Ice maker intelligent control method and system based on AI analysis and storage medium

Abstract

The invention relates to the technical field of intelligent control, and discloses an ice maker intelligent control method and system based on AI analysis, and a storage medium, and the method comprises the steps: real-time collection of water storage state data, AI dynamic analysis of water storage requirements, dynamic adjustment of a refrigeration path and ice making parameters, optimal control of an ice making process, and abnormal state early warning and self-repairing. The system comprises a multi-source sensor module, a data acquisition and preprocessing module, an AI analysis and decision module, a dynamic priority scheduling algorithm, a control execution module and a man-machine interaction module, and a storage medium comprises a real-time control thread, a data storage thread and a remote upgrade interface. The water storage temperature can be adjusted in real time according to actual use requirements, the problem of temperature fluctuation caused by the fact that a traditional ice maker depends on fixed temperature setting is solved, the stability of the water storage temperature is remarkably improved, and therefore the quality of drinking water and ice cubes output by the ice maker is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent control, and specifically provides an intelligent control method, system and storage medium for an ice maker based on AI analysis. Background Art

[0002] With the improvement of people's living standards and the pursuit of a healthy lifestyle, people pay more and more attention to various diets in life. Among them, drinking water, as the basis of diet, is the most important thing for people. Among them, the ice maker, as an indispensable household appliance in daily life, its main function is to provide convenient drinking water services, including normal temperature water, cold water and ice cubes, etc. Most traditional ice makers use a refrigeration system to directly cool the water in the water tank to produce cold water, and further generate ice cubes through an ice making module.

[0003] After retrieval, the patent with the Chinese patent number CN118794177A discloses an intelligent control method and device for an ice maker and an ice maker. By implementing the present invention, when detecting the usage requirement for the ice maker, it can automatically collect the water storage information of the hot water component and determine whether the water storage has been heated to the first water supply temperature before the current water storage. If so, it will generate a control instruction corresponding to the usage requirement, and then execute equipment control on the ice maker according to the control instruction. The collection of the water storage information and the comparison with the first water supply temperature ensure that when performing the equipment control operation, the current water storage of the hot water component will surely meet the subsequent water supply requirements, improving the water supply accuracy of the current water storage; after the current water storage reaches the first water supply temperature, it is beneficial to improve the cleanliness of the drinking water / ice cubes output by the ice maker; in addition, the equipment control operation can also realize the intelligent adjustment of the refrigeration path.

[0004] However, the water storage temperature management of the above-mentioned ice maker usually depends on fixed temperature settings and cannot be dynamically adjusted according to actual usage requirements, resulting in large fluctuations in the water storage temperature and affecting the quality of the drinking water and ice cubes output by the ice maker; in addition, traditional ice makers lack real-time monitoring and intelligent analysis of the water storage state during the ice making process, and it is easy to start ice making when the water storage temperature does not reach the ideal state, thereby affecting the shape and cleanliness of the ice cubes. Based on this, the present invention designs an intelligent control method, system and storage medium for an ice maker based on AI analysis to solve the above problems. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent control method, system and storage medium for an ice maker based on AI analysis, which solves the problem that it cannot be adjusted according to actual usage requirements in the background art.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: An intelligent control method, system and storage medium for an ice maker based on AI analysis, including: Step S1, collect the water storage status data in real time. Collect the water temperature data in the water storage tank through a temperature sensor, with a collection frequency of once per second; monitor the water level of the water storage tank through a water level sensor, and set the trigger threshold to the lowest water level of 100 mm; detect the water quality parameters through a turbidity sensor, and trigger the cleaning mode when the turbidity reaches 50 NTU.

[0007] Step S2, perform AI dynamic analysis of the water storage demand. Input the real-time collected water storage temperature, water level, and water quality data into a prediction model based on a long short-term memory network. The number of nodes in the input layer of the model is 3, the number of nodes in the hidden layer is 64, and the output layer is the classification label of the water storage demand. The classification labels include immediate ice making, delayed ice making, and cleaning priority.

[0008] Step S3, dynamically adjust the refrigeration path and ice making parameters. When the AI determines immediate ice making, start the two-stage compression refrigeration system and reduce the water storage temperature from the initial temperature of 20 °C to the target temperature of 4 °C; when the AI determines delayed ice making, start the energy-saving mode and maintain the water storage temperature at 8 °C; when the AI determines cleaning priority, activate the ultraviolet sterilization module and simultaneously start the drain pump for water replacement.

