Intelligent moisture-proof box humidity control method and system

Abstract

The present invention discloses a method and system for controlling humidity in an intelligent moisture-proof box, comprising the following steps: using a capacitive humidity sensor to monitor the humidity inside the moisture-proof box in real time based on optimizing LoRa, and dynamically adjusting the sampling frequency of the capacitive humidity sensor according to the historical humidity changes in the box to obtain a real-time humidity difference value; when the real-time humidity difference value is greater than a preset humidity difference value, using fast decoupling fuzzy control to quickly adjust the real-time humidity difference value to the preset humidity difference value; when the real-time humidity difference value is less than or equal to the preset humidity difference value, using a prediction controller based on a neural network model to gradually adjust the humidity until the preset humidity value is reached; when the number of times the monitored humidity value exceeds the target humidity range reaches a preset value, the humidification and dehumidification working cycles are adaptively optimized by a control algorithm according to environmental factors, so as to achieve efficient and accurate control of humidity. The present invention achieves more intelligent, accurate, fast and efficient humidity control.

Description

Technical Field

[0001] The present invention relates to the technical field of humidity control, and in particular to a humidity control method and system for an intelligent moisture-proof box. Background Art

[0002] As electronic products, optical instruments, precision instruments, cultural relics, medicines, etc. have higher and higher requirements for storage environment humidity, humidity control has become an important part of protecting these items. Excessive humidity can easily lead to oxidation and short circuit of electronic components, moldy optical instrument lenses, deterioration of medicines, and even corrosion of cultural relics; while low humidity can also cause aging and cracking of certain materials. Therefore, in order to ensure the safe storage of these items in various environments, moisture-proof boxes, as a professional storage equipment, have been widely used.

[0003] Traditional moisture-proof boxes usually use physical moisture-absorbing materials or electronic drying devices to control the humidity inside the box. However, moisture-proof boxes in the prior art have the following problems in terms of humidity control:

[0004] (1) Low humidity control accuracy: Many humidity control systems of moisture-proof boxes can only be adjusted within a large humidity range, which makes it difficult to meet the strict humidity requirements for precision items, especially for some humidity-sensitive items such as optical instruments and semiconductor components. Even slight fluctuations in humidity may cause irreversible damage.

[0005] (2) Slow humidity response: Existing moisture-proof boxes often cannot adjust the humidity in the box in time when the ambient humidity changes, which may cause the stored items to be exposed to an unsuitable humidity environment in a short period of time, increasing the risk of damage;

[0006] (3) High energy consumption: Traditional electronic drying devices usually operate at constant power and cycle when regulating humidity, resulting in energy waste and increasing the operating cost of the equipment;

[0007] (4) Humidity monitoring is not intelligent enough: Most existing moisture-proof boxes rely on simple humidity sensors for monitoring. The control system lacks an intelligent adjustment mechanism and is unable to perform adaptive humidity control based on environmental changes and other factors. It is not convenient to use and lacks flexibility.

[0008] In order to solve the above problems, the prior art urgently needs a more intelligent, accurate, fast and efficient humidity control method and system for use in moisture-proof boxes to better protect the safety and quality of stored items. Summary of the invention

[0009] In order to overcome the shortcomings of the prior art, the present invention provides a method and system for humidity control in an intelligent moisture-proof box, which are used to solve the technical problems of low control accuracy, slow response speed, high energy consumption and lack of intelligence in the existing moisture-proof box in terms of humidity control, and to achieve more intelligent, precise, fast and efficient humidity control.

[0010] To solve the above problems, the technical solution adopted by the present invention is as follows:

[0011] A method for controlling humidity in an intelligent moisture-proof box comprises the following steps:

[0012] The capacitive humidity sensor is used to monitor the humidity inside the moisture-proof box in real time based on the optimized LoRa, and the sampling frequency of the capacitive humidity sensor is dynamically adjusted according to the historical changes in the humidity inside the box;

[0013] Compare the monitored humidity value with the target humidity range to obtain the real-time humidity difference value;

[0014] When the real-time humidity difference value is greater than the preset humidity difference value, the fast decoupling fuzzy control is used to quickly adjust the real-time humidity difference value to the preset humidity difference value;

[0015] When the real-time humidity difference value is less than or equal to the preset humidity difference value, a prediction controller based on a neural network model is used to gradually adjust the humidity until a preset humidity value within the target humidity range is reached;

[0016] When the monitored humidity value exceeds the target humidity range for a number of times that reaches a preset value, the humidification and dehumidification working cycles are adaptively optimized through a control algorithm based on environmental factors to achieve efficient and precise control of humidity.

