Push-pull catkin cleaning device for low-voltage comprehensive distribution box shell and control system of catkin cleaning device

Through the push-pull catkin cleaning device and intelligent control system, the fire risk and poor ventilation problems of low-voltage comprehensive distribution boxes during catkin infringement are solved, efficient and intelligent catkin management is achieved, and cleaning efficiency and equipment life are improved.

CN119209273AInactive Publication Date: 2024-12-27JIANGSU QIHOU INTELLIGENT ELECTRICAL EQUIP CO LTD
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202411573147.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-11-06
Publication Date
2024-12-27
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Low-voltage integrated distribution boxes are susceptible to catkins in late spring and early summer, resulting in fire risks and poor ventilation. The existing cleaning methods are inefficient and inconvenient.

Method used

Design a push-pull catkin cleaning device and control system, including motor modules, brush fixing brackets, polyester silk braided brushes and multi-module collaborative control system, and optimize cleaning strategies using electrostatic adsorption and machine learning prediction models.

Benefits of technology

Effectively clean the catkins in the ventilation holes, reduce the temperature rise inside the box, improve the equipment life and cleaning efficiency, and improve the cleaning effect and user control experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119209273A_ABST
    Figure CN119209273A_ABST
Patent Text Reader

Abstract

The invention discloses a push-pull type catkin cleaning device for a low-voltage comprehensive distribution box shell and a control system of the push-pull type catkin cleaning device, and relates to the field of outdoor high-low-voltage complete switch equipment.The cleaning device comprises the distribution box shell, a fixed bottom plate is arranged at the bottom of the distribution box shell, and fixed guide rails are arranged on the two sides of the bottom end of the fixed bottom plate; a motor module is arranged at one end of each fixed guide rail, a brush fixing support is arranged between the two fixed guide rails, and a bottom plate push-pull type brush is arranged at the top end of the brush fixing support; the fixed bottom plate and the fixed guide rail are fixedly connected through a guide rail base, and a plurality of bottom plate ventilation holes are formed in the bottom of the fixed bottom plate. According to the low-voltage comprehensive distribution box adopting the push-pull type catkin cleaning device, on the basis that the functions of an existing low-voltage comprehensive distribution box are guaranteed, catkin at the ventilation holes can be effectively cleaned only by additionally arranging the cleaning device under the condition that the protection grade is not changed, and the purposes of reducing components in the box body or temperature rise in the box body and improving the working efficiency are achieved. The product service life is prolonged.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the field of outdoor high and low voltage complete switchgear, in particular to a push-pull type catkin cleaning device for a low voltage integrated distribution box housing and a control system thereof. Background Art

[0002] The low-voltage integrated distribution box (JP cabinet) is an integrated distribution product installed on the low-voltage side of the transformer, integrating power distribution, metering, protection, control, and reactive power compensation. It consists of three parts: switch room, metering room, and compensation room. The main components include circuit breakers, metering and measuring current transformers, electric energy meters and data acquisition terminals, distribution integrated control terminals, capacitor switching units, and self-healing low-voltage shunt capacitors. Low-voltage integrated distribution boxes have been widely used in the transformation of urban and new rural power distribution systems due to their advantages in power distribution, control, management, harmonic control, and reactive power compensation. This product can improve the power factor of the low-voltage distribution network, reduce losses, improve the power supply quality at the load end, and improve the utilization rate of transformers and transmission lines. It is of great significance in terms of energy saving and loss reduction, strengthening the security and stable operation of the power system network, and improving the economic benefits of users.

[0003] In late spring and early summer every year, the flying catkins are full of grease, which is a flammable substance. They are very light and drift with the wind, especially drifting into the distribution box and drilling into the distribution cabinet. Once the accumulated catkins encounter an open flame, they burn very quickly and can easily cause a fire. At the same time, catkins can easily block the ventilation holes, resulting in poor ventilation inside the box, which can easily cause the temperature of the components inside the box or the overall temperature inside the box to rise, resulting in an increase in the temperature of the internal components or the box, affecting the service life of the components and the entire cabinet, and causing serious harm to the safe operation of the power grid equipment. The current common distribution cabinets are not equipped with specially designed cleaning mechanisms. They are cleaned by manual cleaning, which is easy to cause catkins to fly, or by sprinkling water, which is easy to cause catkins to adhere to the ground and continue to fly after drying. The cleaning effect is not good, and the catkins after cleaning will also be inconvenient in the subsequent collection and processing process.

[0004] In view of the high frequency use of low-voltage integrated distribution boxes in urban and rural power grids and the adverse effects of willow catkins on low-voltage integrated distribution box equipment, it is necessary to design a willow catkin cleaning device that meets the safety requirements of the low-voltage integrated distribution box shell. Summary of the invention

[0005] Based on this, it is necessary to provide a push-pull catkin cleaning device and a control system for the low-voltage integrated distribution box shell in response to the above-mentioned technical problems.

[0006] In the first aspect, the present invention provides a push-pull willow catkin cleaning device for a low-voltage integrated distribution box shell, comprising a distribution box shell, a fixed bottom plate is arranged at the bottom of the distribution box shell, fixed guide rails are arranged on both sides of the bottom end of the fixed bottom plate, a motor module is arranged at one end of the fixed guide rail, a brush fixing bracket is arranged between the two fixed guide rails, and a bottom plate push-pull brush is arranged on the top of the brush fixing bracket; the fixed bottom plate and the fixed guide rail are maintained in fixed connection through the guide rail base, and a plurality of bottom plate ventilation holes are opened at the bottom of the fixed bottom plate.

[0007] Furthermore, the motor module includes a steering gear set arranged at one end of the fixed guide rail, a driving shaft is arranged between the two steering gear sets, a stepping motor is arranged on one side of one of the steering gear sets, and an output end of the stepping motor is connected to the driving shaft;

[0008] A driving screw matched with the steering gear set is arranged inside the fixed guide rail, and the bottom of the fixed guide rail is connected with both sides of the brush fixing bracket by means of crawler tracks.

[0009] Furthermore, the surface of the bottom plate push-pull brush is made of polyester woven material, and the surface of the polyester woven material is smooth and has polarity, which is used to generate electrostatic adsorption.

