A rice mill control method and system based on voltage detection

Abstract

The present application discloses a rice milling machine control method and system based on voltage detection, which relates to the field of grain processing technology. The disclosed rice milling machine control method and system based on voltage detection collect the main motor voltage signal to extract characteristic parameters, and dynamically adjust the rice milling machine operating parameters in combination with a pre-training model, thereby solving the problems of low monitoring accuracy and delayed response of traditional methods, and can provide processing effects, with the advantages of improving rice milling efficiency, reducing broken rice rate and enhancing equipment adaptability.

Description

Technical Field

[0001] The present application relates to the field of grain processing technology, and in particular to a rice milling machine control method and system based on voltage detection. Background Art

[0002] In the grain processing industry, rice mills are crucial processing equipment, and their performance stability and efficiency directly impact the quality and yield of the final product. Traditional rice mill control methods rely heavily on manual experience or simple sensor feedback, making it difficult to achieve precise load control and intelligent regulation. This not only wastes energy and increases processing costs, but also compromises rice milling performance.

[0003] Existing technologies primarily monitor the load status of rice mills using mechanical pressure sensors or current detection devices. These methods suffer from issues such as delayed response and insufficient measurement accuracy. Particularly when processing rice with varying moisture contents, traditional methods are unable to accurately identify changes in material properties, leading to delayed adjustment of milling parameters, which can easily cause an increase in broken rice or uneven milling. Furthermore, when abnormal operating conditions occur, such as material blockage or equipment overload, existing control systems often lack effective early warning and automatic adjustment mechanisms, requiring manual intervention to restore normal operation, severely impacting production efficiency and equipment life.

[0004] The above content is only used to assist in understanding the technical solution of this application and does not constitute an admission that the above content is prior art. Summary of the Invention

[0005] The main purpose of this application is to provide a rice milling machine control method and system based on voltage detection, aiming to improve the processing effect of the rice milling machine.

[0006] To achieve the above objectives, the present application proposes a rice milling machine control method based on voltage detection, wherein the rice milling machine includes a grinding mechanism, the grinding mechanism includes a grinding roller and a main motor connected to the grinding roller, and the method includes:

[0007] collecting terminal voltage signals of the main motor to generate raw voltage data;

[0008] Preprocessing the raw voltage data to obtain corresponding preprocessed data;

[0009] Extracting the voltage mean, voltage fluctuation amplitude, and high-frequency energy ratio from the preprocessed data to generate corresponding voltage characteristic data;

[0010] Inputting the voltage characteristic data into a pre-trained load prediction model to output rice milling load state parameters;

[0011] The feeding speed of the rice milling machine and the gap between the grinding rollers are adjusted according to the rice milling load state parameters.

[0012] In one embodiment, the step of collecting the terminal voltage signal of the main motor to generate raw voltage data includes:

[0013] The voltage sensor collects the three-phase terminal voltage of the main motor at a preset sampling frequency to generate raw voltage waveform data;

[0014] Performing a sliding average noise reduction process on the original voltage waveform data to obtain noise-reduced voltage data;

[0015] The noise-reduced voltage data is subjected to AD conversion to generate a digital voltage sequence as the original voltage data.

[0016] In one embodiment, the step of preprocessing the raw voltage data to obtain corresponding preprocessed data includes:

[0017] The raw voltage data is smoothed using a Butterworth low-pass filter and segmented into fixed-length windows to extract the effective voltage range of each data segment.

[0018] Normalize the segmented data to eliminate the dimensional differences in voltage amplitude caused by power supply fluctuations;

[0019] The sensor sampling delay is compensated by the time series alignment algorithm to generate time-synchronized preprocessed data.

[0020] In one embodiment, the step of extracting the voltage mean, voltage fluctuation amplitude, and high-frequency energy ratio from the pre-processed data to generate corresponding voltage characteristic data includes:

[0021] Calculating the arithmetic mean of the preprocessed data within a preset time window to generate a voltage mean;

[0022] Calculating the difference between the maximum value and the minimum value of the pre-processed data within a preset time window to generate a voltage fluctuation amplitude;

[0023] The pre-processed data is decomposed by wavelet transform to extract high-frequency components, and the energy value of the high-frequency components in the preset time window and the total energy value in the preset time window are calculated to determine the high-frequency energy ratio;

[0024] The voltage characteristic data is formed by combining the voltage mean value, voltage fluctuation amplitude, and high-frequency energy proportion.

[0025] In one embodiment, the rice milling load state parameters include a required level of milling pressure and an estimated value of rice moisture content, and the method further includes:

[0026] Collecting a historical data set, the historical data set including voltage characteristic data, measured milling pressure level and rice moisture content data, to form a training sample set;

[0027] A two-branch neural network model was constructed, including a fully connected network branch that outputs the required milling pressure level when inputting voltage mean data and fluctuation amplitude data, and a convolutional network branch that outputs an estimated rice moisture content when inputting high-frequency energy ratio data.

[0028] Optimizing the parameters of the dual-branch neural network model by a back-propagation algorithm until the errors between the milling pressure requirement level and the estimated rice moisture content output by the dual-branch neural network model and the measured values ​​in the training sample set reach below a preset threshold;

[0029] The optimized dual-branch neural network model is used as the load forecasting model.