[0017] As a preferred embodiment of the present invention, when real-time monitoring of the humidity inside the moisture-proof box is performed, it includes:

[0018] Introduce the LoRa radio frequency module and connect it with the transmission controller to build a radio frequency layer, and define the communication protocol working frequency band in the radio frequency layer;

[0019] In the physical layer, the continuous time of each frequency modulation signal is determined by the spreading factor and bandwidth, and linear spread spectrum frequency modulation is performed;

[0020] The transmission rate of LoRa is determined by the spreading factor, coding rate and bandwidth at the medium access control layer;

[0021] In the dormant state, communication is performed through the SPI or UART interface. After initialization, the wireless communication module enters the standby and transceiver states, and the carrier frequency is configured for the wireless communication module;

[0022] Initialize data transmission, generate control instructions according to environmental parameters, and send real-time humidity monitoring data of the moisture-proof box.

[0023] As a preferred embodiment of the present invention, when sending the real-time humidity monitoring data of the moisture-proof box, it includes:

[0024] Determine the first and second random number seeds according to the moisture-proof box to be monitored, and generate a third random number seed;

[0025] The first, second and third random number seeds form a random number sequence and add it to the timer for encryption of real-time humidity data;

[0026] When sending next time, the first and second random number seeds are updated, and a new third random number seed and a new random number sequence are generated, and the new random number sequence is used to encrypt the real-time humidity data.

[0027] As a preferred embodiment of the present invention, when generating a control instruction according to environmental parameters, it includes:

[0028] Generate different humidity monitoring instructions based on historical changes in humidity in the box and dynamically adjust the sampling frequency of the capacitive humidity sensor;

[0029] The dynamic adjustment process includes:

[0030] Construct a BP neural network model for dynamically adjusting the sampling frequency, and select sample data for training;

[0031] Sampling data at an initial sampling frequency, and preprocessing the humidity sampling value before inputting it into the BP neural network model for humidity prediction to obtain a humidity prediction value;

[0032] The humidity prediction value is compared with a humidity threshold model interval generated based on historical humidity data, and the sampling frequency is dynamically adjusted.

[0033] As a preferred embodiment of the present invention, when constructing a BP neural network model, it includes:

[0034] Set at the input layer Nodes , used to receive the preprocessed humidity sampling values;

[0035] Set in hidden layer Nodes ,choose The function is used as the activation function, and the connection weights between the input layer and the hidden layer are set to ;

[0036] Set at the output layer Nodes ,choose The function is used as the activation function, and the connection weights between the hidden layer and the output layer are set to ;

[0037] When preprocessing the humidity sampling value, it includes:

[0038] The humidity sampling value is subjected to power frequency interference elimination and wavelet threshold denoising to obtain a denoised humidity sampling value;

[0039] The denoised humidity sampling values ​​are normalized based on the mean and standard deviation of historical humidity data.

[0040] As a preferred embodiment of the present invention, when fast decoupling fuzzy control is used for fast adjustment, it includes:

[0041] According to the real-time humidity difference value and the target humidity range, the humidity deviation change rate is obtained;

[0042] Will Parameter gain , , The basic domain and fuzzy domain are set as the target value, and the real-time humidity difference value and humidity deviation change rate are combined with the quantization factor and the target humidity range as input;

[0043] Based on the control rule table and fuzzy reasoning, the output Parameter gain , , , and after combining with the scale factor, defuzzification is performed to obtain the final Parameter gain , , ;

[0044] According to the final Parameter gain , , Adjustment In the controller , , parameter.

[0045] As a preferred embodiment of the present invention, when the humidity is gradually adjusted by using a predictive controller based on a neural network model, it includes:

[0046] Construct the future input sequence of the predictive controller in the control horizon;

[0047] Input the future input sequence into the predictive controller to obtain the predicted time domain The control output under

[0048] The target control sequence is obtained by minimizing the target function, and a smooth control signal is outputted using the target control sequence to gradually adjust the humidity until a preset humidity value within the target humidity range is reached.

[0049] As a preferred embodiment of the present invention, when minimizing the objective function to obtain the target control sequence, it includes:

[0050] Assigning a certain power constraint to the objective function;

[0051] The objective function is minimized by a sequential quadratic programming method to obtain a control sequence within the power constraint, as shown in Formula 16:

[0052] (16);

[0053] In the formula, To control the cycle The power increase of the internal humidification module, To control the cycle The power increment of the internal dehumidification module, , , and , , The control cycles are based on the humidification module and the dehumidification module respectively. , , The difference between the internal humidity and the preset humidity value, , , and , , Humidifier and dehumidifier Parameters of the controller.

[0054] As a preferred embodiment of the present invention, when adaptively optimizing the working cycle of humidification and dehumidification, it includes:

[0055] The outside temperature, outside humidity and inside temperature of the box are taken as inputs and mapped using a three-input single-output super closed ball CMAC network to obtain a model for estimating the humidity inside the box.