[0010] In a second aspect, the present invention also provides a control system for a push-pull type catkin cleaning device for a low-voltage integrated distribution box housing, the control system comprising: a power supply module, a local control module, a collection and monitoring module, a remote control module, a feedback adjustment module and a wireless communication module;

[0011] Among them, the power supply module is used to connect the power supply line inside the low-voltage integrated distribution box, support multiple power supply modes, provide power supply for the motor module, and monitor the power supply status;

[0012] The local control module is used to set the seasonal initial working time, multiple cleaning modes and corresponding control parameters according to the location of the low-voltage integrated distribution box and the catkin production cycle, and configure the user control interface to provide manual adjustment of the cleaning mode and control parameters;

[0013] The collection and monitoring module is used to configure the distributed monitoring stations in the region and obtain the catkin concentration and meteorological data collected in real time by each monitoring station, and the meteorological data includes temperature, humidity, wind direction, wind speed and precipitation;

[0014] The remote control module is used to predict the concentration and duration of catkins in the current area at a future time based on the real-time collected catkins concentration and meteorological data through data analysis and processing of the catkins prediction model, optimize the catkins cleaning strategy, and issue remote control instructions to the feedback regulation module. The catkins prediction model includes a concentration prediction model and a time prediction model;

[0015] The feedback adjustment module is used to drive and adjust the cleaning mode and control parameters set by the local control module according to the received remote control instructions, and monitor the equipment operation status and catkin cleaning effect in real time, and further optimize the control parameters according to the monitoring feedback information;

[0016] The wireless communication module is used to realize wireless transmission between modules through wireless communication protocols.

[0017] Further, the local control module includes: a timing drive unit, a mode setting unit, a parameter configuration unit, a control interface unit and a local monitoring unit;

[0018] Among them, the timing drive unit is used to activate the motor module to perform the catkin cleaning operation according to the preset initial working time when the catkins are flying, and set the start time of the cleaning operation through the built-in timer to achieve regular opening and closing;

[0019] A mode setting unit is used to set the working mode that the motor module needs to execute during operation, and after the working mode is set, the catkin cleaning operation is performed according to the cleaning strategy pre-configured in the respective working modes. The working modes include manual mode, timing mode and intelligent mode, and the cleaning strategies include high-intensity cleaning operation, medium-intensity cleaning operation and low-intensity cleaning operation;

[0020] The parameter configuration unit is used to set the control parameters of the motor module. Under different working modes and cleaning strategies, different control parameters are used to perform cleaning operations of different intensities. The control parameters include cleaning frequency, push-pull speed and cleaning time.

[0021] The control interface unit is used to provide a user interaction interface, access the operation display interface of the manual mode, and display real-time data, device status and current mode information;

[0022] The local monitoring unit is used to monitor the operating status of the motor module and the cleaning effect of the cleaning operation in real time, and transmit the monitoring data to the feedback adjustment module and the remote control module.

[0023] Furthermore, the remote control module includes: a data transmission unit, a concentration prediction unit, a time prediction unit, a strategy optimization unit and a remote control unit;

[0024] The data transmission unit is used to obtain the catkin concentration and meteorological data collected by the monitoring site, and record and manage them according to the time information and site location information;

[0025] The concentration prediction unit is used to analyze the correlation of catkin concentrations at different monitoring sites, and to build a concentration prediction model based on the time series delay correlation algorithm in combination with the long short-term memory network to predict the catkin concentration at the location of each low-voltage integrated distribution box in the intelligent mode;

[0026] The time prediction unit is used to extract feature sets based on the meteorological data collected in real time, use the duration of willow catkins flying as a label, establish a time prediction model, and use the particle swarm optimization algorithm to optimize the hyperparameters of the time prediction model to output the duration of willow catkins flying;

[0027] A strategy optimization unit is used to optimize the catkin cleaning strategy of the drive motor in the low-voltage integrated distribution box according to the catkin concentration prediction results and the duration prediction results, adjust the intensity of the cleaning operation, and adjust the corresponding control parameters according to the change of the operation intensity;

[0028] The remote control unit is used to generate remote control instructions according to the strategy optimization results and send them to the feedback adjustment module through the wireless communication network.

[0029] Furthermore, the correlation of catkin concentrations at different monitoring sites was analyzed, and a concentration prediction model based on the time series delay correlation algorithm was constructed in combination with the long short-term memory network. The catkin concentrations at the locations of various low-voltage integrated distribution boxes in the intelligent mode were predicted, including:

[0030] Obtain the catkin concentration collected at each monitoring station, normalize it using the maximum-minimum normalization method, and convert it into a time series according to the monitoring time, with each monitoring station corresponding to a time series;

[0031] The catkin concentration to be predicted in the low-voltage integrated distribution box and the catkin concentration at the monitoring stations around the low-voltage integrated distribution box at the same time are set, and the input sample data matrix is ​​obtained after calculation and screening using the time series delay correlation algorithm;

[0032] The historical catkins concentration data were divided into training set and test set. The training set was used to establish a concentration prediction model based on long short-term memory network and train it. The input sample data matrix was then input into the concentration prediction model to output the catkins concentration at the location of the low-voltage integrated distribution box.

[0033] Furthermore, the catkin concentration to be predicted in the low-voltage integrated distribution box and the catkin concentration at the monitoring stations around the low-voltage integrated distribution box at the same time are set, and the input sample data matrix is ​​obtained after calculation and screening using the time series delay correlation algorithm, including:

[0034] Assume that at time t, the catkin concentration at the location of the low-voltage integrated power distribution station is C t, and the catkin concentration C collected by the surrounding i-th monitoring station at time t it ;

[0035] The delay time window is set, and the catkin concentration at the current time and the past N times is used as the input sample. The catkin concentration at the low-voltage integrated distribution box and each monitoring station is combined to form a sample data matrix.

[0036] For each monitoring station, the cross-correlation function is used to calculate the delayed correlation between different monitoring stations. By calculating the cross-correlation function between different monitoring stations, the monitoring station most relevant to the predicted catkin concentration is screened out. The expression of the cross-correlation function is:

[0037]

[0038] Where CCF i,j (τ) represents the correlation coefficient between the i-th and j-th monitoring stations; μ i and μ j Respectively represent the mean of the i-th and j-th monitoring stations; σ i and σ j They represent the standard deviation of the i-th and j-th monitoring stations respectively; E represents the expected value; τ represents the time delay;

[0039] An input sample data matrix is ​​formed based on the catkin concentrations at the screened monitoring sites.

[0040] Furthermore, based on the meteorological data collected in real time, the feature set is extracted, and the duration of willow catkins flying is used as a label to establish a time prediction model. The particle swarm optimization algorithm is used to optimize the hyperparameters of the time prediction model. The output duration of willow catkins flying includes:

[0041] Obtain historical meteorological data for normalization, scale the original data to a specific space, extract effective features of meteorological data, and divide them into meteorological training set and meteorological verification set;

[0042] The data window size, dropout ratio, number of hidden layer neurons and learning rate of the gated recurrent network are set as the hyperparameter optimization targets, and the predicted duration of catkins flying is substituted into the target fitness function. The particle swarm optimization algorithm is used to find the optimal hyperparameters of the gated recurrent network.