[0009] Step S4, optimize the control of the ice making process. Use the fuzzy PID algorithm to adjust the blade speed of the ice making module, and dynamically adjust the cutting frequency according to the feedback data of the ice thickness sensor; adjust the ice making cycle according to the ambient temperature, shorten the ice making cycle when the ambient temperature is higher than 25 °C, and extend the ice making cycle when the ambient temperature is lower than 15 °C.

[0010] Preferably, it further includes Step S5, abnormal state warning and self-repair. Monitor the operating state of the compressor through a vibration sensor, and trigger a warning when the vibration amplitude reaches 0.5 g; if the water temperature does not drop by 2 °C within 5 minutes after the refrigeration starts, it is determined that the refrigerant leaks and a system shutdown command is triggered.

[0011] Preferably, in the said Step S2, the training process of the long short-term memory network prediction model is as follows: Step S201, collect the water storage temperature, water level, and water quality data every 15 minutes in the past at least 1 year, as well as the corresponding water storage demand labels, as the training data set.

[0012] Step S202, divide the training data set into a training set, a validation set, and a test set, with proportions of 70%, 15%, and 15% respectively.

[0013] Step S203, use the cross-entropy loss function as the loss function to measure the difference between the classification label predicted by the model and the actual label.

[0014] Step S204, during the training process, when the loss value on the validation set no longer decreases within 5 consecutive training epochs, stop the training in advance to prevent overfitting.

[0015] Preferably, in step S2, the two-stage compression refrigeration system includes: Two groups of compressors, namely a high-pressure compressor and a low-pressure compressor. The displacement of the high-pressure compressor is 10 m³ / h, and the displacement of the low-pressure compressor is 8 m³ / h.

[0016] Two groups of condensers, namely a high-pressure condenser and a low-pressure condenser, with condensation areas of 5 m² and 4 m² respectively.

[0017] Two groups of evaporators, namely a high-pressure evaporator and a low-pressure evaporator, with evaporation areas of 3 m² and 2 m² respectively.

[0018] The refrigerant of the two-stage compression refrigeration system uses R410A, and its charging amount is accurately calculated and adjusted according to the system design and actual operation conditions to ensure the efficient operation of the refrigeration system.

[0019] Preferably, in step S4, the specific parameter settings of the fuzzy PID algorithm are as follows: the initial value of the proportional coefficient is 0.6, which is dynamically adjusted according to the deviation of the ice thickness and does not exceed 1.2 at most; the initial value of the integral coefficient is 0.1, which is used to eliminate the steady-state error of the ice thickness and is adjusted according to the accumulated amount of the deviation, not exceeding 0.3 at most; the initial value of the differential coefficient is 0.05, which is used to predict the change trend of the ice thickness and is adjusted according to the change rate of the deviation, not exceeding 0.1 at most.

[0020] Preferably, an intelligent control system for an ice maker based on AI analysis includes the following modules: Module S1, a multi-source sensor module, including a temperature sensor group, which includes a water storage tank temperature sensor and an ambient temperature sensor; it also includes a water quality monitoring module, which integrates a turbidity sensor and a pH sensor.

[0021] Module S2, a data acquisition and preprocessing module, which uses a microcontroller for data acquisition with a sampling frequency of 1 kHz; data preprocessing includes Kalman filtering and median filtering.

[0022] Module S3, an AI analysis and decision-making module, a long short-term memory network model deployed on edge computing devices, with an inference delay not exceeding 50 milliseconds.

[0023] Module S4, a dynamic priority scheduling algorithm, which adjusts the refrigeration strategy according to the user's preset preferences.

[0024] Module S5, a control execution module, a refrigeration system drive circuit, which uses PWM voltage regulation technology to control the compressor speed; a mechanical execution mechanism, including a blade assembly driven by a stepper motor and a drainage passage controlled by a solenoid valve.

[0025] Module S6, a human-machine interaction module, with a touch screen to display real-time data and an operation interface; supporting voice command recognition, and the response time does not exceed 0.5 seconds.

[0026] Preferably, in the module S4, the dynamic priority scheduling algorithm further includes the following parameters: according to the user's preset preferences, the refrigeration strategies are divided into three modes: high-efficiency priority, low-energy consumption priority, and fast refrigeration priority; in the high-efficiency priority mode, the energy efficiency ratio of the refrigeration system is not less than 3.5; in the low-energy consumption priority mode, the energy consumption reduction of the refrigeration system is not less than 20%.