[0056] Among them, input data is collected through a sliding data window to obtain an input data set;

[0057] Based on the temperature inside and outside the box, the humidity inside and outside the box, and the air flow rate between the box and the outside, establish Control functions;

[0058] The estimated value of the humidity in the box output by the humidity estimation model is input into In the control function, we get Control optimization value;

[0059] Based on the Control optimization value optimizes the duty cycle of humidification and dehumidification.

[0060] An intelligent moisture-proof box humidity control system, comprising:

[0061] Sampling frequency dynamic adjustment unit: used to dynamically adjust the sampling frequency of the capacitive humidity sensor according to the historical changes in humidity in the box;

[0062] The first humidity adjustment unit is used to compare the monitored humidity value with the target humidity range to obtain a real-time humidity difference value; when the real-time humidity difference value is greater than the preset humidity difference value, the real-time humidity difference value is quickly adjusted to the preset humidity difference value by using fast decoupling fuzzy control;

[0063] A second humidity adjustment unit: when the real-time humidity difference value is less than or equal to the preset humidity difference value, a prediction controller based on a neural network model is used to gradually adjust the humidity until a preset humidity value within the target humidity range is reached;

[0064] Humidity control optimization unit: When the monitored humidity value exceeds the target humidity range for a preset number of times, the humidification and dehumidification working cycles are adaptively optimized through the control algorithm according to environmental factors to achieve efficient and accurate humidity control;

[0065] Among them, the capacitive humidity sensor is used to monitor the humidity inside the moisture-proof box in real time based on the optimized LoRa.

[0066] Compared with the prior art, the present invention has the following beneficial effects:

[0067] (1) The present invention introduces a LoRa radio frequency module to construct a radio frequency layer, and combines the radio frequency layer with the physical layer and the media access control layer to achieve real-time monitoring and transmission of humidity data in the moisture-proof box, thereby accurately regulating the humidity in the moisture-proof box based on accurate real-time humidity data;

[0068] (2) The present invention constructs a BP neural network model for dynamically adjusting the sampling frequency, and generates a humidity threshold model interval based on historical humidity data. The humidity prediction is output through the BP neural network model, and the humidity prediction value is compared with the humidity threshold model interval. The sampling frequency is dynamically adjusted according to the comparison result, thereby effectively reducing the sampling error generated in the sampling process and obtaining a more accurate humidity sampling value.

[0069] (3) The present invention eliminates power frequency interference and wavelet threshold denoising on the humidity sampling value based on the reference capacitor, effectively eliminates the fluctuation of the sensor capacitance signal, and better retains the useful signal, thereby obtaining more effective normalized data for humidity prediction;

[0070] (4) When the monitored humidity value differs greatly from the target humidity range, the present invention uses fast decoupling fuzzy control to quickly adjust the real-time humidity difference value to the preset humidity difference value (first-level humidity adjustment), and then cooperates with the prediction controller based on the neural network model to accurately adjust the humidity (second-level humidity adjustment), so that the humidity in the box gradually approaches the target value and accurately reaches the target value;

[0071] (5) The present invention constructs a future input sequence for the predictive controller, inputs it into the predictive controller to obtain the control output in the prediction time domain, and obtains the target control sequence by minimizing the objective function, thereby using the target control sequence to output a smooth control signal, so that the humidity in the box gradually and smoothly approaches the target value and finally reaches the target value, thereby realizing precise control of the humidity in the box;

[0072] (6) The present invention estimates the humidity inside the box by constructing a humidity estimation model inside the box, and combines Control function output Control optimization value to optimize the humidification and dehumidification working cycle, achieve efficient and accurate control of long-term humidity, and optimize energy consumption.

[0073] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 It is a step diagram of the humidity control method of the intelligent moisture-proof box provided by the present invention;

[0075] Figure 2 This is a hardware framework diagram based on optimizing LoRa real-time monitoring provided by the present invention;

[0076] Figure 3 It is a structural diagram of the BP neural network model for dynamically adjusting the sampling frequency provided by the present invention.

[0077] Explanation of the figure numbers: 1. Moisture-proof box; 2. Capacitive humidity sensor; 3. Transmission controller; 4. LoRa radio frequency module; 5. Wireless communication module; 6. Input layer; 7. Hidden layer; 8. Output layer. DETAILED DESCRIPTION

[0078] The intelligent moisture-proof box humidity control method provided by the present invention is as follows: Figure 1 As shown, the following steps are included:

[0079] Step S1: monitor the humidity inside the moisture-proof box 1 in real time through the capacitive humidity sensor 2 based on the optimized LoRa, and dynamically adjust the sampling frequency of the capacitive humidity sensor 2 according to the historical changes in the humidity inside the box;

[0080] Step S2: Compare the monitored humidity value with the target humidity range to obtain a real-time humidity difference value;

[0081] Step S3: When the real-time humidity difference value is greater than the preset humidity difference value, the real-time humidity difference value is quickly adjusted to the preset humidity difference value by using fast decoupling fuzzy control;

[0082] Step S4: when the real-time humidity difference value is less than or equal to the preset humidity difference value, the humidity is gradually adjusted by a prediction controller based on a neural network model until a preset humidity value within a target humidity range is reached;

[0083] Step S5: When the number of times the monitored humidity value exceeds the target humidity range reaches a preset value, the humidification and dehumidification working cycles are adaptively optimized through the control algorithm according to environmental factors to achieve efficient and accurate control of humidity.

