the flywheel energy storage monitoring system of Example 1, the present invention will be further described and explained below.

[0033] like Figure 1 or Figure 3 or Figure 4As shown, the flywheel energy storage monitoring system includes a sensing and detection unit, a signal processing unit, a PLC control unit, and a converter integrated in the same control cabinet. All parts work together, and the sensing and detection unit and the PLC control unit are connected through the signal processing unit.

[0034] The control cabinet adopts a layered structure design, consisting of two separable layers. The functional areas of the upper and lower layers are strategically laid out to ensure efficient collaboration and independent maintenance of each module. The upper cabinet integrates a signal processing unit, a PLC control unit, and a converter. Each unit is fixed to the upper cabinet mounting plate via a dedicated mounting bracket. The signal processing unit and the PLC control unit are connected via shielded twisted-pair cables. The RS485 communication module of the PLC control unit is connected to the communication port of the converter via shielded twisted-pair cables. The cables between the devices are arranged in an orderly manner through the internal cable trays of the cabinet, ensuring the compact integration of the upper-level low-voltage control and converter devices.

[0035] The lower cabinet houses the flywheel energy storage motor, vacuum pump, and sensor detection unit. The flywheel energy storage motor and vacuum pump are fixed to the bottom of the lower cabinet via a shock-absorbing base. The vacuum pump is connected to the exhaust port flange of the flywheel vacuum chamber. Various sensors of the sensor detection unit are deployed on the motor in preset positions. The sensor cables are led through the wiring holes to the upper and lower connection interfaces and connected to the input end of the upper signal processing unit. The bottom of the lower cabinet is equipped with reinforcing ribs to improve structural stability and meet the load-bearing and shock-absorbing requirements when the motor rotates at high speed.

[0036] The system deploys a series of sensors for temperature, vibration, noise, and vacuum levels to collect flywheel motor operating parameters in real time. The specific configurations of each sensor are as follows: The temperature sensor uses a PT100 platinum resistance temperature sensor, which is embedded in three symmetrical phase points of the flywheel motor stator winding, the outer walls of the front and rear bearing housings, and the inner wall of the vacuum chamber using specialized fixtures. The sensor and the measured part are filled with thermally conductive silicone grease to ensure efficient heat transfer. A three-wire connection is used to connect to the signal processing unit to eliminate the influence of lead resistance on measurement accuracy, covering a measurement range of -50℃ to 200℃. The vibration sensor is a piezoelectric accelerometer, rigidly fixed to the flywheel base in the horizontal (X-axis) and vertical (Z-axis) directions and the radial (Y-axis) position of the motor end cover via threads. The range is set to ±50g, and the frequency response covers 1Hz to 10kHz. To accurately capture abnormal bearing vibration and rotor resonance signals; the noise sensor adopts a long rod-type noise transmitter, which is installed inside the flywheel cabinet (away from interference sources such as cooling fans), and is wrapped with sound insulation cotton to reduce the impact of environmental noise. The output signal is connected to the signal processing unit via cable, and works with the spectrum analysis algorithm built into the PLC control unit to identify mechanical abnormal noise; the vacuum sensor adopts a capacitive thin film vacuum gauge, which is integrated into the exhaust port flange of the flywheel vacuum chamber and is connected to the inside of the chamber through a special sealing joint.

[0037] The signal processing unit adopts a modular design, with each module operating sequentially as follows: First, the signal input interface module, serving as the front end of the entire system, is responsible for receiving raw signals from field sensors. This module routes different types of signals to their corresponding processing channels, processes them, and then transmits them to the PLC control unit.