[0043] By configuring the optimized hyperparameters, a time prediction model based on a gated recurrent network is constructed. The processed meteorological data and concentration data are used as model inputs. After processing by the fully connected layer, the duration of willow catkin flying at the corresponding willow catkin concentration is output.

[0044] Furthermore, the data window size, discard ratio, number of hidden layer neurons and learning rate of the gated recurrent network are set as the hyperparameter optimization targets, and the predicted duration of catkins flying is substituted into the target fitness function. The particle swarm optimization algorithm is used to find the optimal hyperparameters of the gated recurrent network, including:

[0045] Set the initial parameters of the particle swarm optimization algorithm, including the number of particles, number of iterations, initial inertia weight and learning factor, and set the optimization range of the hyperparameters of the gated recurrent network and the particle velocity threshold, as well as the target fitness function;

[0046] Initialize the population randomly, determine the initial population and individual optimal values, use the particle group individual position and velocity update formula to obtain the latest position and velocity of the particle, and build a time prediction model based on the gated recurrent network for training according to the hyperparameters corresponding to the latest position of each particle, then return the model prediction value and substitute it into the target fitness function for verification, and generate the optimal population and individual values ​​of this iteration according to the fitness value of the particle;

[0047] Determine whether the iteration termination conditions are met. If not, continue to iterate to find the optimal solution. If yes, terminate the iteration and output the optimal hyperparameters of the time prediction model.

[0048] The beneficial effects of the present invention are:

[0049] 1. By adopting a low-voltage integrated distribution box with a push-pull catkin cleaning device, on the basis of ensuring the functions of the existing low-voltage integrated distribution box, as long as this cleaning device is added, while ensuring that the protection level remains unchanged, the catkins at the ventilation holes can be effectively cleaned, thereby reducing the temperature rise of components inside the box or inside the box, increasing the service life of the product, and ensuring the safe operation of power grid equipment.

[0050] 2. Through the integration of multi-module collaboration, efficient and intelligent willow catkin cleaning management is achieved. According to the seasonal changes in the production of catkins, the cleaning time and mode can be flexibly set, and dynamic adjustments based on weather and concentration forecasts are made to improve the cleaning efficiency and energy-saving effects. Remote control and wireless communication technology are used to optimize the user control experience, and machine learning models are used to improve the accuracy and adaptability of predictions, making the cleaning strategy more flexible and intelligent. The overall control system has stronger intelligent decision-making capabilities, adaptability and real-time response capabilities, which effectively reduces energy consumption and improves cleaning effects, greatly enhancing the scientificity and efficiency of catkin cleaning work. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0052] Figure 1 is a front view of a push-pull type catkin cleaning device for a low-voltage integrated distribution box housing according to an embodiment of the present invention;

[0053] Figure 2 2. It is a bottom view of a push-pull type catkin cleaning device for a low-voltage integrated distribution box housing according to an embodiment of the present invention;

[0054] Figure 3 It is a schematic diagram of the fixed guide rail and motor module structure of a push-pull type catkin cleaning device for a low-voltage integrated distribution box housing according to an embodiment of the present invention;

[0055] Figure 4 The present invention is a system principle block diagram of a control system of a push-pull type catkin cleaning device for a low-voltage integrated distribution box housing according to an embodiment of the present invention.

[0056] Figure numbers: 1. Distribution box housing; 2. Fixed bottom plate; 3. Fixed guide rail; 4. Motor module; 401. Steering gear set; 402. Driving coupling; 403. Stepping motor; 404. Driving screw; 5. Brush fixing bracket; 6. Bottom plate push-pull brush; 7. Guide rail base; 8. Bottom plate ventilation hole. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0058] See also Figure 1-Figure 3 , a push-pull type catkin cleaning device for a low-voltage integrated distribution box housing and a control system thereof are provided, including a distribution box housing 1, a fixed bottom plate 2 is arranged at the bottom of the distribution box housing 1, fixed guide rails 3 are arranged on both sides of the bottom end of the fixed bottom plate 2, a motor module 4 is arranged at one end of the fixed guide rail 3, a brush fixing bracket 5 is arranged between the two fixed guide rails 3, and a bottom plate push-pull type brush 6 is arranged at the top of the brush fixing bracket 5. The fixed bottom plate 2 and the fixed guide rail 3 are fixedly connected by a guide rail base 7, and a plurality of bottom plate ventilation holes 8 are opened at the bottom of the fixed bottom plate 2.

[0059] In addition, a certain number of ventilation holes 8 are evenly arranged on the fixed bottom plate 2 of the low-voltage integrated distribution box shell, a ventilation net is welded on the back of the ventilation hole 8, and a dust filter is welded on the ventilation and heat dissipation holes, which meets the protection level IP44 requirements. While ensuring smooth air circulation inside the box, it can effectively prevent catkins, branches or other external objects from entering the shell.

[0060] In the description of the present invention, the motor module 4 includes a steering gear set 401 disposed at one end of the fixed guide rail 3, a driving shaft 402 is disposed between the two steering gear sets 401, a stepper motor 403 is disposed on one side of one of the steering gear sets 401, and the output end of the stepper motor 403 is connected to the driving shaft 402. A driving screw 404 matched with the steering gear set 401 is disposed inside the fixed guide rail 3, and the bottom of the fixed guide rail 3 is connected to both sides of the brush fixing bracket 5 by a crawler.

[0061] Specifically, the motor module uses M5*25 bolts to fix the guide rail base 7 and the fixed base plate 2 together. The fixed guide rail 3 is above the guide rail base 7 and is controlled by the stepper motor 403 to drive the slider to move upward. The fixed guide rail 3 installed on the fixed base plate 2 of the low-voltage integrated distribution box shell adopts an aluminum profile self-lubricating cleaning system. The friction coefficient of the linear module is low during reciprocating motion, which reduces equipment maintenance and facilitates long-term use of the equipment; the guide rail surface is connected by a crawler track, which drives the brush to move stably and reliably; the screw rod is made of a solid round rod made of stainless steel to improve the motor operation accuracy; the end cover is made of cast iron and equipped with an anti-collision strip for buffering and protection.

[0062] The bottom plate push-pull brush 6 is integrally mounted on the fixed rail 3. The width of the bottom plate push-pull brush 6 is the same as the width of the ventilation hole 8. The bottom plate push-pull brush 6 is driven by the stepper motor 403 to move on the fixed rail 3. The bottom plate push-pull brush 6 is retractable and can be adjusted in width according to the width of the ventilation hole. When the single-axis stepper motor controller controls the angular displacement or linear displacement of the stepper motor by emitting an electric pulse signal, the stepper motor will rotate a fixed angle or move forward one step to carry the bottom plate push-pull brush to achieve the catkin cleaning operation.