[0030] In one embodiment, the rice milling machine further includes a feeding motor; the step of adjusting the feeding speed of the rice milling machine and the gap between the grinding rollers according to the rice milling load state parameter comprises:

[0031] According to the milling pressure requirement level of the rice milling load state parameter, the preset pressure-speed mapping table is queried to generate the corresponding feed motor speed control signal to adjust the feed speed;

[0032] According to the estimated value of rice moisture content in the rice milling load state parameter, the gap compensation amount of the grinding roller is determined, and a corresponding grinding roller displacement control signal is generated to adjust the gap of the grinding roller.

[0033] In one embodiment, the method further comprises:

[0034] Determine the maximum allowable load current, and apply milling pressure at equal intervals from no-load to full-load on the rice mill, and record the steady-state load current of the feed motor at the corresponding milling pressure requirement level;

[0035] Calculate the speed proportional coefficient corresponding to each grinding pressure requirement level according to the following formula:

[0036] ;

[0037] in, Indicates the rice milling pressure requirement level; Indicates Speed ​​proportional coefficient under the grinding pressure requirement level; Indicates the maximum allowable load current; Indicates The steady-state load current of the motor under the grinding pressure requirement level;

[0038] The speed of the feed motor is determined according to the speed proportional coefficient corresponding to the grinding pressure requirement level, which specifically includes:

[0039] Determine the speed of the feed motor according to the following formula:

[0040] ;

[0041] in, Indicates the speed of the feed motor; Indicates the speed proportional coefficient; Indicates the rated speed of the feed motor;

[0042] A corresponding relationship between the grinding pressure requirement level and the feed motor speed is established to construct the preset pressure-speed mapping table.

[0043] In one embodiment, the method further comprises:

[0044] When the measured load current is greater than the steady-state load current at a certain grinding pressure requirement level, the speed is corrected according to the following formula:

[0045] ;

[0046] in, Indicates the corrected speed; Indicates the speed of the feed motor; Indicates the measured load current, Indicates the maximum allowable load current; Indicates The steady-state load current of the motor under the grinding pressure requirement level.

[0047] In one embodiment, the step of determining the gap compensation amount of the grinding roller according to the estimated value of the rice moisture content in the rice milling load state parameter includes:

[0048] When the estimated value of the rice moisture content is greater than a preset first threshold value and less than a preset second threshold value, determining the gap compensation amount of the grinding roller to be 0;

[0049] When the estimated value of the rice moisture content is less than a preset first threshold, determining the gap compensation amount of the grinding roller to be 0.1 mm;

[0050] When the estimated value of the rice moisture content is greater than a preset second threshold value and less than a preset third threshold value, determining the gap compensation amount of the grinding roller to be 0.2 mm;

[0051] When the estimated rice moisture content is greater than a preset third threshold, the gap compensation amount of the grinding roller is determined to be 0.3 mm; wherein the preset first threshold is less than the preset second threshold, and the preset second threshold is less than the preset third threshold.

[0052] In one embodiment, the method further comprises:

[0053] Calculating a voltage change rate based on the raw voltage data;

[0054] Collect the current data of the main motor and calculate the current change rate;

[0055] When the current change rate is greater than a preset current change rate and the voltage change rate is less than a preset voltage change rate, the feed motor speed is controlled to decrease by a preset speed, and the roller gap is controlled to increase by a preset distance;

[0056] If, after a preset time period, the current change rate remains greater than the preset current change rate and the voltage change rate remains less than the preset voltage change rate, the grinding roller is controlled to reverse.

[0057] In addition, to achieve the above-mentioned purpose, the present application also proposes a rice milling machine control system based on voltage detection, the system comprising: a memory, a processor, and a rice milling machine control program based on voltage detection stored in the memory and executable on the processor, the rice milling machine control program based on voltage detection being configured to implement the steps of the rice milling machine control method based on voltage detection.

[0058] The present application provides a rice mill control method and system based on voltage detection. By collecting the voltage signal of the main motor to extract characteristic parameters, the rice mill operating parameters are dynamically adjusted in combination with a pre-trained model. This solves the problems of low monitoring accuracy and delayed response of traditional methods, can provide processing effects, and has the advantages of improving rice milling efficiency, reducing broken rice rate, and enhancing the adaptability of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0060] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0061] Figure 1 A flow chart of an embodiment of a rice milling machine control method based on voltage detection provided in this application;

[0062] Figure 2 For this application Figure 1 Detailed flow chart of step S100;

[0063] Figure 3 For this application Figure 1 Detailed flow diagram of step S200;

[0064] Figure 4 For this application Figure 1 Detailed flowchart of step S300;

[0065] Figure 5 A flow chart of another embodiment of the rice milling machine control method based on voltage detection provided in this application;

[0066] Figure 6 For this application Figure 1 Detailed flowchart of step S500;

[0067] Figure 7 For this application Figure 6 Detailed flowchart of step S520;

[0068] Figure 8 A flow chart of another embodiment of the rice milling machine control method based on voltage detection provided by the present application;

[0069] Figure 9 This is a structural diagram of an embodiment of a rice mill control system based on voltage detection provided in this application.

[0070] Description of Figure Numbers:

[0071] 10. Memory; 20. Processor.

[0072] The purpose, features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0073] The technical solutions in this application will be clearly and completely described below in conjunction with the drawings in this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all of the embodiments. The components of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application for which protection is claimed, but merely represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of this application.

[0074] It should be understood that similar reference numerals and letters in the following drawings represent similar items. Therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings. At the same time, in the description of this application, the terms "first", "second", etc. are only used to distinguish the description and should not be understood as indicating or implying relative importance.