[0038] The three leads of the PT100 RTD sensor are connected to a dedicated three-wire measurement terminal, providing a foundation for subsequent accurate measurements. Then, the signal isolation and conditioning module receives the raw signal from the input interface module and performs crucial first-stage processing. Its operation is divided into two parallel and isolated paths: Then, the standardized analog voltage signal, after isolation and conditioning, is sent to the core processing module (MCU). This module is the "brain" of signal conversion and calculation. The standard analog signal (voltage) output by the MCU is then transmitted to the output driver module. The core task of this module is to convert this signal into an industry-standard signal and drive the output. Throughout this process, the power supply module provides power to all the above modules. It receives an external 24VDC power supply and, through DC-DC isolation conversion, generates multiple sets of mutually isolated stable power supplies. The PLC control unit is a programmable logic controller with a microprocessor as its core, which integrates multiple digital input / output channels and at least one serial communication interface. The serial communication interface preferably adopts the RS485 protocol and is connected to the corresponding communication port of the converter through a shielded twisted pair cable to obtain the operating status information of the converter in real time and send power adjustment commands to it. The control unit performs real-time analysis and processing of the received sensor data and converter status information, thereby realizing comprehensive monitoring of the operating status of the flywheel energy storage system and triggering corresponding early warning and protection mechanisms according to preset logic and thresholds. The specific deployment and response of the early warning execution unit are as follows: The primary early warning component consists of LED indicators mounted on the front panel of the flywheel cabinet, displaying green and yellow indicators (green for normal operation, yellow for warning, and a buzzer). Controlled via the digital output port of the PLC control unit, when a parameter exceeds the upper limit but does not reach the danger threshold, the yellow light illuminates and the buzzer sounds briefly at 1-second intervals. The secondary early warning component connects to an industrial switch via the Ethernet port of the PLC control unit, sending TCP / IP messages to the host computer. These messages include fault codes, real-time parameters, and time. A built-in timer has a preset response timeout of 3 minutes. If the host computer does not acknowledge the fault code signal, the PLC control unit directly sends commands to specific locations in the converter's control and status words. This can command the converter to "emergency stop," etc.

[0039] System operation process Taking the complete cycle of a flywheel energy storage system participating in power system frequency regulation as an example, the system operation process is illustrated as follows: After the system is powered on, the PLC control unit automatically acquires the sensor connection status and communication link connectivity, and loads preset parameters: safety thresholds for each sensor and the charging / discharging power range of the converter. After the self-test passes, the LED indicator lights up green, and the system enters standby mode.

[0040] When the power system frequency is higher than the rated value, the PLC control unit determines that electrical energy needs to be absorbed and sends a charging command to the converter. The converter converts the DC power supply into frequency-adjustable three-phase AC power to drive the permanent magnet synchronous motor, and the electrical energy is converted into kinetic energy of the flywheel and stored. When the power system frequency is lower than the rated value, the PLC control unit sends a "discharge command." The converter converts the kinetic energy of the flywheel rotation into electrical energy, inverts it, and feeds it back to the grid. The flywheel speed gradually decreases as the kinetic energy is released. During the above process, the sensing and detection unit collects parameter signals in real time, which are converted by the signal processing unit and transmitted to the PLC control unit. The PLC control unit analyzes the received digital signals.

[0041] For temperature data trend calculations, if the analysis reveals that the stator temperature has reached the safety threshold, a Level 1 warning is triggered: a yellow indicator light on the cabinet illuminates, a buzzer sounds intermittently, and the PLC control unit simultaneously sends a power reduction command to the inverter to slow the temperature rise. If the temperature continues to rise, a Level 2 warning is triggered: the PLC control unit sends a fault code and real-time temperature data to the host computer; if no confirmation signal is received from the host computer within 3 minutes, the PLC control unit directly sends commands to specific positions in the inverter's control and status words. This can include commands such as "emergency shutdown" of the inverter.

[0042] In summary, this invention not only improves the accuracy of monitoring the health status of motors, but also solves the problem of frequency regulation lag in traditional energy storage systems, significantly improving the frequency stability of power systems.

[0043] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A flywheel energy storage monitoring system, characterized by, The system comprises a sensing detection unit, a signal processing unit, a PLC control unit and a converter integrated in the same control cabinet, the sensing detection unit and the PLC control unit are connected through the signal processing unit, the sensing detection unit comprises a temperature sensor, a vibration sensor, a noise sensor and a vacuum degree sensor, the sensing detection unit transmits the collected analog signals to the signal processing unit, after preprocessing by the signal processing unit, the signals are transmitted to the PLC control unit in the form of digital signals through an industrial communication bus, the converter is connected with the PLC control unit through a control cable for transmitting control and feedback signals, and the PLC control unit collects, processes and controls the running parameters of the flywheel energy storage motor, the built-in long short-term memory neural network algorithm based on the mink algorithm in the PLC control unit is used to predict the abnormal state value of the flywheel energy storage motor, and data information of the abnormal state value of the flywheel energy storage motor is output.