[0063] In the description of the present invention, the surface of the bottom plate push-pull brush 6 is made of polyester yarn woven material, and the surface of the polyester yarn woven material is smooth and has polarity, which is used to generate electrostatic adsorption.

[0064] Specifically, the surface of the bottom plate push-pull brush 6 is woven from polyester yarn, which has a smooth surface and a certain polarity. This polarity makes it easy for the polyester yarn to accumulate charges, thus becoming a potential adsorption object of static electricity. When the polyester yarn contacts catkins or other foreign objects with opposite charges, charge transfer will occur, causing the polyester yarn to have corresponding charges. Static electricity has an adsorption force, which can attract objects with opposite charges. When the polyester yarn is charged, it will attract surrounding objects with opposite charges, thus causing static electricity adsorption.

[0065] The long-term operation of the bottom push-pull brush 6 will cause the accumulation of foreign matter such as catkins. Since the present invention is a bottom push-pull cleaning device, foreign matter such as catkins will fall off according to their own gravity and will not stick to the brush to affect the efficiency and effect of the next cleaning work.

[0066] See also Figure 4 , and also provides a control system for a push-pull willow catkin cleaning device for a low-voltage integrated distribution box shell. The control system includes: a power supply module, a local control module, a collection and monitoring module, a remote control module, a feedback adjustment module and a wireless communication module.

[0067] The power supply module is used to connect the power supply lines inside the low-voltage integrated distribution box, supports multiple power supply modes, provides power supply for the motor module, and monitors the power supply status.

[0068] Specifically, the power supply module is equipped with a switching power supply, which provides DC24V to the single-axis stepper motor controller. The controller transmits a pulse signal to the stepper motor, and the stepper motor drives the entire motor module to move up and down. The controller gives a signal to rotate a set distance in the target direction. After reaching the target distance, the motor reverses and moves in the opposite direction, thereby performing a reciprocating motion.

[0069] The local control module is used to set the seasonal initial working time, multiple cleaning modes and corresponding control parameters according to the region where the low-voltage integrated distribution box is located and the catkin production cycle, and configure the user control interface to provide manual adjustment of the cleaning mode and control parameters.

[0070] In the description of the present invention, the local control module includes: a timing drive unit, a mode setting unit, a parameter configuration unit, a control interface unit and a local monitoring unit.

[0071] Among them, the timing drive unit is used to activate the motor module to perform the willow catkin cleaning operation according to the preset initial working time when the catkins are flying, and set the starting time of the cleaning operation through the built-in timer to achieve regular opening and closing.

[0072] The mode setting unit is used to set the working mode that the motor module needs to execute during operation, and after the working mode is set, the catkin cleaning operation is performed according to the cleaning strategy pre-configured in the respective working mode. The working modes include manual mode, timing mode and intelligent mode, and the cleaning strategies include high-intensity cleaning operation, medium-intensity cleaning operation and low-intensity cleaning operation.

[0073] The following is an introduction to the working mode:

[0074] 1. Manual mode

[0075] Function: The user directly controls the start and stop of the cleaning device and adjusts the cleaning parameters in real time.

[0076] Implementation method: Through a user interface (such as a button or a touch screen), the user can manually start or stop the motor module and input specific parameters (such as cleaning time).

[0077] 2. Timing mode

[0078] Function: The system automatically starts the cleanup operation according to the set schedule.

[0079] Implementation method: A built-in timer (such as an RTC module) sets the start and end time of cleaning, and the system automatically activates the cleaning operation at the set time.

[0080] 3. Smart Mode

[0081] Function: The remote control module makes predictions based on real-time monitoring data (such as catkin concentration and meteorological conditions) and automatically selects a cleaning strategy that suits each low-voltage integrated distribution box.

[0082] Implementation method: Combined with the real-time data of the collection and monitoring module, intelligent algorithm analysis and selection of high, medium and low intensity cleaning operations are carried out to achieve the best cleaning effect.

[0083] For example, the timing mode is a pre-set automatic cleaning strategy, which sets different cleaning times according to different months and regions. According to the cycle, seasonality and differences between the north and south regions of poplar and willow catkins, the relationship between the flying of poplar and willow catkins and the weather is inseparable, mainly affected by temperature and humidity. The climate in the north is dry and there is less rain in the high-incidence season. The high-incidence season is divided into three stages: early April to mid-April, late April to early May and mid-May. The high-incidence period is from 10 to 16 o'clock every day. According to this feature, each stage is cleaned 3-5 times, and the single operation time is 5-10 minutes; the humidity in the south is generally high, and poplar and willow catkins are not as easy to blow everywhere as in the north, but there are still a small amount of poplar and willow catkins that need to be cleaned. The high-incidence season is at the end of April and May and mid-June every year. There is no high-incidence period. According to this feature, each stage is cleaned 1-2 times, and the single operation time is 5-10 minutes.

[0084] The parameter configuration unit is used to set the control parameters of the motor module. Under different working modes and cleaning strategies, different specifications of control parameters are adopted to perform cleaning operations of different intensities. The control parameters include cleaning frequency, push-pull speed and cleaning time.

[0085] Specifically, the cleaning frequency is to set the starting frequency of the motor module, which affects the interval time of the cleaning operation. The user can enter the frequency parameter in different working modes, and the system will control the start and stop frequency of the motor according to this parameter. The push-pull speed is to set the moving speed of the push-pull device of the motor module, which directly affects the cleaning effect and efficiency. It can preset different speed parameters according to different cleaning strategies (high, medium, and low intensity), and automatically adjust after the user selects the mode. The cleaning duration is to set the duration of each cleaning operation to ensure that the catkins are thoroughly cleaned. After the mode is selected, the user can enter the specific duration according to the cleaning strategy, and the system controls the operation of the motor according to this duration.

[0086] The control interface unit is used to provide a user interaction interface, access the manual mode operation display interface, and display real-time data, equipment status and current mode information.

[0087] The local monitoring unit is used to monitor the operating status of the motor module and the cleaning effect of the cleaning operation in real time, and transmit the monitoring data to the feedback adjustment module and the remote control module.

[0088] The data collection and monitoring module is used to configure the distributed monitoring stations in the region and obtain the catkin concentration and meteorological data collected in real time by each monitoring station. The meteorological data includes temperature, humidity, wind direction, wind speed and precipitation.

[0089] The remote control module is used to predict the catkin concentration and duration in the current area at a future time based on the real-time collected catkin concentration and meteorological data through data analysis and processing of the catkin prediction model, optimize the catkin cleaning strategy, and issue remote control instructions to the feedback regulation module. The catkin prediction model includes a concentration prediction model and a time prediction model.