[0075] Conventional rice milling machine control relies primarily on operator experience or single sensor feedback, resulting in insufficient load state sensing accuracy. Manual adjustments can lag when material properties change or equipment wear occurs, leading to an imbalance between grinding pressure and material supply, potentially causing equipment overload or insufficient grinding. In one grain processing workshop, a sudden change in rice moisture content led to a failure to adjust the roller gap in a timely manner. This resulted in the rice mill experiencing excessive broken rice and motor shutdowns due to overheating, resulting in significant raw material waste.

[0076] To address these issues, researchers discovered that the voltage signal at the main motor terminal contains implicit information about the load status. Analysis revealed a correlation between the mean voltage and grinding pressure, that the amplitude of voltage fluctuation reflects material flow stability, and that energy variations in the high-frequency component indicate differences in rice grain hardness. This led to a technical strategy: constructing a multidimensional feature system for voltage signals and using machine learning models to establish a mapping between voltage characteristics and load status, thereby achieving dynamic closed-loop control.

[0077] Therefore, the present application proposes a rice milling machine control method based on voltage detection, the rice milling machine includes a grinding mechanism, the grinding mechanism includes a grinding roller and a main motor connected to the grinding roller. Figure 1 , the method includes steps S100 to S500, wherein:

[0078] Step S100, collecting the terminal voltage signal of the main motor to generate raw voltage data;

[0079] Step S200, preprocessing the original voltage data to obtain corresponding preprocessed data;

[0080] Step S300, extracting the voltage mean, voltage fluctuation amplitude, and high-frequency energy ratio from the pre-processed data to generate corresponding voltage characteristic data;

[0081] Step S400: inputting the voltage characteristic data into a pre-trained load prediction model to output rice milling load state parameters;

[0082] Step S500: adjusting the feeding speed of the rice milling machine and the gap between the grinding rollers according to the rice milling load state parameters.

[0083] In this embodiment, the voltage mean refers to the average level of the voltage signal within a preset time window, which can be calculated by a sliding average algorithm and is used to reflect the overall load level of the milling pressure. The voltage fluctuation amplitude refers to the peak-to-valley difference of the voltage signal within the time window, which can be obtained by range calculation and is used to characterize the stability of the material flow during the milling process. The high-frequency energy proportion refers to the energy proportion of the high-frequency component of the voltage signal after wavelet decomposition, which can be achieved by fast wavelet transform and is used to identify the transient impact characteristics caused by changes in the moisture content of rice. The load prediction model refers to a two-branch neural network trained with historical data, in which the convolutional network branch processes the high-frequency energy characteristics, and the fully connected network processes the mean and fluctuation characteristics to achieve a joint prediction of the milling pressure and moisture content.

[0084] Specifically, during the operation of the rice mill, the voltage sensor continuously collects the voltage signal at the main motor end at a fixed sampling frequency. The original signal is low-pass filtered to eliminate high-frequency interference, and then normalized to eliminate the influence of grid voltage fluctuations. The pre-processed data is divided into time windows, and the voltage mean, fluctuation amplitude, and high-frequency energy ratio within the window are calculated synchronously. These three types of features are fused and analyzed through a neural network model to output the required level of milling pressure and the estimated moisture content of the rice. The control system calls the preset speed mapping table according to the pressure level to control the feed speed, and at the same time calculates the roller gap compensation based on the estimated moisture content, and adjusts the milling pressure in real time through the servo mechanism.

[0085] Compared with existing technologies, traditional methods rely on a single current sensor or vibration sensor to determine load status, and are unable to distinguish the effects of changes in grinding pressure from those of changes in material properties. This solution, however, uses multidimensional voltage feature extraction to simultaneously capture information about both load intensity and material property changes, and combines it with a dual-branch neural network to achieve precise state identification. While existing threshold-based alarm-based control methods suffer from response lag, this solution implements feedforward control through a predictive model, significantly improving the timeliness of regulation.

[0086] Through the above technical solution, this application effectively solves the problem of low load identification accuracy in traditional control methods. The load prediction model based on multi-dimensional characteristic analysis of voltage signals can accurately distinguish the combined effects of changes in grinding pressure and changes in rice moisture content, and achieve coordinated adjustment of feed speed and roller gap. This solution avoids the uncertainty of manual adjustment through experience. Through a data-driven closed-loop control mechanism, it ensures that the rice mill maintains a stable operating state under different material characteristics, improves processing results, and reduces energy waste and abnormal equipment loss.

[0087] In one possible implementation, reference Figure 2 , the step S100 includes steps S110 to S130, wherein:

[0088] Step S110, collecting the three-phase terminal voltage of the main motor at a preset sampling frequency through a voltage sensor to generate original voltage waveform data;

[0089] Step S120, performing sliding average noise reduction processing on the original voltage waveform data to obtain noise-reduced voltage data;

[0090] Step S130 , performing AD conversion on the noise-reduced voltage data to generate a digital voltage sequence as the original voltage data.

[0091] In this embodiment, the preset sampling frequency refers to the time interval parameter for data collection, which can be implemented by a sampling frequency of 1000 to 2000 times per second. This frequency range can cover the complete voltage fluctuation spectrum when the motor is operating. Sliding average noise reduction processing refers to an algorithm that dynamically calculates the local average value of time series data using a moving window containing 50 to 200 sampling points, thereby improving data quality by suppressing high-frequency noise components. AD conversion refers to the quantization process of converting an analog voltage signal into a digital signal, which can be implemented using a 12-bit or 16-bit precision analog-to-digital conversion chip to eliminate distortion errors during analog signal transmission.