2. The flywheel energy storage monitoring system of claim 1, wherein, The long short-term memory neural network algorithm based on the mink algorithm is used to predict the abnormal state value of the flywheel energy storage motor, and comprises the following steps: L1. Real-time acquisition of the temperature data information of the flywheel energy storage motor based on the temperature sensor, real-time acquisition of the vibration frequency data information of the flywheel energy storage motor based on the vibration sensor, real-time acquisition of the running noise data information of the flywheel energy storage motor based on the noise sensor, and real-time acquisition of the cavity vacuum state parameter data information of the flywheel energy storage motor based on the vacuum degree sensor; L2. Inputting the temperature, vibration frequency, running noise and cavity vacuum state data information of the flywheel energy storage motor into the long short-term memory neural network model to initialize the weight parameters of the input gate, the forgetting gate and the output gate of the model, and obtaining the data information of the weight parameters of the input gate, the forgetting gate and the output gate of the initialized model; L3. Based on the data information of the weight parameters of the input gate, the forgetting gate and the output gate of the initialized model, initializing the mink population, determining the parameters of the population individuals, and obtaining the data information of the initialized mink population; L4. Based on the data information of the initialized mink population, establishing the fitness value function W_R of the mink population individuals, ， wherein, x i is the parameter of the i-th population individual of the initialized raccoon-bellwort population, a, β and σ are any constant parameters between 0 and 1, n is the sample capacity, the fitness value of the raccoon-bellwort population individual is calculated, the data information of the fitness value of the raccoon-bellwort population individual is obtained, and the individual with the fitness value in the preset range (p1, p2) is selected to update the raccoon-bellwort population; L5. Based on the updated mink population, establishing the target function QF, ， Wherein, y k The weight parameters of the output gate, the forget gate and the output gate of the model corresponding to the population individual are optimized, and an optimized long short-term memory neural network model is obtained.

3. The flywheel energy storage monitoring system of claim 2, wherein: The long short-term memory neural network algorithm based on the mink algorithm is used to predict the abnormal state value of the flywheel energy storage motor, and comprises the following steps: L6. Based on the optimized long short-term memory neural network model, obtaining the optimized long short-term memory neural network model, inputting the temperature, vibration frequency, running noise and cavity vacuum state data information of the flywheel energy storage motor, predicting the abnormal state value of the flywheel energy storage motor, and outputting the data information of the abnormal state value of the flywheel energy storage motor.

4. The flywheel energy storage monitoring system of claim 2, wherein: The constraint conditions of the constant parameters a, β and σ are 。 5. The flywheel energy storage monitoring system of claim 1, wherein: Based on the data information of the abnormal state value of the flywheel energy storage motor, a preset safety threshold is set, if the abnormal state value of the flywheel energy storage motor is greater than the safety threshold, the flywheel energy storage motor is in an abnormal state, and timely maintenance and repair are required, and if the abnormal state value of the flywheel energy storage motor is less than the safety threshold, the motor is in a normal state.

6. The flywheel energy storage monitoring system of claim 1, wherein: The sensing detection unit includes flywheel body and motor. Each sensor is physically connected with the signal processing unit hardware interface analog input module channel through shielded cable. The signal processing unit is built-in signal conditioning circuit. After filtering, amplifying and analog-digital conversion of the weak electric signal output by the sensor, data interaction is established with the PLC control unit through industrial Ethernet bus to realize real-time collection and transmission of flywheel operation parameters.

7. The flywheel energy storage monitoring system of claim 1, wherein: The sensing detection unit, signal processing unit and control unit adopt distributed monitoring architecture. The sensors in the sensing detection unit are equipped with standard aviation plug interfaces and connected in plug-and-play mode. The electrical connection and mechanical structure of each unit in the distributed monitoring architecture meet the electromagnetic compatibility requirements of industrial control equipment specified in IEC 61131-2 standard.

8. The flywheel energy storage monitoring system of claim 1, wherein: The temperature sensor adopts PT100 platinum resistance temperature sensor, which is embedded in the flywheel motor stator winding, bearing seat and vacuum cavity wall, respectively. The measurement range is-50℃-200℃, the accuracy is ±0.1℃, and the three-wire connection mode is used to eliminate the measurement error introduced by lead resistance.

9. The flywheel energy storage monitoring system of claim 1, wherein: The vibration sensor selects an integrated piezoelectric vibration sensor, which is installed on the flywheel base and motor end cover, has vertical and horizontal bidirectional measurement capability, and is used to collect abnormal bearing vibration and mechanical resonance signals.

10. The flywheel energy storage monitoring system of claim 1, wherein: The noise sensor adopts a long rod type noise sensor, which is deployed inside the flywheel control cabinet. Combined with the frequency spectrum analysis algorithm, the characteristic identification of mechanical abnormal sound is realized. The vacuum degree sensor selects a resistance type vacuum gauge, which is integrated and installed at the exhaust port of the vacuum cavity and has temperature compensation function, and is used to accurately monitor the vacuum state of the cavity.

Images