[0027] Preferably, in the module S5, the refrigeration system drive circuit uses PWM voltage regulation technology to control the compressor speed, the frequency of the PWM signal is 20kHz, and the duty cycle adjustment range is 10%-90%; the response time of the drainage passage controlled by the solenoid valve does not exceed 50ms to improve the accuracy of the solenoid valve.

[0028] Preferably, an intelligent control storage medium for an ice maker based on AI analysis, the storage medium includes the following program codes: Program S1, a real-time control thread, with the priority set to the highest task level of the real-time operating system, and the scheduling period is 1ms; including a fuzzy PID control instruction set for adjusting the compressor speed and the blade cutting frequency.

[0029] Program S2, a data storage thread, using a circular buffer mechanism, the buffer size is 1MB, the data persistence period is 60 minutes, and the stored content includes sensor raw data and control logs.

[0030] Program S3, a remote upgrade interface, supporting communication with the OTA server using the HTTPS protocol and a firmware differential upgrade algorithm; the firmware verification uses the SHA-256 hash algorithm, and automatically rolls back to the previous stable version when the upgrade fails.

[0031] Preferably, in the program S2, the sensor raw data and control logs further include a log hierarchical storage strategy. The running status log retention period is 30 days, stored in CSV format; the abnormal alarm log is stored permanently, encrypted and uploaded to the cloud; the log retrieval function supports time range query and keyword filtering.

[0032] Compared with the prior art, the beneficial effects achieved by the present invention are: 1. In the present invention, by dynamically analyzing the water storage demand through AI, the water storage temperature can be adjusted in real time according to the actual usage demand, avoiding the temperature fluctuation problem caused by the traditional ice maker relying on fixed temperature settings. During high-demand periods, the system quickly cools to the target temperature; during low-demand periods, it maintains an energy-saving mode, significantly improving the stability of the water storage temperature, thereby ensuring the quality of the drinking water and ice cubes output by the ice maker.

[0033] 2. In the present invention, the state data of the water storage tank and the compressor are collected in real time through various sensors such as temperature, water level, turbidity, pH value, and vibration, and intelligent analysis is carried out in combination with the long short-term memory network model. This multi-dimensional real-time monitoring and dynamic decision-making mechanism can effectively avoid the situation of starting to make ice when the water storage temperature does not reach the ideal state, thus significantly improving the shape and cleanliness of the ice cubes.

[0034] 3. In the present invention, by adopting a two-stage compression refrigeration system, the high-pressure compressor and the low-pressure compressor work together, and cooperate with the optimized condenser and evaporator, significantly improving the refrigeration efficiency. In the high-efficiency priority mode, at the same time, by dynamically adjusting the refrigeration path and ice-making parameters, the system quickly cools during high demand and reduces energy consumption during low demand, achieving a balance between high efficiency and energy saving. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 is the flow chart of the intelligent control method of the ice maker of the present invention; Figure 2 is the flow chart of the training of the long short-term memory network prediction model of the present invention; Figure 3 is the structural schematic diagram of the two-stage compression refrigeration system of the present invention; Figure 4 is the schematic diagram of the fuzzy PID algorithm of the present invention; Figure 5 is the structural composition diagram of the intelligent control system module of the ice maker of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0036] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0037] Embodiment 1; Please refer to Figures 1 - 5, An intelligent control method for an ice maker based on AI analysis, including: Step S1, collecting water storage state data in real time, collecting the water temperature data in the water storage tank through a temperature sensor, with a collection frequency of 1 time per second; monitoring the water level in the water storage tank through a water level sensor, and setting the trigger threshold to the lowest water level of 100 mm; detecting water quality parameters through a turbidity sensor, and triggering the cleaning mode when the turbidity reaches 50 NTU; Step S2, AI dynamically analyzes the water storage demand, inputs the real-time collected water storage temperature, water level and water quality data into a prediction model based on a long short-term memory network. The number of nodes in the input layer of the model is 3, the number of nodes in the hidden layer is 64, and the output layer is the classification label of the water storage demand. The classification labels include immediate ice making, delayed ice making and cleaning priority; Step S3, dynamically adjusting the refrigeration path and ice making parameters. When AI determines immediate ice making, start the two-stage compression refrigeration system to reduce the water storage temperature from the initial temperature of 20 °C to the target temperature of 4 °C; when AI determines delayed ice making, start the energy-saving mode to maintain the water storage temperature at 8 °C; when AI determines cleaning priority, activate the ultraviolet sterilization module and simultaneously start the drainage pump to change the water; Step S4, optimizing the control of the ice making process, using the fuzzy PID algorithm to adjust the blade speed of the ice making module, and dynamically adjusting the cutting frequency according to the feedback data of the ice thickness sensor; adjusting the ice making cycle according to the ambient temperature, shortening the ice making cycle when the ambient temperature is higher than 25 °C, and lengthening the ice making cycle when the ambient temperature is lower than 15 °C. It also includes Step S5, abnormal state warning and self-repair. Monitor the operating state of the compressor through a vibration sensor, and trigger a warning when the vibration amplitude reaches 0.5 g; if the water temperature does not drop by 2 °C within 5 minutes after the refrigeration starts, it is determined that the refrigerant leaks and a system shutdown command is triggered.