[0092] Specifically, the synchronous acquisition of the three-phase terminal voltages fully captures the interphase voltage differences during motor operation, avoiding feature extraction errors caused by missing data from a single phase. The sliding average algorithm dynamically adjusts the window length to preserve the valid signal trend while eliminating random interference, such as brush contact noise or high-frequency glitches caused by electromagnetic interference. The noise-reduced analog signal is converted into a digital sequence using a high-precision analog-to-digital converter. This process not only enables digital storage of the signal but also ensures accuracy consistency in subsequent data processing stages through quantization resolution control.

[0093] Compared to existing technologies, traditional methods typically use single-phase voltage acquisition without dynamic noise reduction, resulting in phase deviation and residual noise in the data. While existing technologies use fixed-window mean filtering to easily distort the effective signal, this solution utilizes a sliding average algorithm that adaptively adjusts the window parameters based on noise characteristics. Furthermore, existing analog-to-digital conversion often overlooks the matching relationship between signal preprocessing and quantization accuracy. This solution, however, utilizes pre-processing noise reduction to fully utilize the effective resolution of the A / D converter.

[0094] Through the above technical solution, this application solves the problem of waveform distortion caused by noise interference during voltage signal acquisition, eliminates the amplitude dimension differences caused by power supply fluctuations, and improves the quantization accuracy of the signal conversion process. The complete acquisition of three-phase voltage provides the data foundation for subsequent feature extraction. The sliding average algorithm effectively suppresses the contamination of the effective signal by high-frequency interference. The high-precision AD conversion ensures the true restoration of the digital sequence, thus establishing a reliable data source for accurately determining the load status of the rice mill.

[0095] In one possible implementation, reference Figure 3 , the step S200 includes steps S210 to S230, wherein:

[0096] Step S210: smoothing the original voltage data using a Butterworth low-pass filter, segmenting the data into fixed-length windows, and extracting the effective voltage range of each data segment;

[0097] Step S220 , normalizing the segmented data to eliminate the dimensional difference of the voltage amplitude caused by power supply fluctuations;

[0098] Step S230 : Compensating for sensor sampling delays through a time series alignment algorithm to generate time-synchronized pre-processed data.

[0099] In this embodiment, the Butterworth low-pass filter refers to a filter having the maximum flat amplitude characteristic within the passband, which can be specifically implemented by an eighth-order filter with a cutoff frequency of 50 Hz, and is used to filter out high-frequency interference noise and retain effective low-frequency voltage signals. Fixed-time window segmentation refers to dividing continuous voltage data into independent analysis units according to a fixed time length, which can be specifically implemented by a non-overlapping window of 200 milliseconds, and is used to distinguish the data distribution characteristics of different milling stages. Normalization processing refers to linearly mapping the voltage amplitude to a uniform numerical interval, which can be specifically implemented by a maximum and minimum value normalization method, and is used to eliminate amplitude differences caused by grid voltage fluctuations. The time series alignment algorithm refers to the synchronous correction of the timestamps of multi-source data, and can specifically use the cubic spline interpolation method to compensate for sensor delay data, and is used to eliminate data phase offsets caused by sampling delays.

[0100] Specifically, the preprocessing process first uses a Butterworth low-pass filter to filter the noisy voltage data in the frequency domain to suppress contamination of the valid signal by high-frequency electromagnetic interference. The filtered data is then divided into multiple analysis units according to fixed-length windows. The effective operating state is identified by calculating the voltage extreme value difference within each window. Each data segment is normalized to eliminate absolute value differences caused by grid voltage fluctuations. Finally, an interpolation algorithm is used to compensate for sensor sampling delays, ensuring that the preprocessed data remains synchronized in the time dimension. These three processing steps successively eliminate signal noise, dimensional differences, and timing deviations, forming a high-precision preprocessed data sequence.

[0101] Compared with existing technologies, traditional methods typically only perform a single filtering process on the voltage signal, without considering the dimensional differences and sensor sampling delays caused by power supply fluctuations. The use of moving average filtering in existing technologies is prone to signal phase distortion, and the lack of a data segmentation mechanism limits feature extraction accuracy. This application achieves multi-dimensional data quality optimization by combining low-pass filtering, normalization processing, and time synchronization algorithms, effectively solving the signal distortion problem under complex working conditions.

[0102] Through the above technical solution, this application can eliminate high-frequency noise interference in voltage signals, suppress the impact of power supply system fluctuations on data stability, and compensate for timing errors caused by sensor sampling delays, thereby obtaining high-precision preprocessed data. This provides a reliable data foundation for the subsequent accurate extraction of voltage characteristic parameters, ultimately improving the accuracy of rice milling load state prediction.

[0103] In one possible implementation, reference Figure 4 , the step S300 includes steps S310 to S340, wherein:

[0104] Step S310, calculating the arithmetic mean of the pre-processed data within a preset time window to generate a voltage mean;

[0105] Step S320, calculating the difference between the maximum value and the minimum value of the pre-processed data within a preset time window to generate a voltage fluctuation amplitude;

[0106] Step S330 , performing wavelet transform decomposition on the pre-processed data to extract high-frequency components, calculating the energy value of the high-frequency components within a preset time window and the total energy value within the preset time window to determine the high-frequency energy ratio;

[0107] Step S340 : Combining the voltage mean value, voltage fluctuation amplitude, and high-frequency energy ratio to form the voltage characteristic data.