[0090] In the description of the present invention, the remote control module includes: a data transmission unit, a concentration prediction unit, a time prediction unit, a strategy optimization unit and a remote control unit.

[0091] Among them, the data transmission unit is used to obtain the catkin concentration and meteorological data collected by the monitoring site, and record and manage them according to the time information and site location information.

[0092] The concentration prediction unit is used to analyze the correlation of willow catkin concentrations at different monitoring sites, and to build a concentration prediction model based on the time series delayed correlation algorithm in combination with the long short-term memory network to predict the willow catkin concentration at the location of each low-voltage integrated distribution box in the intelligent mode.

[0093] Specifically, the Long Short-Term Memory (LSTM) network is a special recurrent neural network (RNN) that can learn and remember dependencies in long time series. LSTM effectively solves the gradient vanishing or exploding problem faced by traditional RNN in long time series learning by introducing the forget gate, input gate, and output gate mechanism. This enables LSTM to retain key information and ignore irrelevant information when processing time series data.

[0094] LSTM is particularly suitable for processing time-dependent sequence data, because the catkin concentration data itself changes over time, and LSTM can capture these time dependencies. At the same time, considering the catkin concentrations at different monitoring sites and their correlations, LSTM can process multivariate time series data and mine potential interactive relationships.

[0095] In the description of the present invention, the correlation of catkin concentrations at different monitoring sites is analyzed, and a concentration prediction model based on a time series delay correlation algorithm is constructed in combination with a long short-term memory network. The catkin concentrations at the locations of various low-voltage integrated distribution boxes in the intelligent mode are predicted, including:

[0096] Step S11, obtaining the catkin concentration collected at each monitoring site, normalizing the catkin concentration by the maximum-minimum normalization method, and converting the catkin concentration into a time series according to the order of monitoring time, with each monitoring site corresponding to a time series.

[0097] Specifically, the concentration of catkins can be monitored in real time through a particle concentration sensor (such as a PM2.5 / PM10 sensor), and the sensor data can be filtered and averaged to remove abnormal values ​​to obtain a more accurate concentration value. Then set the standard for catkin concentration, using the mass concentration of catkin particles in the air (such as micrograms / cubic meter) as the measurement standard. Define the concentration threshold according to the actual situation, such as: low concentration: 0-30μg / m 3 ; Medium concentration: 30-100 μg / m 3 ; High concentration: >100μg / m 3 Step S12: Set the catkin concentration to be predicted in the low-voltage integrated distribution box and the catkin concentration of the monitoring stations around the low-voltage integrated distribution box at the same time, and use the time series delay correlation algorithm to obtain an input sample data matrix through calculation and screening.

[0098] In the description of the present invention, the catkin concentration to be predicted of the low-voltage integrated distribution box and the catkin concentration of the monitoring stations around the low-voltage integrated distribution box at the same time are set, and the time series delay correlation algorithm is used. After calculation and screening, the input sample data matrix is ​​obtained, including:

[0099] Step S121: Set the catkin density at the location of the low-voltage integrated power distribution station at time t to Ct , and the catkin concentration C collected by the surrounding i-th monitoring station at time t it .

[0100] Step S122, set the delay time window, take the catkin concentration at the current time and the past N times as input samples, and combine the catkin concentration of the low-voltage integrated distribution box and each monitoring station to form a sample data matrix. Thus, a time series sample is created, so that the model can learn the dynamic changes and trends in time. The input samples will be used to train the prediction model to help capture the change pattern of catkin concentration over time.

[0101] Step S123: for each monitoring site, the delayed correlation between different monitoring sites is calculated using the cross-correlation function. By calculating the cross-correlation function between different monitoring sites, the monitoring site most relevant to the catkin concentration to be predicted is selected, wherein the expression of the cross-correlation function is:

[0102]

[0103] Where CCF i,j τ represents the correlation coefficient between the ith and jth monitoring stations; μ i and μ j Respectively represent the mean of the i-th and j-th monitoring stations; σ i and σ j They represent the standard deviations of the ith and jth monitoring stations respectively; E represents the expected value; τ represents the time delay.

[0104] Specifically, the time series delay correlation algorithm is used to capture and analyze the correlation between different monitoring sites in time series data, especially when there is a time delay. The algorithm can identify potential influencing factors, thereby providing more accurate input features for subsequent prediction models and ensuring that the model can fully utilize the time series information of historical data.

[0105] The Cross-Correlation Function (CCF) is a statistical tool used to analyze the relationship between two time series. It can quantify the degree of influence of one time series on another at different time delays. The cross-correlation function helps identify the similarity or dependency between two signals by calculating the correlation between them at different time points. The cross-correlation function can identify the strength of the correlation between two time series at different time delays. By using different delay values, it is possible to analyze whether the impact of one series on another is immediate or delayed. In machine learning and predictive modeling, the cross-correlation function helps identify and select the most influential features to improve the performance of the model.

[0106] In the present invention, for the prediction of catkin concentration, there may be correlations between catkin concentrations at different monitoring sites. Through the cross-correlation function, these correlations can be quantified to help identify which monitoring sites have a significant impact on the catkin concentration at the predicted location. The cross-correlation function can process time series data and is suitable for analyzing the dynamic relationship of catkin concentration over time. It can capture the influence relationship of different monitoring points in time and provide the time series characteristics required by the model.

[0107] In addition, the mutual interference function can ensure the quality of the input sample data matrix and improve the accuracy and robustness of the prediction model by screening out the monitoring sites that are most relevant to the catkin concentration to be predicted, so as to ensure that the input sample data matrix contains the most influential data, improve the prediction accuracy of the model, and provide a basis for selecting relevant features.

[0108] Step S124: forming an input sample data matrix according to the catkin concentrations of the screened monitoring sites.

[0109] Specifically, the input sample data matrix obtained through calculation and screening includes the catkin concentration of the low-voltage integrated distribution box to be predicted and the catkin concentration data of each monitoring station related to it in the same time window, providing structured input data for the concentration prediction model and helping the model learn the dynamic relationship and correlation in the time series. It helps to enhance the generalization ability of the model, improve the prediction accuracy, ensure that the changing law of catkin concentration can be effectively reflected in practical applications, and then support the formulation of intelligent cleaning strategies.

[0110] Step S13: divide the historical catkin concentration data into a training set and a test set, use the training set to establish a concentration prediction model based on the long short-term memory network and train it, then input the input sample data matrix into the concentration prediction model, and output the catkin concentration corresponding to the location of the low-voltage integrated distribution box.