[0038] In Step S2, the training process of the long short-term memory network prediction model is as follows: Step S201, collect the water storage temperature, water level and water quality data at intervals of 15 minutes in the past at least 1 year, as well as the corresponding water storage demand labels, as the training data set; Step S202, divide the training data set into a training set, a validation set and a test set, with proportions of 70%, 15% and 15% respectively; Step S203, use the cross-entropy loss function as the loss function to measure the difference between the classification label predicted by the model and the actual label; Step S204, during the training process, when the loss value on the validation set no longer decreases in 5 consecutive training cycles, stop training in advance to prevent overfitting.

[0039] In step S2, the two-stage compression refrigeration system includes: two groups of compressors, namely a high-pressure compressor and a low-pressure compressor. The displacement of the high-pressure compressor is 10 m³ / h, and the displacement of the low-pressure compressor is 8 m³ / h; two groups of condensers, namely a high-pressure condenser and a low-pressure condenser, with condensation areas of 5 m² and 4 m² respectively; two groups of evaporators, namely a high-pressure evaporator and a low-pressure evaporator, with evaporation areas of 3 m² and 2 m² respectively. The refrigerant of the two-stage compression refrigeration system uses R410A, and its charging amount is accurately calculated and adjusted according to the system design and actual operation conditions to ensure the efficient operation of the refrigeration system.

[0040] In step S4, the specific parameter settings of the fuzzy PID algorithm are as follows: the initial value of the proportional coefficient is 0.6, which is dynamically adjusted according to the deviation of the ice thickness and does not exceed 1.2 at most; the initial value of the integral coefficient is 0.1, which is used to eliminate the steady-state error of the ice thickness and is adjusted according to the accumulated amount of the deviation and does not exceed 0.3 at most; the initial value of the differential coefficient is 0.05, which is used to predict the change trend of the ice thickness and is adjusted according to the change rate of the deviation and does not exceed 0.1 at most.

[0041] The working principle of the embodiment of the present invention is: the system collects key parameters in the water storage tank in real time through the multi-source sensor module, including water temperature, water level, turbidity and pH value, and simultaneously monitors the environmental temperature and the vibration state of the compressor. The acquisition frequency of the sensor data is 1 time per second to ensure the real-time and accuracy of the data. After the data is collected, it is preprocessed by the microcontroller, and Kalman filtering and median filtering technologies are used to eliminate noise interference to ensure the data quality input to the AI analysis module. For example, when the turbidity sensor detects that the turbidity reaches 50 NTU, the system will trigger the cleaning mode and start the ultraviolet sterilization module and the drainage pump to automatically change the water.

[0042] The collected water storage temperature, water level and water quality data are input into the prediction model based on the long short-term memory network. This model learns the dynamic change law of water storage demand by training historical data; the number of nodes in the input layer of the model is 3, the number of nodes in the hidden layer is 64, and the output layer is the classification label of the water storage demand; the model optimizes the prediction accuracy through the cross-entropy loss function and stops training when the loss value of the validation set no longer decreases for 5 consecutive cycles to prevent overfitting.