[0084] In the above step S1, if Figure 2 As shown, when the internal humidity of the moisture-proof box 1 is monitored in real time based on the optimization of LoRa, it includes:

[0085] Introduce LoRa radio frequency module 4, connect LoRa radio frequency module 4 with transmission controller 3, build radio frequency layer, and define communication protocol working frequency band in radio frequency layer;

[0086] In the physical layer, the continuous time of each frequency modulation signal is determined by the spreading factor and bandwidth, and linear spread spectrum frequency modulation is performed;

[0087] At the medium access control layer, the transmission rate of LoRa is determined by the spreading factor, coding rate and bandwidth, as shown in Formula 1:

[0088] (1);

[0089] In the formula, is the transmission rate of LoRa, is the spreading factor, is the bandwidth, is the coding rate;

[0090] In the dormant state, communication is performed through the SPI or UART interface. After initialization, the wireless communication module 5 enters the standby and transceiver states, and the carrier frequency is configured for the wireless communication module 5, as shown in Formula 2:

[0091] (2);

[0092] In the formula, is the carrier frequency, is the crystal frequency, is the RF frequency, is the preset index value;

[0093] Initialize data transmission, generate control instructions according to environmental parameters, and send real-time humidity monitoring data of moisture-proof box 1;

[0094] Among them, the control instructions include: humidity monitoring instructions, status register read and write instructions.

[0095] Specifically, the preset index value The value of is 17, 18, 19, preferably, the value is 18.

[0096] Furthermore, when sending the real-time humidity monitoring data of the moisture-proof box 1, it includes:

[0097] Determine the first and second random number seeds according to the moisture-proof box 1 to be monitored;

[0098] According to the first and second random number seeds, a third random number seed is generated, as shown in Formula 3:

[0099] (3);

[0100] In the formula, is the first random number seed, is the second random number seed, is the third random number seed, is a preset constant value, is a 32-bit hexadecimal constant. is a bitwise AND operation;

[0101] The first, second and third random number seeds are used to form a random number sequence, and the random number sequence is added to a timer for encryption of real-time humidity data;

[0102] Among them, when sending next time, the first and second random number seeds are updated, as shown in formula 4:

[0103] (4);

[0104] (5);

[0105] In the formula, is the first random number seed after update, is the updated second random number seed, is the modulus, To add value, is the modulo operation;

[0106] A new third random number seed and a new random number sequence are obtained according to the new first and second random number seeds, and the new random number sequence is used for encrypting the real-time humidity data.

[0107] Specifically, the preset constant value The value of is between 65500 and 65550, preferably, the value is 65530.

[0108] Specifically, the present invention constructs a radio frequency layer by introducing a LoRa radio frequency module 4, and combines the radio frequency layer with a physical layer and a media access control layer, thereby realizing real-time monitoring and transmission of humidity data in the moisture-proof box 1, so that the humidity in the moisture-proof box 1 can be accurately regulated based on accurate real-time humidity data. In addition, the present invention also equips different moisture-proof boxes 1 with different first and second random number seeds according to the moisture-proof boxes 1 to be monitored, thereby generating different third random number seeds and random number sequences, thereby improving the security of data transmission, and updating the random number seeds and random number sequences for the same moisture-proof box 1 at each transmission, thereby further improving the security of data transmission.

[0109] Furthermore, when generating control instructions according to environmental parameters, it includes:

[0110] Generate different humidity monitoring instructions based on historical changes in humidity in the box and dynamically adjust the sampling frequency of the capacitive humidity sensor;

[0111] The dynamic adjustment process includes:

[0112] Construct a BP neural network model for dynamically adjusting the sampling frequency, and select sample data to train the BP neural network model;

[0113] Generate humidity threshold model interval based on historical humidity data, as shown in Formula 6:

[0114] (6);

[0115] In the formula, The maximum value of humidity data within a certain period of time The average value of The minimum value of humidity data within a certain period of time The average value of is the first floating value that follows a normal distribution, is the second floating value that follows a normal distribution;

[0116] The data is sampled at the initial sampling frequency, and after preprocessing the humidity sampling value, it is input into the BP neural network model for humidity prediction to obtain the humidity prediction value;

[0117] Compare the humidity predictions to the humidity threshold model intervals;

[0118] If it is less than the humidity threshold model interval, the sampling frequency is increased; if it is greater than the humidity threshold model interval, the sampling frequency is reduced, as shown in Formula 7 and Formula 8:

[0119] (7);

[0120] (8);

[0121] In the formula, is the adjusted sampling frequency, is the initial sampling frequency, is the sampling frequency adjustment value;

[0122] If it is within the humidity threshold model interval, there is no need to adjust the sampling frequency.