[0108] In this embodiment, the voltage mean refers to the arithmetic mean of the voltage data within a preset time window, which can be implemented by a moving average algorithm or a weighted average algorithm, and is used to reflect the average load level when the motor is running. The voltage fluctuation amplitude refers to the difference between the maximum and minimum values ​​of the voltage data, which can be implemented by sliding window range calculation, and is used to characterize the dynamic disturbance characteristics of the voltage caused by load changes. The high-frequency energy proportion refers to the proportion of the energy of the high-frequency component to the total energy after wavelet transform decomposition, which can be implemented by multi-scale decomposition using the Daubechies wavelet basis, and is used to capture transient abnormal fluctuations in the voltage waveform related to the moisture content of rice.

[0109] Specifically, the calculation of the voltage mean eliminates random interference signals to establish a quantitative indicator of the steady-state load level. The calculation of the voltage fluctuation range effectively reflects the dynamic impact of material density changes on the motor during the rice milling process. The extraction of the high-frequency energy fraction uses time-frequency analysis to identify voltage waveform distortion caused by differences in rice moisture content. These three methods construct composite feature vectors from the steady-state, dynamic, and transient dimensions, respectively, enabling subsequent models to simultaneously perceive both macro-trends and micro-anomalies in rice mill load changes.

[0110] Compared with existing technologies, traditional methods use only a single voltage mean or fluctuation amplitude as a feature, which cannot distinguish different voltage variation patterns caused by milling pressure and rice moisture content. This solution introduces the time-frequency domain feature of high-frequency energy proportion and combines it with multi-dimensional feature fusion to effectively distinguish the type of load variation and solve the problem of prediction bias caused by a single feature dimension.

[0111] Through the above technical solution, this application realizes the multi-dimensional feature extraction of voltage signals. In the load state prediction of rice mills, it can simultaneously capture the steady-state load level, dynamic disturbance intensity and transient abnormal fluctuations, and provide subsequent models with more discriminative input data, thereby improving the prediction accuracy of the grinding pressure demand level and the estimated value of rice moisture content.

[0112] In a feasible implementation manner, the rice milling load state parameters include the milling pressure requirement level and the estimated value of the rice moisture content. Figure 5 The method further includes steps S410 to S440, wherein:

[0113] Step S410, collecting a historical data set, wherein the historical data set includes voltage characteristic data, measured milling pressure level and rice moisture content data to form a training sample set;

[0114] Step S420, constructing a two-branch neural network model, including a first branch of a fully connected network, for outputting a milling pressure requirement level when inputting voltage mean data and fluctuation amplitude data, and a second branch of a convolutional network, for outputting an estimated value of rice moisture content when inputting high-frequency energy proportion data;

[0115] Step S430, optimizing the parameters of the dual-branch neural network model by a back propagation algorithm until the errors between the milling pressure requirement level and the estimated rice moisture content output by the dual-branch neural network model and the measured values ​​in the training sample set reach below a preset threshold;

[0116] Step S440: Using the optimized dual-branch neural network model as the load forecasting model.

[0117] In this embodiment, the historical data set refers to a collection of associated data including voltage characteristics and measured parameters, which can be implemented specifically by database storage and used to establish a mapping relationship between voltage characteristics and milling load status. The dual-branch neural network model refers to a composite model architecture including independent feature processing paths, which can be implemented specifically by a combination of a fully connected layer and a convolutional layer structure. The fully connected network captures the steady-state characteristics of the voltage mean, and the convolutional network extracts the dynamic pattern of high-frequency energy. The backpropagation algorithm refers to a method of adjusting network weights through gradient calculation, which can be implemented specifically by a chain derivation rule so that the error between the model output and the measured value gradually converges.

[0118] Specifically, the voltage mean data reflects the average load level of the main motor during operation, and is directly mapped to the grinding pressure requirement level through the fully connected network. The voltage fluctuation amplitude data characterizes the severity of the load change and is used by the fully connected network to correct the tolerance range for pressure level determination. The high-frequency energy proportion data reflects the high-frequency oscillation characteristics of the voltage caused by material friction during the grinding process. The convolutional network extracts its correlation pattern with the moisture content of the rice through local perception. The dual-branch structure allows the pressure level and moisture content prediction tasks to share the underlying voltage features but maintain independent parameter updates, avoiding feature interference caused by multi-task learning of a single network. The preset error threshold is set as the termination condition for model training. For example, the iteration can be stopped when the mean square error is lower than 0.05 to ensure that the prediction accuracy meets the engineering control requirements.

[0119] Compared to existing technologies, traditional methods typically use a single-output model to predict a single parameter, such as grinding pressure or moisture content, which cannot simultaneously reflect mechanical load and material properties. Conventional neural networks, when processing complex features, can easily overwhelm high-frequency signal features with steady-state data. However, this solution utilizes a dual-branch structure to process different frequency domain features separately, preserving the ability to identify high-frequency dynamic information. Existing technologies often rely on humidity sensors for moisture content prediction. This solution indirectly infers moisture content through high-frequency voltage characteristics, reducing the need for dedicated sensor deployment.

[0120] Through the above technical solution, this application achieves the simultaneous and accurate prediction of milling pressure demand and rice moisture content, solving the load regulation lag problem caused by single-parameter prediction in traditional methods. The dual-path processing mechanism of voltage characteristics effectively distinguishes steady-state load characteristics from dynamic material characteristics, avoiding feature confusion of a single model. The joint optimization of model parameters produces a synergistic effect between pressure level and moisture content prediction. For example, high-frequency voltage fluctuations caused by changes in moisture content can synchronously correct the pressure level determination result, improving the overall accuracy of load state identification.