[0043] Embodiment 2; Please refer to Figures 1 - 5, in the embodiments of the present invention, an intelligent control system for an ice maker based on AI analysis includes the following modules: Module S1, a multi-source sensor module, including a temperature sensor group, which includes a water storage tank temperature sensor and an ambient temperature sensor; it also includes a water quality monitoring module, integrating a turbidity sensor and a pH sensor; Module S2, a data acquisition and preprocessing module, using a microcontroller for data acquisition, with a sampling frequency of 1 kHz; data preprocessing includes Kalman filtering and median filtering; Module S3, an AI analysis and decision-making module, a long short-term memory network model deployed on an edge computing device, with an inference delay not exceeding 50 milliseconds; Module S4, a dynamic priority scheduling algorithm, which adjusts the refrigeration strategy according to user preset preferences; Module S5, a control execution module, a refrigeration system drive circuit, using PWM voltage regulation technology to control the compressor speed; a mechanical execution mechanism, including a blade assembly driven by a stepper motor and a drainage passage controlled by a solenoid valve; Module S6, a human-machine interaction module, a touch screen for displaying real-time data and an operation interface; supporting voice command recognition, with a response time not exceeding 0.5 seconds.

[0044] In Module S4, the dynamic priority scheduling algorithm further includes the following parameters: According to user preset preferences, the refrigeration strategy is divided into three modes: high efficiency priority, low energy consumption priority, and fast refrigeration priority; in the high efficiency priority mode, the energy efficiency ratio of the refrigeration system is not less than 3.5; in the low energy consumption priority mode, the energy consumption of the refrigeration system is reduced by not less than 20%. In Module S5, the refrigeration system drive circuit uses PWM voltage regulation technology to control the compressor speed, the frequency of the PWM signal is 20 kHz, and the duty cycle adjustment range is 10% - 90%; the response time of the drainage passage controlled by the solenoid valve does not exceed 50 ms to improve the accuracy of the solenoid valve.

[0045] An intelligent control storage medium for an ice maker based on AI analysis, the storage medium includes the following program codes: Program S1, a real-time control thread, with the priority set to the highest task level of the real-time operating system, and a scheduling period of 1 ms; including a fuzzy PID control instruction set for adjusting the compressor speed and the blade cutting frequency; Program S2, a data storage thread, using a circular buffer mechanism, with a buffer size of 1 MB and a data persistence period of 60 minutes, and the stored content includes sensor raw data and control logs; Program S3, a remote upgrade interface, supporting communication with an OTA server using the HTTPS protocol, and a firmware differential upgrade algorithm; firmware verification uses the SHA-256 hash algorithm, and automatically rolls back to the previous stable version when the upgrade fails. In Program S2, the sensor raw data and control logs further include a log hierarchical storage strategy, the running status log retention period is 30 days, stored in CSV format; the abnormal alarm log is stored permanently, encrypted and uploaded to the cloud; a log retrieval function, supporting time range query and keyword filtering.

[0046] The working principle of the embodiments of the present invention is as follows: According to the AI analysis results, the system dynamically adjusts the refrigeration path and ice-making parameters. Immediate ice-making mode: When the AI determines "immediate ice-making", the two-stage compression refrigeration system is started. The system includes a high-pressure compressor (displacement 10 m³ / h) and a low-pressure compressor (displacement 8 m³ / h), cooperating with a high-pressure condenser (5 m²), a low-pressure condenser (4 m²), a high-pressure evaporator (3 m²), and a low-pressure evaporator (2 m²), using R410A as the refrigerant. The system reduces the water storage temperature from the initial temperature of 20°C to the target temperature of 4°C to meet the rapid ice-making demand; Delayed ice-making mode: When the AI determines "delayed ice-making", the system starts the energy-saving mode to maintain the water storage temperature at 8°C and reduce energy consumption; Cleaning priority mode: When the water quality parameters (such as turbidity) reach the threshold, the system preferentially starts the cleaning mode, activates the ultraviolet sterilization module and simultaneously starts the drain pump to change the water, ensuring that the water quality in the water storage tank meets the standards.

[0047] During the ice-making process, the system uses a fuzzy PID algorithm to optimize the ice-making parameters. Ice thickness control: The ice thickness is monitored in real time through an ice thickness sensor, and the blade rotation speed and cutting frequency are dynamically adjusted. In the fuzzy PID algorithm, the initial value of the proportional coefficient is 0.6 (not exceeding 1.2 at most), the initial value of the integral coefficient is 0.1 (not exceeding 0.3 at most), and the initial value of the differential coefficient is 0.05 (not exceeding 0.1 at most), ensuring uniform ice thickness and reasonable cutting frequency; Ambient temperature adaptation: The ice-making cycle is dynamically adjusted according to the ambient temperature. When the ambient temperature is higher than 25°C, the ice-making cycle is shortened to improve the refrigeration efficiency; when the ambient temperature is lower than 15°C, the ice-making cycle is extended to reduce energy consumption.