[0123] Furthermore, if Figure 3 As shown, when constructing a BP neural network model for dynamically adjusting the sampling frequency, it includes:

[0124] In input layer 6 set Nodes , used to receive the preprocessed humidity sampling values;

[0125] In hidden layer 7, set Nodes ,choose function as the activation function, and set the connection weight between the input layer 6 and the hidden layer 7 to ;

[0126] At output layer 8 set Nodes ,choose function as the activation function, and set the connection weight between hidden layer 7 and output layer 8 to ;

[0127] When preprocessing the humidity sampling value, it includes:

[0128] The humidity sampling value is subjected to power frequency interference elimination and wavelet threshold denoising to obtain a denoised humidity sampling value;

[0129] Among them, the power frequency interference elimination includes: removing the fluctuation of the sensor capacitance signal through the reference capacitor, as shown in Formula 9:

[0130] (9);

[0131] In the formula, is the standard sampling digital quantity of the analog-to-digital converter, is the reference capacitance of the capacitive humidity sensor 2, is the capacitance signal of the capacitive humidity sensor 2, The capacitance signal of the capacitive humidity sensor 2 after removing the fluctuation;

[0132] Based on the mean value and standard deviation of historical humidity data, the denoised humidity sampling value is normalized, as shown in Formula 10:

[0133] (10);

[0134] In the formula, yes The humidity sampling value after de-noising at this moment, The BP neural network model The average value of historical humidity data at the input end. The BP neural network model The standard deviation of historical humidity data at the input end, is the normalized humidity sampling value.

[0135] Specifically, the present invention constructs a BP neural network model for dynamically adjusting the sampling frequency, and generates a humidity threshold model interval based on historical humidity data. The humidity prediction is output through the BP neural network model, and the humidity prediction value is compared with the humidity threshold model interval. The sampling frequency is dynamically adjusted according to the comparison result, thereby effectively reducing the sampling error generated in the sampling process and obtaining a more accurate humidity sampling value.

[0136] Specifically, the present invention performs power frequency interference elimination and wavelet threshold denoising on the humidity sampling value based on the reference capacitor, thereby effectively eliminating the fluctuation of the sensor capacitance signal and better retaining the useful signal, thereby obtaining more effective normalized data for humidity prediction.

[0137] In the above step S3, when the fast decoupling fuzzy control is used for fast adjustment, it includes:

[0138] According to the real-time humidity difference value and the target humidity range, the humidity deviation change rate is obtained, as shown in Formula 11:

[0139] (11);

[0140] In the formula, is the humidity deviation change rate, is the real-time humidity difference value, A preset humidity value within the target humidity range;

[0141] Will Parameter gain , , The basic domain and fuzzy domain are set as the target value, and the real-time humidity difference value and humidity deviation change rate are combined with the quantization factor and the target humidity range as input;

[0142] Based on the control rule table and fuzzy reasoning, the output Parameter gain , , ;

[0143] Output Parameter gain , , Combined with the scale factor and defuzzified to get the final Parameter gain , , ;

[0144] According to the final Parameter gain , , Adjustment In the controller , , parameter.

[0145] Specifically, the target value of the basic domain is , the target value of the fuzzy domain is The quantization factor of the real-time humidity difference value is 0.4, the quantization factor of the humidity deviation change rate is 0.7, and the proportional factor is 0.6.

[0146] Specifically, when the monitored humidity value differs greatly from the target humidity range, the present invention quickly adjusts the real-time humidity difference value to the preset humidity difference value (first-level humidity adjustment) through fast decoupling fuzzy control, and then cooperates with the predictive controller based on the neural network model to accurately adjust the humidity (second-level humidity adjustment), so that the humidity in the box gradually approaches the target value and accurately reaches the target value.