[0121] In a feasible embodiment, the rice milling machine further includes a feeding motor. Figure 6 , the step S500 includes steps S510 to S520, wherein:

[0122] Step S510: according to the milling pressure requirement level of the rice milling load state parameter, query the preset pressure-speed mapping table to generate a corresponding feed motor speed control signal to adjust the feed speed;

[0123] Step S520: determining the gap compensation amount of the grinding roller according to the estimated value of the rice moisture content in the rice milling load state parameter, and generating a corresponding grinding roller displacement control signal to adjust the gap of the grinding roller.

[0124] In this embodiment, the pressure-speed mapping table refers to a pre-established reference table of the correspondence between the grinding pressure level and the feed motor speed. Specifically, it can be constructed by recording the correspondence between the steady-state load current and the speed of the motor at different pressure levels, and is used to achieve standardized matching between the load state and the actuator action. The estimated value of the moisture content of the rice refers to the predicted value of the moisture content of the material obtained by analyzing the voltage characteristic data. Specifically, it can be estimated using a neural network model obtained by training the high-frequency energy ratio and the sample moisture content data, and is used to guide the dynamic adjustment of the grinding gap. The gap compensation amount refers to the amount of adjustment of the grinding roller position set according to the difference in the moisture content of the material. Specifically, it can be achieved by setting the moisture content threshold interval to correspond to different compensation amounts. For example, the preset first threshold is set to 12%, the second threshold is set to 15%, and the third threshold is set to 18%.

[0125] Specifically, the feed speed adjustment process is automatically matched through a pressure-speed mapping table. When the required level of grinding pressure is detected, the system directly calls the corresponding speed parameter in the mapping table to generate a control signal, avoiding the response delay of manual adjustment based on experience. The adjustment of the gap between the grinding rollers is performed by comparing the moisture content estimate with the preset threshold range. When a low-moisture material is detected, the gap compensation amount is automatically reduced to prevent insufficient grinding. When a high-moisture material is detected, the gap compensation amount is increased to avoid an increase in the broken rice rate. This dual-parameter collaborative control method enables the feed speed and grinding intensity to form a dynamic balance based on the real-time load status.

[0126] Compared to existing technologies, traditional methods rely on operators visually observing the material state to make mechanical adjustments, which can lead to adjustment lag and insufficient precision. Existing technologies adjust feed speed solely through a single motor current parameter, failing to address fluctuations in milling quality caused by moisture content variations. This solution, however, achieves complex closed-loop control of rice mill operating parameters by establishing a pressure-speed mapping relationship and a moisture content gap compensation mechanism.

[0127] Through the above technical solution, this application achieves automatic coordinated control of the rice mill's feed speed and grinding gap, effectively solving the problems of low adjustment accuracy and slow response speed caused by traditional equipment's reliance on manual experience. The precise matching of grinding pressure and feed speed avoids the risk of motor overload, and the gap compensation mechanism based on moisture content estimation reduces the rate of broken rice during the grinding of high-moisture materials, thereby improving the quality of the finished product while ensuring grinding efficiency.

[0128] In a feasible embodiment, the method further includes determining a maximum allowable load current, applying milling pressure at equally spaced pressure levels within a range from no-load to full-load of the rice mill, and recording the steady-state load current of the feed motor at each pressure level;

[0129] Calculate the speed proportional coefficient corresponding to each grinding pressure requirement level according to the following formula:

[0130] ;

[0131] in, Indicates the rice milling pressure requirement level; Indicates Speed ​​proportional coefficient under the grinding pressure requirement level; Indicates the maximum allowable load current; Indicates The steady-state load current of the motor under the grinding pressure requirement level;

[0132] The speed of the feed motor is determined according to the speed proportional coefficient corresponding to the grinding pressure requirement level, which specifically includes:

[0133] Determine the speed of the feed motor according to the following formula:

[0134] ;

[0135] in, Indicates the speed of the feed motor; Indicates the speed proportional coefficient; Indicates the rated speed of the feed motor;

[0136] A corresponding relationship between the grinding pressure requirement level and the feed motor speed is established to construct the preset pressure-speed mapping table.

[0137] In this embodiment, the maximum allowable load current refers to the maximum current threshold of the rice mill's drive motor under safe operating conditions. This can be determined based on the motor's rated power and the overload protection device's set value, and is used to define the safe operating margins of the system. Equally spaced pressure levels divide the rice mill's operating pressure range into multiple evenly distributed test points. This can be achieved by using no-load pressure as the starting point and full-load pressure as the end point, with fixed increments, to establish benchmark data for different load conditions. The steady-state load current refers to the current value when the feed motor reaches a stable operating state under constant milling pressure. This can be obtained by continuously collecting current data and calculating its average value, and is used to reflect the load characteristics under a specific pressure level. The speed proportionality factor refers to the ratio of the actual speed to the rated speed. This can be calculated using a formula based on the difference between the steady-state load current and the maximum allowable current, and is used to achieve nonlinear speed regulation with load changes.

[0138] Specifically, by pre-calibrating the steady-state current at different pressure levels, a corresponding relationship between the load current and the grinding pressure is established. When the current pressure level is detected, the speed proportional coefficient is dynamically calculated based on the ratio of the steady-state current value recorded at that level to the maximum allowable current. For example, when the steady-state current of a certain pressure level reaches 80% of the maximum allowable current, the proportional coefficient will be adjusted to 0.8, and the actual speed is 80% of the rated speed. This nonlinear adjustment method based on the load current ratio can automatically reduce the feed speed to prevent overload at high loads, and maintain a higher speed at low loads to improve processing efficiency. The final pressure-speed mapping table converts the abstract pressure level into an executable speed instruction, realizing closed-loop control of the feed speed during the processing process.