[0048] Embodiment 3; Please refer to Figures 1 - 5 , in the embodiments of the present invention, a specific application of an ice maker is provided: Multi-source data acquisition. An ice maker is deployed in the back kitchen of a restaurant. Its sensor module (module S1) collects data at a frequency of 1 kHz. The water temperature sensor in the water storage tank (accuracy of ±0.2°C) monitors the water temperature from the initial 25°C to the target 4°C. The trigger threshold of the water level sensor is set to 100 mm, and when the water level is lower than the threshold, it automatically replenishes water. The turbidity sensor detects the water quality in real time, and when the NTU value exceeds 50, it triggers the cleaning mode. The vibration sensor monitors the operation of the compressor, and the 0.5g amplitude warning threshold prevents mechanical failures.

[0049] AI dynamic decision-making. The edge computing device (module S3) runs an LSTM prediction model (3 nodes in the input layer and 64 nodes in the hidden layer), outputs decision labels every 50 ms. In the immediate ice-making mode, when the predicted demand surges during the dinner peak period, the two-stage compressor (high-pressure exhaust volume of 10 m³ / h + low-pressure exhaust volume of 8 m³ / h) is started, and the water temperature is reduced from 20 °C to 4 °C within 5 minutes. In the delayed ice-making mode, the water temperature is maintained at 8 °C during the low-demand period at noon, the compressor speed is reduced by 40%, and the energy consumption is reduced by 22%. In the cleaning priority mode, when the turbidity exceeds the standard, the ultraviolet germicidal module (wavelength 254 nm) is activated and the drain pump (flow rate 3 L / s) is started to complete the water replacement.

[0050] Execute optimization control. The fuzzy PID algorithm (program S1) dynamically adjusts the blade speed (adjustable from 500 to 1200 rpm). When the ice thickness sensor detects the standard value of 8 mm, the proportional coefficient is adjusted to 0.8 and the integral coefficient is 0.15 to achieve a cutting accuracy of ±0.3 mm. When the ambient temperature sensor detects 28 °C, the ice-making cycle is shortened to 15 minutes per batch, and the efficiency is increased by 35% compared to the normal temperature mode.

[0051] Abnormal handling and upgrade. When the temperature drop is less than 2 °C within 5 minutes due to refrigerant leakage (step S5), the system automatically shuts down and pushes an alarm to the human-machine interaction screen (module S6). The storage medium (program S3) completes OTA upgrade through the HTTPS protocol, and uses SHA-256 verification to ensure the security of the firmware. When the upgrade fails, it rolls back to the stable version within 50 ms.

[0052] The system modules operate in coordination. For dynamic priority scheduling (module S4), when the user selects the "high efficiency priority" mode, the two-stage condenser (area of 5 m² + 4 m²) works together, and the energy efficiency ratio reaches 3.8, which is 42% higher than that of the single-stage system. The PWM drive control (module S5) adjusts the compressor duty cycle (30% - 85%) at a frequency of 20 kHz, saving 18% energy compared to the traditional relay control. The data persistence (program S2) uses a circular buffer to store 30 days of operation logs, encrypts the abnormal records and uploads them to the cloud, and shortens the fault diagnosis response time to 2 minutes.

[0053] The working principle of the embodiments of the present invention is as follows: The AI prediction model reduces the start-stop times of the refrigeration system by 62%, and reduces the annual electricity cost by about 1,200 yuan (calculated at 0.8 yuan / kWh). The vibration warning mechanism reduces the incidence rate of compressor mechanical failures from 5% to 0.3%. In the water quality compliance rate, the ultraviolet germicidal module cooperates with real-time water replacement, and the water quality qualification rate is increased from 82% to 99.6%. The OTA remote upgrade reduces the on-site maintenance frequency, and the annual operation and maintenance cost is reduced by 45%.