[0147] In the above step S4, when the humidity is gradually adjusted by using a prediction controller based on a neural network model, it includes:

[0148] Construct the future input sequence of the predictive controller in the control domain, as shown in Formula 12:

[0149] (12);

[0150] In the formula, For future input sequences, is the power sequence of the humidification module, is the power sequence of the dehumidification module, To control the time domain, for The humidification module power at the moment, for Dehumidification module power at the moment;

[0151] Input the future input sequence into the predictive controller to obtain the predicted time domain The control output under is shown in Formula 13:

[0152] (13);

[0153] In the formula, For the prediction time domain The control output under For the prediction time domain The humidification module power sequence is as follows: For the prediction time domain Dehumidification module power sequence under; For the prediction time domain Down Humidification module power at the moment, For the prediction time domain Down Dehumidification module power at the moment;

[0154] Among them, the target control sequence is obtained by minimizing the objective function, and the objective function is shown in formula 14:

[0155] (14);

[0156] In the formula, is the target value of the output; is the transpose symbol; is the tracking error, where is the tracking error of the humidification module, is the tracking error of the dehumidification module; is the weight of the control signal, where is the weight of the humidification module control signal, is the weight of the control signal of the dehumidification module; is the power increment, where is the power increment of the humidification module, is the power increment of the dehumidification module;

[0157] The target control sequence is used to output a smooth control signal to gradually adjust the humidity until a preset humidity value within the target humidity range is reached.

[0158] Furthermore, when minimizing the objective function to obtain the target control sequence, it includes:

[0159] A certain power constraint is given to the objective function, as shown in Formula 15:

[0160] (15);

[0161] In the formula, , The maximum and minimum power allowed for the humidification module to operate. , The maximum and minimum power allowed for the dehumidification module to operate. , It is the maximum rising power and falling power of the humidification module in each control cycle; , is the maximum rising power and falling power of the dehumidification module in each control cycle, is the power of the humidification module. is the power of the dehumidification module. is the power change value of the humidification module in each control cycle, is the power change value of the dehumidification module in each control cycle;

[0162] The objective function is minimized by the sequential quadratic programming method to obtain a control sequence within the power constraint, as shown in Formula 16:

[0163] (16);

[0164] In the formula, To control the cycle The power increase of the internal humidification module, To control the cycle The power increment of the internal dehumidification module, , , Based on the control cycle of the humidification module , , The difference between the internal humidity and the preset humidity value, , , Based on the dehumidification module control cycle , , The difference between the internal humidity and the preset humidity value, , , For humidifier The controller parameters, , , For dehumidifier Parameters of the controller.

[0165] Specifically, the neural network model used in the present invention is a recursive neural network, whose input is the control cycle under the control of the humidification module or the dehumidification module. , , The difference between the internal humidity and the preset humidity value is the output of the control cycle The power increase of the internal humidification module or dehumidification module.

[0166] Specifically, the present invention constructs a future input sequence for the predictive controller, inputs it into the predictive controller to obtain the control output in the predicted time domain, and obtains the target control sequence by minimizing the objective function, thereby using the target control sequence to output a smooth control signal, so that the humidity in the box gradually and smoothly approaches the target value, and finally reaches the target value, thereby realizing precise regulation of the humidity in the box.

[0167] In the above step S5, when the working cycle of humidification and dehumidification is adaptively optimized, it includes:

[0168] The outside temperature, outside humidity and inside temperature of the box are taken as inputs and mapped using a three-input single-output super closed ball CMAC network to obtain the inside humidity estimation model, as shown in Formula 17:

[0169] (17);

[0170] In the formula, is the estimated humidity in the box, is the temperature outside the box, is the humidity outside the box, is the temperature inside the box; It is a three-input single-output super-closed sphere CMAC network mapping;

[0171] Among them, input data is collected through a sliding data window to obtain an input data set;

[0172] Based on the outside temperature value, outside humidity value, inside temperature value, inside humidity value, and air flow rate value between the box and the outside, establish The control function is shown in Formula 18:

[0173] (18);

[0174] In the formula, is the air flow between the box and the outside world, is the humidity in the box;

[0175] Input the estimated value of the humidity inside the box output by the humidity estimation model into In the control function, we get Control optimization value;

[0176] based on Control optimization value optimizes the duty cycle of humidification and dehumidification.

[0177] Specifically, the present invention estimates the humidity in the box by constructing a box humidity estimation model, and combines Control function output Control optimization value to optimize the humidification and dehumidification working cycle, achieve efficient and accurate control of long-term humidity, and optimize energy consumption.

[0178] The intelligent moisture-proof box humidity control system provided by the present invention comprises: a frequency sampling dynamic adjustment unit, a first humidity adjustment unit, a second humidity adjustment unit and a humidity control optimization unit.

[0179] The sampling frequency dynamic adjustment unit is used to dynamically adjust the sampling frequency of the capacitive humidity sensor 2 according to the historical changes in humidity in the box;

[0180] The first humidity adjustment unit is used to compare the monitored humidity value with the target humidity range to obtain a real-time humidity difference value; when the real-time humidity difference value is greater than the preset humidity difference value, the real-time humidity difference value is quickly adjusted to the preset humidity difference value by using fast decoupling fuzzy control;

[0181] A second humidity adjustment unit: when the real-time humidity difference value is less than or equal to the preset humidity difference value, a prediction controller based on a neural network model is used to gradually adjust the humidity until a preset humidity value within the target humidity range is reached;

[0182] Humidity control optimization unit: When the monitored humidity value exceeds the target humidity range for a number of times that reaches a preset value, the humidification and dehumidification working cycles are adaptively optimized through a control algorithm based on environmental factors to achieve efficient and precise control of humidity.