[0139] Compared to existing technologies, traditional methods often use fixed speeds or simple linear speed regulation strategies, which are unable to adapt to load fluctuations caused by changes in rice moisture content and hardness during the rice milling process. This method, however, establishes a nonlinear mapping relationship between pressure level and speed, automatically matching the optimal feed speed based on real-time load conditions. For example, when encountering high-hardness rice, the system automatically reduces the feed speed by detecting an increase in load current. Existing fixed speed solutions increase the risk of motor overload.

[0140] Through the above technical solution, this application achieves dynamic adaptation of the rice mill's feed rate and grinding pressure, effectively preventing motor overload caused by sudden load changes. Furthermore, a nonlinear speed regulation mechanism based on a proportional coefficient optimizes energy efficiency while ensuring process safety. A preset pressure-speed mapping table simplifies the control system's parameter calculation process and improves real-time response speed.

[0141] In a feasible embodiment, the method further includes correcting the rotation speed according to the following formula when the measured load current is greater than the steady-state load current at a certain grinding pressure requirement level:

[0142] ;

[0143] in, Indicates the corrected speed; Indicates the speed of the feed motor; Indicates the measured load current, Indicates the maximum allowable load current; Indicates The steady-state load current of the motor under the grinding pressure requirement level.

[0144] In this embodiment, the measured load current refers to the feed motor operating current, collected in real time by a current sensor during rice milling operation. Specifically, this can be achieved using a Hall effect current sensor and is used to characterize the dynamic changes in the current milling load. The steady-state load current refers to the average current value of the feed motor in a stable operating state at a specific milling pressure requirement level. This can be obtained through pressure level calibration experiments and used to establish benchmark current parameters under different operating conditions. The maximum allowable load current refers to the maximum current value that the rice mill's main motor can withstand within a safe operating range. This value is set based on the motor's rated parameters and is used to prevent equipment overload damage. The square root operation in the speed correction coefficient is used to balance the nonlinear relationship between current deviation and speed adjustment amplitude. Its convergence is verified through mathematical modeling to ensure that the corrected speed can both quickly respond to abnormal currents and avoid material flow disruptions caused by sudden speed changes.

[0145] Specifically, when it is detected that the measured load current exceeds the steady-state current value corresponding to the current grinding pressure requirement level, the speed correction coefficient is generated by calculating the ratio of the difference between the maximum allowable load current and the measured current to the difference between the maximum allowable load current and the steady-state current. This coefficient is multiplied by the current speed to obtain the corrected speed, so that the feed speed is dynamically reduced as the abnormality of the load current increases. For example, when the measured current is close to the maximum allowable current, the correction coefficient approaches zero, and the feed speed drops significantly to relieve the motor load pressure. This mechanism realizes adaptive speed adjustment through dynamic current feedback, which not only takes into account the differences in the normal load range under the current grinding pressure level, but also avoids misadjustment caused by instantaneous current fluctuations.

[0146] Compared with existing technologies, traditional methods usually use fixed thresholds to trigger speed regulation or adjust the speed through linear proportions, which cannot adapt to the dynamic changes in load current characteristics under different grinding pressure levels. This solution introduces steady-state current parameters based on pressure level calibration and optimizes the speed correction ratio in combination with a square root function. While reducing the instantaneous load of the motor, it maintains the processing continuity of the rice mill under complex working conditions. For example, in existing technologies, when the current exceeds the limit, the machine is directly shut down or the speed is reduced in steps, which can easily lead to material blockage or a sudden drop in processing efficiency. However, this solution maintains the stability of the processing process while protecting equipment safety through progressive speed correction.

[0147] Through the above technical solution, the present application realizes dynamic speed compensation of the rice mill when the load current increases abnormally, effectively avoiding the risk of motor overload caused by instantaneous accumulation of materials or sudden changes in grinding resistance. By combining the difference ratio between the steady-state current parameter and the maximum allowable current at the current pressure level, the speed adjustment range is accurately controlled, which not only prevents the decline in processing efficiency caused by excessive speed reduction, but also ensures a rapid response to abnormal current conditions. For example, when the moisture content of rice fluctuates and causes a sudden increase in grinding resistance, this solution can automatically balance the grinding load and motor output power by correcting the feed speed in real time, avoiding the impact of equipment protective shutdown on continuous production.

[0148] In one possible implementation, reference Figure 7 , the step S520 includes steps S521 to S524, wherein:

[0149] Step S521: when the estimated value of the rice moisture content is greater than a preset first threshold value and less than a preset second threshold value, determining the gap compensation amount of the grinding roller to be 0;

[0150] Step S522: when the estimated value of the rice moisture content is less than a preset first threshold, determining the gap compensation amount of the grinding roller to be 0.1 mm;

[0151] Step S523: when the estimated rice moisture content is greater than the preset second threshold value and less than the preset third threshold value, determining the gap compensation amount of the grinding roller to be 0.2 mm;

[0152] Step S524: When the estimated rice moisture content is greater than a preset third threshold, the gap compensation amount of the grinding roller is determined to be 0.3 mm; wherein the preset first threshold is less than the preset second threshold, and the preset second threshold is less than the preset third threshold.