[0054] Working principle: The system collects key parameters such as water temperature, water level, turbidity, pH value in the water storage tank, and environmental temperature, compressor vibration status, etc. in real time through a multi-source sensor module. After being preprocessed by the microcontroller using Kalman filtering and median filtering techniques, they are input into the long short-term memory network model of the AI analysis and decision-making module. This model learns the dynamic change law of water storage demand based on historical data and outputs water storage demand classification labels. According to the AI analysis results, the system dynamically adjusts the refrigeration path and ice-making parameters: In the immediate ice-making mode, the two-stage compression refrigeration system is started, using high-pressure and low-pressure compressors, condensers, evaporators, and R410A refrigerant to reduce the water storage temperature from 20°C to 4°C; in the delayed ice-making mode, the energy-saving mode is started to maintain the water storage temperature at 8°C; in the cleaning priority mode, the ultraviolet sterilization module is activated and the drain pump is started to change water. During the ice-making process, the fuzzy PID algorithm is used to optimize the ice-making parameters. The ice thickness sensor is used to monitor the ice thickness in real time, dynamically adjusting the blade speed and cutting frequency. At the same time, the ice-making cycle is dynamically adjusted according to the environmental temperature, shortening when it is higher than 25°C and lengthening when it is lower than 15°C to achieve efficient, energy-saving and high-quality ice-making effects. In addition, the system has an abnormal state warning and self-repair function, monitoring the operation status of the compressor through a vibration sensor, and judging refrigerant leakage and shutting down the system when the water temperature is abnormal.

[0055] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirits of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent control method for ice making machine based on AI analysis, characterized in that: include: Step S1, real-time collection of water storage status data, the water temperature data in the water tank is collected through the temperature sensor, and the collection frequency is once per second; the water level of the water tank is monitored through the water level sensor, and the trigger threshold is set to the minimum water level of 100 mm; the water quality parameters are detected through the turbidity sensor, and the cleaning mode is triggered when the turbidity reaches 50NTU; Step S2, AI dynamically analyzes water storage demand, and inputs the real-time collected water storage temperature, water level and water quality data into a prediction model based on a long short-term memory network. The number of model input layer nodes is 3, the number of hidden layer nodes is 64, and the output layer is the classification label of water storage demand, which includes immediate ice making, delayed ice making and cleaning priority; Step S3, dynamically adjust the refrigeration path and ice-making parameters. When AI determines that ice-making is immediate, start the two-stage compression refrigeration system to reduce the water storage temperature from the initial temperature of 20 degrees Celsius to the target temperature of 4 degrees Celsius; when AI determines that ice-making is delayed, start the energy-saving mode to maintain the water storage temperature at 8 degrees Celsius; when AI determines that cleaning is prioritized, activate the ultraviolet disinfection module and synchronously start the drainage pump to change the water; Step S4, optimizing the control of the ice-making process, using the fuzzy PID algorithm to adjust the blade speed of the ice-making module, and dynamically adjusting the cutting frequency according to the feedback data from the ice thickness sensor; adjusting the ice-making cycle according to the ambient temperature, shortening the ice-making cycle when the ambient temperature is higher than 25 degrees Celsius, and extending the ice-making cycle when the ambient temperature is lower than 15 degrees Celsius.

2. The ice making machine intelligent control method based on AI analysis according to claim 1 is characterized in that: It also includes step S5, abnormal state warning and self-repair, which monitors the operating status of the compressor through a vibration sensor and triggers a warning when the vibration amplitude reaches 0.5g; if the water temperature does not drop by 2 degrees Celsius within 5 minutes after the refrigeration is started, it is determined to be a refrigerant leak and a system shutdown command is triggered.

3. The ice making machine intelligent control method based on AI analysis according to claim 1 is characterized in that: In step S2, the training process of the long short-term memory network prediction model is as follows: Step S201, collecting water storage temperature, water level and water quality data at intervals of 15 minutes for at least the past year, as well as corresponding water storage demand labels, as a training data set; Step S202, dividing the training data set into a training set, a validation set, and a test set, with the proportions being 70%, 15%, and 15% respectively; Step S203: The loss function adopts a cross entropy loss function to measure the difference between the classification label predicted by the model and the actual label; Step S204: During the training process, when the loss value on the validation set no longer decreases within five consecutive training cycles, the training is stopped in advance to prevent overfitting.