[0183] Among them, the capacitive humidity sensor 2 is used to monitor the humidity inside the moisture-proof box 1 in real time based on the optimized LoRa.

[0184] The above-mentioned embodiments are only preferred embodiments of the present invention and cannot be used to limit the scope of protection of the present invention. Any non-substantial changes and substitutions made by technicians in this field on the basis of the present invention shall fall within the scope of protection required by the present invention.

Claims

1. A method for controlling humidity in an intelligent moisture-proof box, characterized in that: The following steps are involved: The capacitive humidity sensor is used to monitor the humidity inside the moisture-proof box in real time based on LoRa, and the sampling frequency of the capacitive humidity sensor is dynamically adjusted according to the historical changes in the humidity inside the box; Compare the monitored humidity value with the target humidity range to obtain the real-time humidity difference value; When the real-time humidity difference value is greater than the preset humidity difference value, the fast decoupling fuzzy control is used to quickly adjust the real-time humidity difference value to the preset humidity difference value; When the real-time humidity difference value is less than or equal to the preset humidity difference value, a prediction controller based on a neural network model is used to gradually adjust the humidity until a preset humidity value within the target humidity range is reached; When the monitored humidity value exceeds the target humidity range for a preset number of times, the humidification and dehumidification working cycles are adaptively optimized through the control algorithm according to environmental factors to achieve efficient and accurate humidity control; Among them, when real-time monitoring of the humidity inside the moisture-proof box is carried out, it includes: Introduce the LoRa radio frequency module and connect it with the transmission controller to build a radio frequency layer, and define the communication protocol working frequency band in the radio frequency layer; In the physical layer, the continuous time of each frequency modulation signal is determined by the spreading factor and bandwidth, and linear spread spectrum frequency modulation is performed; The transmission rate of LoRa is determined by the spreading factor, coding rate and bandwidth at the medium access control layer; In the dormant state, communication is performed through the SPI or UART interface. After initialization, the wireless communication module enters the standby and transceiver states, and the carrier frequency is configured for the wireless communication module; Initialize data transmission, generate control instructions based on environmental parameters, and send real-time humidity monitoring data of the moisture-proof box; When generating control instructions based on environmental parameters, it includes: Generate different humidity monitoring instructions based on historical changes in humidity in the box and dynamically adjust the sampling frequency of the capacitive humidity sensor; The dynamic adjustment process includes: Construct a BP neural network model for dynamically adjusting the sampling frequency, and select sample data for training; Sampling data at an initial sampling frequency, and preprocessing the humidity sampling value before inputting it into the BP neural network model for humidity prediction to obtain a humidity prediction value; Compare the humidity prediction value with the humidity threshold model interval generated based on historical humidity data, and dynamically adjust the sampling frequency; When adaptively optimizing the humidification and dehumidification working cycles, including: The outside temperature, outside humidity and inside temperature of the box are taken as inputs and mapped using a three-input single-output super closed ball CMAC network to obtain a model for estimating the humidity inside the box. Among them, input data is collected through a sliding data window to obtain an input data set; Based on the temperature inside and outside the box, the humidity inside and outside the box, and the air flow rate between the box and the outside, establish Control functions; The estimated value of the humidity in the box output by the humidity estimation model is input into In the control function, we get Control optimization value; Based on the Control optimization value optimizes the duty cycle of humidification and dehumidification.

2. The intelligent moisture-proof box humidity control method according to claim 1 is characterized in that: When sending the real-time humidity monitoring data of the moisture proof box, it includes: Determine the first and second random number seeds according to the moisture-proof box to be monitored, and generate a third random number seed; The first, second and third random number seeds form a random number sequence and add it to the timer for encryption of real-time humidity data; When sending next time, the first and second random number seeds are updated, and a new third random number seed and a new random number sequence are generated, and the new random number sequence is used to encrypt the real-time humidity data.

3. The intelligent moisture-proof box humidity control method according to claim 2 is characterized in that: When building a BP neural network model, it includes: Set at the input layer Nodes , used to receive the preprocessed humidity sampling values; Set in hidden layer Nodes ,choose The function is used as the activation function, and the connection weights between the input layer and the hidden layer are set to ; Set at the output layer Nodes ,choose The function is used as the activation function, and the connection weights between the hidden layer and the output layer are set to ; When preprocessing the humidity sampling value, it includes: The humidity sampling value is subjected to power frequency interference elimination and wavelet threshold denoising to obtain a denoised humidity sampling value; The denoised humidity sampling values ​​are normalized based on the mean and standard deviation of historical humidity data.