[0111] It can be seen from the above steps S11 to S13 that when predicting the catkin concentration at the location of the low-voltage integrated distribution box, the concentration prediction model combining the long short-term memory network (LSTM) and the time series delay correlation algorithm can effectively capture the time dependence of the catkin concentration and the correlation between the monitoring sites. LSTM is a special recurrent neural network with the ability of long-term memory and is suitable for processing time series data, so that it can accurately model the temporal trend of the catkin concentration.

[0112] Through the time series delayed correlation algorithm, the willow catkin concentration data of different monitoring stations are first normalized and converted into time series. Then, the delayed correlation between the monitoring stations is analyzed, and the stations most relevant to the predicted concentration are screened out, ensuring that the construction of the input sample data matrix contains sufficient information, thereby improving the prediction accuracy of the model.

[0113] During the training process, historical concentration data is divided into training sets and test sets to ensure that the model learns potential patterns and verifies its effect on the test set. Ultimately, the model can output the catkin concentration prediction results for a specific low-voltage integrated distribution box, providing a scientific basis for the formulation of cleaning strategies, thereby improving cleaning efficiency and intelligence.

[0114] Overall, the application of LSTM combined with the time series delayed correlation algorithm can fully consider the temporal dependence and spatial correlation, and is an ideal choice for the prediction of catkin concentration.

[0115] The time prediction unit is used to extract feature sets based on real-time collected meteorological data, establish a time prediction model with the duration of willow catkins flying as a label, and use the particle swarm optimization algorithm to optimize the hyperparameters of the time prediction model to output the duration of willow catkins flying.

[0116] In the description of the present invention, based on the meteorological data collected in real time, the feature set is extracted, the duration of willow catkins flying is used as a label, a time prediction model is established, and the hyperparameters of the time prediction model are optimized using a particle swarm optimization algorithm. The output duration of willow catkins flying includes:

[0117] Step S21, obtain historical meteorological data for normalization processing, scale the original data to a specific space, extract effective features of the meteorological data, and divide the meteorological training set and the meteorological verification set.

[0118] Step S22, setting the data window size, discard ratio, number of hidden layer neurons and learning rate of the gated recurrent network as the hyperparameter optimization target, substituting the predicted duration of catkin flying into the target fitness function, and using the particle swarm optimization algorithm to find the optimal hyperparameters of the gated recurrent network.

[0119] Specifically, the hyperparameters of the gated recurrent network (GRU / LSTM) include:

[0120] 1. Data Window Size: Set the input time step, such as 5, 10, etc., to determine the short-term memory capacity of the gated recurrent network model.

[0121] 2. Dropout Rate: Prevent overfitting, usually set to 0.2 to 0.5.

[0122] 3. Number of hidden layer neurons (Hidden Units): affects the complexity and learning ability of the gated recurrent network model.

[0123] 4. Learning Rate: Controls the speed at which the parameters of the gated recurrent network model are updated, usually set between 0.001 and 0.01.

[0124] In addition, the particle swarm optimization algorithm is an optimization algorithm based on swarm intelligence, which searches by simulating the foraging behavior of bird flocks. Each "particle" represents a potential solution, and over time, the particle continuously updates its position in the search space to find the best solution. When optimizing the hyperparameters of the GRU model, PSO can effectively search the hyperparameter space and find the hyperparameter combination that optimizes the prediction performance, thereby improving the prediction ability and accuracy of the model.

[0125] In the description of the present invention, the data window size, discard ratio, number of hidden layer neurons and learning rate of the gated recurrent network are set as the hyperparameter optimization targets, and the predicted duration of catkins flying is substituted into the target fitness function. The optimal hyperparameters of the gated recurrent network are found by using the particle swarm optimization algorithm, including:

[0126] Step S221, setting the initial parameters of the particle swarm optimization algorithm, including the number of particles, the number of iterations, the initial inertia weight and the learning factor, and setting the optimization range of the hyperparameters of the gated recurrent network and the particle velocity threshold, as well as the target fitness function.

[0127] Step S222, randomly initialize the population, determine the initial population and individual optimal values, use the particle swarm individual position and velocity update formula to obtain the latest position and velocity of the particle, and build a time prediction model based on the gated recurrent network for training according to the hyperparameters corresponding to the latest position of each particle, then return the model prediction value and substitute it into the target fitness function for verification, and generate the population and individual optimal values ​​of this iteration according to the fitness value of the particle.

[0128] Step S223, determine whether the iteration termination condition is met, if not, continue to iterate to find the optimal solution, if yes, terminate the iteration and output the optimal hyperparameters of the time prediction model.

[0129] Step S23, by configuring the optimized hyperparameters, a time prediction model based on a gated recurrent network is constructed, the processed meteorological data and concentration data are used as model inputs, and after processing by a fully connected layer, the duration of catkin flying at the corresponding catkin concentration is output.

[0130] Specifically, the Gated Recurrent Unit (GRU) is an improved recurrent neural network that aims to solve the gradient vanishing and exploding problems of traditional RNN in long sequence learning. GRU can control the flow of information in the network by introducing the update gate and reset gate mechanism. The update gate determines how much information from the previous moment is retained, while the reset gate controls the impact of the current input on the previous information. This flexible mechanism enables GRU to effectively capture long-term and short-term dependencies, thereby performing well in sequence data modeling.

[0131] The principle of building a GRU-based time prediction model is to use historical time series data to predict future outputs. Through its gating mechanism, GRU is able to keep a memory of important information during the learning process and forget irrelevant information when necessary. This mechanism allows the model to capture the potential time dependencies in the input data, so that it can make more accurate predictions based on previous historical information when faced with new inputs. In the prediction of the duration of willow catkins flying, GRU can comprehensively consider the complex relationship between meteorological conditions and historical concentration data to generate reasonable duration prediction results.

[0132] The strategy optimization unit is used to optimize the catkin cleaning strategy of the drive motor in the low-voltage integrated distribution box according to the catkin concentration prediction results and the duration prediction results, adjust the intensity of the cleaning operation, and adjust the corresponding control parameters according to the changes in the operation intensity.

[0133] Specifically, according to the real-time prediction results of catkin concentration and duration, the cleaning strategy of the drive motor in the low-voltage integrated distribution box is intelligently adjusted to ensure the efficiency and effectiveness of the cleaning operation. According to the prediction results, the intensity of the cleaning operation can be automatically adjusted to decide whether to use high, medium or low intensity cleaning operations. Then, according to the selected operation intensity, the relevant control parameters (such as cleaning frequency, push and pull speed and cleaning duration) are dynamically adjusted to optimize the cleaning effect and energy consumption.

[0134] For example, cleanup strategy decisions can be divided into:

[0135] 1. High concentration and long duration: Select a high-intensity cleaning strategy and set corresponding control parameters (such as high frequency and fast speed).