4. The ice making machine intelligent control method based on AI analysis according to claim 1 is characterized in that: In step S2, the two-stage compression refrigeration system comprises: There are two sets of compressors, namely high-pressure compressor and low-pressure compressor. The exhaust volume of high-pressure compressor is 10m 3 / h, the exhaust volume of the low-pressure compressor is 8m 3 / h; There are two sets of condensers, namely high pressure condenser and low pressure condenser, with condensation areas of 5m 2 and 4m 2 ; There are two sets of evaporators, namely high pressure evaporator and low pressure evaporator, with evaporation areas of 3m 2 and 2m 2 ; The refrigerant of the two-stage compression refrigeration system is R410A, and its charging amount is accurately calculated and adjusted according to the design of the system and the actual operation conditions to ensure the efficient operation of the refrigeration system.

5. The ice making machine intelligent control method based on AI analysis according to claim 1 is characterized in that: In step S4, the specific parameters of the fuzzy PID algorithm are set as follows: the initial value of the proportional coefficient is 0.6, which is dynamically adjusted according to the deviation of the ice thickness, and the maximum value does not exceed 1.2; the initial value of the integral coefficient is 0.1, which is used to eliminate the steady-state error of the ice thickness, and is adjusted according to the accumulated amount of deviation, and the maximum value does not exceed 0.3; the initial value of the differential coefficient is 0.05, which is used to predict the changing trend of the ice thickness, and is adjusted according to the rate of change of the deviation, and the maximum value does not exceed 0.

1.

6. An intelligent control system for ice making machine based on AI analysis, characterized in that: The system includes the following modules: Module S1, a multi-source sensor module, includes a temperature sensor group, including a water tank temperature sensor and an ambient temperature sensor; and also includes a water quality monitoring module, integrating a turbidity sensor and a pH sensor; Module S2, data acquisition and preprocessing module, uses a microcontroller for data acquisition with a sampling frequency of 1 kHz; data preprocessing includes Kalman filtering and median filtering; Module S3, AI analysis and decision-making module, a long short-term memory network model deployed on edge computing devices, with an inference delay of no more than 50 milliseconds; Module S4, dynamic priority scheduling algorithm, adjusts the cooling strategy according to the user's preset preferences; Module S5, control execution module, refrigeration system drive circuit, using PWM voltage regulation technology to control the compressor speed; mechanical actuator, including a blade assembly driven by a stepper motor and a drainage path controlled by a solenoid valve; Module S6, human-computer interaction module, touch screen displays real-time data and operation interface; supports voice command recognition, and the response time does not exceed 0.5 seconds.

7. The ice making machine intelligent control system based on AI analysis according to claim 6 is characterized in that: In the module S4, the dynamic priority scheduling algorithm also includes the following parameters: according to the user's preset preferences, the refrigeration strategy is divided into three modes: high efficiency priority, low energy consumption priority and fast refrigeration priority; in the high efficiency priority mode, the energy efficiency ratio of the refrigeration system is not less than 3.5; in the low energy consumption priority mode, the energy consumption of the refrigeration system is reduced by not less than 20%.

8. The ice machine intelligent control system based on AI analysis according to claim 6 is characterized in that: In the module S5, the refrigeration system driving circuit uses PWM voltage regulation technology to control the compressor speed, the frequency of the PWM signal is 20kHz, and the duty cycle adjustment range is 10%-90%; the response time of the drainage path controlled by the solenoid valve does not exceed 50ms to improve the accuracy of the solenoid valve.

9. An ice machine intelligent control storage medium based on AI analysis, characterized in that: The storage medium is used to execute the ice-making machine intelligent control method based on AI analysis according to any one of claims 1 to 5, and the storage medium includes the following program code: Program S1, real-time control thread, has a priority level set to the highest task level of the real-time operating system and a scheduling period of 1ms; it contains a fuzzy PID control instruction set for adjusting the compressor speed and blade cutting frequency; Program S2, data storage thread, uses a circular buffer mechanism with a buffer size of 1MB and a data persistence period of 60 minutes. The storage content includes sensor raw data and control logs; Program S3, remote upgrade interface, supports HTTPS protocol to communicate with OTA server, firmware differential upgrade algorithm; The firmware verification uses the SHA-256 hash algorithm, and automatically rolls back to the previous stable version if the upgrade fails.

10. The ice machine intelligent control storage medium based on AI analysis according to claim 9, characterized in that: In the program S2, the sensor raw data and control log further include a log classification storage strategy, the operation status log is kept for 30 days and is stored in CSV format; the abnormal alarm log is permanently stored and uploaded to the cloud after encryption; the log retrieval function supports time range query and keyword filtering.

Images