4. The intelligent moisture-proof box humidity control method according to any one of claims 1 to 3, characterized in that: When using fast decoupling fuzzy control for fast adjustment, it includes: According to the real-time humidity difference value and the target humidity range, the humidity deviation change rate is obtained; Will Parameter gain , , The basic domain and fuzzy domain are set as the target value, and the real-time humidity difference value and humidity deviation change rate are combined with the quantization factor and the target humidity range as input; Based on the control rule table and fuzzy reasoning, the output Parameter gain , , , and after combining with the scale factor, defuzzification is performed to obtain the final Parameter gain , , ; According to the final Parameter gain , , Adjustment In the controller , , parameter.

5. The intelligent moisture-proof box humidity control method according to claim 4, characterized in that: When using a predictive controller based on a neural network model to gradually adjust humidity, it includes: Construct the future input sequence of the predictive controller in the control horizon; Input the future input sequence into the predictive controller to obtain the predicted time domain The control output under The target control sequence is obtained by minimizing the target function, and a smooth control signal is outputted using the target control sequence to gradually adjust the humidity until a preset humidity value within the target humidity range is reached.

6. The intelligent moisture-proof box humidity control method according to claim 5, characterized in that: When minimizing the objective function to obtain the target control sequence, it includes: Assigning a certain power constraint to the objective function; The objective function is minimized by a sequential quadratic programming method to obtain a control sequence within the power constraint, as shown in Formula 16: （16）； In the formula, To control the cycle The power increase of the internal humidification module, To control the cycle The power increment of the internal dehumidification module, , , and , , The control cycles are based on the humidification module and the dehumidification module respectively. , , The difference between the internal humidity and the preset humidity value, , , and , , Humidifier and dehumidifier Parameters of the controller.

7. An intelligent moisture-proof box humidity control system, characterized in that: include: Sampling frequency dynamic adjustment unit: used to dynamically adjust the sampling frequency of the capacitive humidity sensor according to the historical changes in humidity in the box; The first humidity adjustment unit is used to compare the monitored humidity value with the target humidity range to obtain a real-time humidity difference value; when the real-time humidity difference value is greater than the preset humidity difference value, the real-time humidity difference value is quickly adjusted to the preset humidity difference value by using fast decoupling fuzzy control; A second humidity adjustment unit: when the real-time humidity difference value is less than or equal to the preset humidity difference value, a prediction controller based on a neural network model is used to gradually adjust the humidity until a preset humidity value within the target humidity range is reached; Humidity control optimization unit: When the monitored humidity value exceeds the target humidity range for a preset number of times, the humidification and dehumidification working cycles are adaptively optimized through the control algorithm according to environmental factors to achieve efficient and accurate humidity control; Among them, the humidity inside the moisture-proof box is monitored in real time based on LoRa through a capacitive humidity sensor; When real-time monitoring of the humidity inside the moisture-proof box, including: Introduce the LoRa radio frequency module and connect it with the transmission controller to build a radio frequency layer, and define the communication protocol working frequency band in the radio frequency layer; In the physical layer, the continuous time of each frequency modulation signal is determined by the spreading factor and bandwidth, and linear spread spectrum frequency modulation is performed; The transmission rate of LoRa is determined by the spreading factor, coding rate and bandwidth at the medium access control layer; In the dormant state, communication is performed through the SPI or UART interface. After initialization, the wireless communication module enters the standby and transceiver states, and the carrier frequency is configured for the wireless communication module; Initialize data transmission, generate control instructions based on environmental parameters, and send real-time humidity monitoring data of the moisture-proof box; When generating control instructions based on environmental parameters, it includes: Generate different humidity monitoring instructions based on historical changes in humidity in the box and dynamically adjust the sampling frequency of the capacitive humidity sensor; The dynamic adjustment process includes: Construct a BP neural network model for dynamically adjusting the sampling frequency, and select sample data for training; Sampling data at an initial sampling frequency, and preprocessing the humidity sampling value before inputting it into the BP neural network model for humidity prediction to obtain a humidity prediction value; Compare the humidity prediction value with the humidity threshold model interval generated based on historical humidity data, and dynamically adjust the sampling frequency; When adaptively optimizing the humidification and dehumidification working cycles, including: The outside temperature, outside humidity and inside temperature of the box are taken as inputs and mapped using a three-input single-output super closed ball CMAC network to obtain a model for estimating the humidity inside the box. Among them, input data is collected through a sliding data window to obtain an input data set; Based on the temperature inside and outside the box, the humidity inside and outside the box, and the air flow rate between the box and the outside, establish Control functions; The estimated value of the humidity in the box output by the humidity estimation model is input into In the control function, we get Control optimization value; Based on the Control optimization value optimizes the duty cycle of humidification and dehumidification.

Images