[0136] 2. Medium concentration, medium duration: Select a medium-intensity cleaning strategy and configure moderate control parameters.

[0137] 3. Low concentration and short duration: Select a low-intensity cleaning strategy and set lower control parameters.

[0138] The remote control unit is used to generate remote control instructions according to the strategy optimization results and send them to the feedback adjustment module through the wireless communication network.

[0139] The feedback adjustment module is used to drive and adjust the cleaning mode and control parameters set by the local control module according to the received remote control instructions, and to monitor the equipment operation status and catkin cleaning effect in real time, and further optimize the control parameters according to the monitoring feedback information.

[0140] Specifically, the feedback adjustment module adjusts the cleaning mode and control parameters set by the local control module according to the received remote control instructions to ensure that the equipment performs the cleaning operation according to the optimization strategy. The operating status of the equipment is continuously monitored, including key indicators such as the motor's operating speed, frequency, and cleaning effect, to ensure the effectiveness of the cleaning operation. Finally, based on real-time monitoring data, the operating status of the equipment and the cleaning effect of catkins are analyzed, and the control parameters are adjusted and optimized in time to cope with different environmental changes and cleaning needs.

[0141] The wireless communication module is used to realize wireless transmission between modules through wireless communication protocols.

[0142] In summary, with the help of the above technical scheme of the present invention, by adopting a low-voltage integrated distribution box with a push-pull catkin cleaning device, on the basis of ensuring the function of the existing low-voltage integrated distribution box, as long as this cleaning device is added, while ensuring that the protection level remains unchanged, the catkins at the ventilation holes can be effectively cleaned, so as to reduce the temperature rise of the components inside the box or the inside of the box, improve the service life of the product, and ensure the safe operation of the power grid equipment. By integrating multi-module collaboration, efficient and intelligent catkin cleaning management is realized, and the cleaning time and mode are flexibly set according to the seasonal changes of catkins, and the dynamic adjustment based on meteorological and concentration prediction is improved to improve the cleaning efficiency and energy saving effect; remote control and wireless communication technology are adopted to optimize the user control experience, and the accuracy and adaptability of the prediction are improved by using machine learning models, so that the cleaning strategy is more flexible and intelligent, and the overall control system has stronger intelligent decision-making ability, adaptability and real-time response ability, thereby effectively reducing energy consumption and improving the cleaning effect, greatly enhancing the scientificity and efficiency of catkin cleaning work.

[0143] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

Claims

1. A push-pull catkin cleaning device for a low-voltage integrated distribution box housing, comprising a distribution box housing, a fixed bottom plate being arranged at the bottom of the distribution box housing, characterized in that: Fixed guide rails are arranged on both sides of the bottom end of the fixed bottom plate, a motor module is arranged at one end of the fixed guide rail, a brush fixing bracket is arranged between the two fixed guide rails, and a bottom plate push-pull brush is arranged at the top of the brush fixing bracket; The fixed bottom plate is fixedly connected to the fixed guide rail via a guide rail base, and a plurality of bottom plate ventilation holes are provided at the bottom of the fixed bottom plate.

2. The push-pull catkin cleaning device for the low-voltage integrated distribution box housing according to claim 1 is characterized in that: The motor module includes a steering gear set arranged at one end of the fixed guide rail, a driving shaft is arranged between the two steering gear sets, a stepping motor is arranged on one side of one of the steering gear sets, and the output end of the stepping motor is connected to the driving shaft; A driving screw matched with the steering gear set is arranged inside the fixed guide rail, and the bottom of the fixed guide rail is connected to both sides of the brush fixing bracket by means of crawlers.

3. The push-pull catkin cleaning device for the low-voltage integrated distribution box housing according to claim 1 is characterized in that: The surface of the bottom plate push-pull brush is made of polyester woven material, and the surface of the polyester woven material is smooth and has polarity, which is used to generate electrostatic adsorption.

4. A control system for a push-pull willow catkin cleaning device for a low-voltage integrated distribution box housing, used to realize the control of the push-pull willow catkin cleaning device for a low-voltage integrated distribution box housing according to any one of claims 1 to 3, characterized in that: The control system includes: a power supply module, a local control module, a collection and monitoring module, a remote control module, a feedback adjustment module and a wireless communication module; The power supply module is used to connect the power supply line inside the low-voltage integrated distribution box, support multiple power supply modes, provide power supply for the motor module, and monitor the power supply status; The local control module is used to set seasonal initial working time, multiple cleaning modes and corresponding control parameters according to the region where the low-voltage integrated distribution box is located and the catkin production cycle, and configure a user control interface to provide manual adjustment of the cleaning mode and control parameters; The collection and monitoring module is used to configure the monitoring stations established in a distributed manner in the region, and obtain the catkin concentration and meteorological data collected in real time by each monitoring station, and the meteorological data includes temperature, humidity, wind direction, wind speed and precipitation; The remote control module is used to predict the concentration and duration of catkins in the current area at a future time based on the real-time collected catkins concentration and meteorological data, through data analysis and processing of the catkins prediction model, optimize the catkins cleaning strategy, and issue remote control instructions to the feedback regulation module, and the catkins prediction model includes a concentration prediction model and a time prediction model; The feedback adjustment module is used to drive and adjust the cleaning mode and control parameters set by the local control module according to the received remote control instructions, and monitor the equipment operation status and catkin cleaning effect in real time, and further optimize the control parameters according to the monitoring feedback information; The wireless communication module is used to realize wireless transmission between modules through a wireless communication protocol.

5. The control system of the push-pull type catkin cleaning device for the low-voltage integrated distribution box housing according to claim 4 is characterized in that: The local control module includes: a timing drive unit, a mode setting unit, a parameter configuration unit, a control interface unit and a local monitoring unit; The timing drive unit is used to activate the motor module to clean the catkins according to the preset initial working time during the process of catkins flying, and set the start time of the cleaning operation through the built-in timer to achieve regular opening and closing; The mode setting unit is used to set the working mode that the motor module needs to execute when running, and after the working mode is set, the catkin cleaning operation is performed according to the cleaning strategy pre-configured in the respective working modes, the working modes include manual mode, timing mode and intelligent mode, and the cleaning strategy includes high-intensity cleaning operation, medium-intensity cleaning operation and low-intensity cleaning operation; The parameter configuration unit is used to set the control parameters of the motor module. Under different working modes and cleaning strategies, different control parameters are used to perform cleaning operations of different intensities. The control parameters include cleaning frequency, push-pull speed and cleaning duration. The control interface unit is used to provide a user interaction interface, access the operation display interface of the manual mode, and display real-time data, device status and current mode information; The local monitoring unit is used to monitor the operating status of the motor module and the cleaning effect of the cleaning operation in real time, and transmit the monitoring data to the feedback adjustment module and the remote control module.

6. The control system of the push-pull type catkin cleaning device for the low-voltage integrated distribution box housing according to claim 4 is characterized in that: The remote control module includes: a data transmission unit, a concentration prediction unit, a time prediction unit, a strategy optimization unit and a remote control unit; The data transmission unit is used to obtain the catkin concentration and meteorological data collected by the monitoring site, and record and manage them according to the time information and site location information; The concentration prediction unit is used to analyze the correlation of catkin concentrations at different monitoring sites, and to construct a concentration prediction model based on a time series delay correlation algorithm in combination with a long short-term memory network to predict the catkin concentration at the location of each low-voltage integrated distribution box in the intelligent mode; The time prediction unit is used to extract a feature set based on the meteorological data collected in real time, use the duration of willow catkins flying as a label, establish a time prediction model, and use a particle swarm optimization algorithm to optimize the hyperparameters of the time prediction model to output the duration of willow catkins flying; The strategy optimization unit is used to optimize the catkin cleaning strategy of the drive motor in the low-voltage integrated distribution box according to the catkin concentration prediction result and the duration prediction result, adjust the intensity of the cleaning operation, and adjust the corresponding control parameters according to the change of the operation intensity; The remote control unit is used to generate a remote control instruction according to the strategy optimization result, and send it to the feedback adjustment module through the wireless communication network.

7. The control system of the push-pull type catkin cleaning device for the low-voltage integrated distribution box housing according to claim 6 is characterized in that: The analysis of the correlation of catkin concentrations at different monitoring sites and the construction of a concentration prediction model based on a time series delay correlation algorithm in combination with a long short-term memory network predict the catkin concentrations at the locations of various low-voltage integrated distribution boxes in the intelligent mode, including: Obtain the catkin concentration collected at each monitoring station, normalize it using the maximum-minimum normalization method, and convert it into a time series according to the monitoring time, with each monitoring station corresponding to a time series; The catkin concentration to be predicted in the low-voltage integrated distribution box and the catkin concentration at the monitoring stations around the low-voltage integrated distribution box at the same time are set, and the input sample data matrix is ​​obtained after calculation and screening using the time series delay correlation algorithm; The historical catkins concentration data were divided into training set and test set. The training set was used to establish a concentration prediction model based on long short-term memory network and train it. The input sample data matrix was then input into the concentration prediction model to output the catkins concentration at the location of the low-voltage integrated distribution box.

8. The control system of the push-pull type catkin cleaning device for the low-voltage integrated distribution box housing according to claim 7 is characterized in that: The catkin concentration to be predicted in the low-voltage integrated distribution box and the catkin concentration at the monitoring stations around the low-voltage integrated distribution box at the same time are set, and the input sample data matrix is ​​obtained after calculation and screening using the time series delay correlation algorithm, including: Assume that at time t, the catkin concentration at the location of the low-voltage integrated power distribution station is C t , and at time t, the catkin concentration C collected by the surrounding i-th monitoring station it ; The delay time window is set, and the catkin concentration at the current time and the past N times is used as the input sample. The catkin concentration at the low-voltage integrated distribution box and each monitoring station is combined to form a sample data matrix. For each monitoring site, the delayed correlation between different monitoring sites is calculated using the cross-correlation function. By calculating the cross-correlation function between different monitoring sites, the monitoring site most relevant to the catkin concentration to be predicted is screened out, wherein the expression of the cross-correlation function is: Where CCF i,j (τ) represents the mutual correlation coefficient between the ith and jth monitoring stations; μ i and μ j represent the mean values ​​of the ith and jth monitoring stations respectively; σ i and σ j represent the standard deviation of the ith and jth monitoring stations respectively; E represents expected value; τ represents the time delay; An input sample data matrix is ​​formed based on the catkin concentrations at the screened monitoring sites.

9. The control system of the push-pull type catkin cleaning device for the low-voltage integrated distribution box housing according to claim 8, characterized in that: According to the meteorological data collected in real time, the feature set is extracted, the duration of willow catkins flying is used as a label, a time prediction model is established, and the hyper parameters of the time prediction model are optimized by using a particle swarm optimization algorithm. The duration of willow catkins flying is output, including: Obtain historical meteorological data for normalization, scale the original data to a specific space, extract effective features of meteorological data, and divide them into meteorological training set and meteorological verification set; The data window size, dropout ratio, number of hidden layer neurons and learning rate of the gated recurrent network are set as the hyperparameter optimization targets, and the predicted duration of catkins flying is substituted into the target fitness function. The particle swarm optimization algorithm is used to find the optimal hyperparameters of the gated recurrent network. By configuring the optimized hyperparameters, a time prediction model based on a gated recurrent network is constructed. The processed meteorological data and concentration data are used as model inputs. After processing by the fully connected layer, the duration of willow catkin flying at the corresponding willow catkin concentration is output.

10. The control system of the push-pull type catkin cleaning device for the low-voltage integrated distribution box housing according to claim 9, characterized in that: The data window size, discard ratio, number of hidden layer neurons and learning rate of the gated recurrent network are set as the hyperparameter optimization target, and the predicted duration of catkins flying is substituted into the target fitness function. The optimal hyperparameters of the gated recurrent network are found by using the particle swarm optimization algorithm, including: Setting the initial parameters of the particle swarm optimization algorithm, including the number of particles, the number of iterations, the initial inertia weight and the learning factor, and setting the optimization range of the hyperparameters of the gated recurrent network and the particle velocity threshold, as well as the target fitness function; Initialize the population randomly, determine the initial population and individual optimal values, use the particle group individual position and velocity update formula to obtain the latest position and velocity of the particle, and build a time prediction model based on the gated recurrent network for training according to the hyperparameters corresponding to the latest position of each particle, then return the model prediction value and substitute it into the target fitness function for verification, and generate the optimal population and individual values ​​of this iteration according to the fitness value of the particle; Determine whether the iteration termination conditions are met. If not, continue to iterate to find the optimal solution. If yes, terminate the iteration and output the optimal hyperparameters of the time prediction model.

Citation Information

Patent Citations

  • Intelligent blast dust removal control system of distribution box

    CN117559244A

  • Low-voltage power distribution cabinet facilitating access of generator car

    CN217720410U

  • Self-cleaning type willow catkin preventing mechanism

    CN220005235U

  • Heat dissipation device of low-voltage power distribution cabinet

    CN221305352U

  • Anti-pollution distribution panel based on static electricity generation, and control method thereof

    KR101804091B1