[0153] In this embodiment, the preset threshold refers to the critical value of rice moisture content determined experimentally. Specifically, it can be set by measuring the milling performance of samples with different moisture contents using laboratory testing equipment. The gap compensation refers to the mechanical displacement of the grinding roller relative to the reference position and can be achieved through a servo motor-driven adjustment mechanism. Multi-level threshold division refers to mapping continuously changing moisture content data into discrete control intervals. This can be achieved using a conditional judgment logic module.

[0154] Specifically, when the moisture content is detected within the intermediate threshold range, indicating that the rice moisture is within the normal processing range, maintaining the baseline gap ensures grinding efficiency. When the moisture content is detected below the minimum threshold, the system automatically increases the gap compensation to reduce the breakage rate of the dried rice. When the moisture content exceeds the intermediate threshold but does not reach the maximum threshold, appropriately increasing the gap can prevent a sudden increase in pressure caused by material adhesion. When the moisture content exceeds the maximum threshold, further increasing the gap can effectively prevent clogging of the grinding chamber. This multi-stage adjustment mechanism maintains stable grinding pressure by matching moisture content changes with the mechanical gap in real time.

[0155] In one specific embodiment, the first threshold can be set at 12%, the second threshold at 15%, and the third threshold at 18%. The gap compensation actuator can utilize a stepper motor coupled with a ball screw structure. The number of motor drive pulses corresponding to each 0.1 mm of compensation is determined through calibration testing. The threshold parameters can be dynamically adjusted through a human-machine interface to accommodate the processing requirements of different rice varieties.

[0156] Compared to existing technologies, traditional methods often rely on fixed gaps or single threshold adjustments, which can easily lead to imbalanced grinding pressure when responding to large fluctuations in moisture content. Gap adjustment devices disclosed in existing literature only perform feedback control based on the motor load current, failing to predict the impact of changes in material properties on the grinding process. This solution establishes a multi-level correspondence between moisture content and gap size, combining feedforward and feedback control, significantly improving system response speed and control accuracy.

[0157] Through this technical solution, the present application can dynamically adjust the grinding gap based on the moisture content of the rice, effectively avoiding an increase in broken rice due to over-drying of the material, preventing equipment overload failures caused by high-moisture materials, and reducing the frequency of manual adjustments. The multi-level threshold mechanism balances control accuracy and system stability, improving the equipment's continuous operation capacity while ensuring rice milling quality.

[0158] In one possible implementation, reference Figure 8 The method further includes steps S610 to S640, wherein:

[0159] Step S610, calculating the voltage change rate based on the original voltage data;

[0160] Step S620, collecting the current data of the main motor and calculating the current change rate;

[0161] Step S630, when the current change rate is greater than a preset current change rate and the voltage change rate is less than a preset voltage change rate, controlling the feed motor speed to decrease by a preset speed and controlling the roller gap to increase by a preset distance;

[0162] Step S640: If the current change rate remains greater than the preset current change rate and the voltage change rate remains less than the preset voltage change rate after the preset time period, the grinding roller is controlled to reverse.

[0163] In this embodiment, the voltage change rate refers to the amount of change in voltage per unit time, which can be achieved by performing differential operations on the original voltage data or sliding window differential calculations, and is used to reflect the dynamic fluctuations in the power supply status of the main motor. The current change rate refers to the amount of change in current per unit time, which can be achieved by performing first-order derivative calculations after collecting the current signal through a Hall sensor, and is used to characterize sudden load changes or changes in mechanical resistance. The preset speed can be set to 10%-30% of the current speed of the feed motor, which can be achieved by adjusting the output frequency of the inverter to quickly reduce the material supply. The preset distance can be set to a range of 0.05-0.15 mm, which is specifically implemented by a servo motor driving the roller displacement mechanism to relieve instantaneous mechanical pressure. The preset duration can be set to 3-5 seconds, which is specifically implemented by a timer module to verify the effectiveness of intervention measures.

[0164] Specifically, when the rice milling machine experiences a sudden load change due to abnormal moisture content in the material or the mixing of foreign matter, the current change rate of the main motor will quickly exceed the preset threshold, while the voltage change rate fails to rise synchronously due to the grid compensation mechanism. At this time, the first stage of protection is formed by reducing the amount of material accumulation by reducing the feed speed and increasing the roller gap to reduce the grinding resistance. If the abnormal state is not eliminated within the set time, it indicates that there is a risk of persistent blockage. At this time, the roller is controlled to rotate in the opposite direction by a specific angle, such as 30-60 degrees, to remove the stuck material from the grinding area. The entire process uses a multi-parameter collaborative judgment mechanism to implement a control strategy from early warning to active intervention in stages, while maintaining the continuous operation of the equipment.

[0165] Compared to existing technologies, traditional methods typically only monitor the absolute value of current or use fixed-time delays for shutdowns, failing to distinguish between transient fluctuations and persistent anomalies. Single-parameter threshold control in existing technologies can easily lead to malfunctions, such as triggering unnecessary shutdowns when grid voltage fluctuates. This solution, however, accurately identifies mechanical jamming characteristics through dynamic correlation analysis of the rate of change of voltage and current. It also employs a progressive control approach, avoiding efficiency losses caused by frequent shutdowns while proactively eliminating the source of the jam through reversal operations.

[0166] Through the above technical solution, this application can promptly eliminate abnormal loads without interrupting the processing flow, preventing excessive accumulation of material in the grinding area that could cause roller jamming. By implementing speed regulation, gap adjustment, and reversal operations in stages, this effectively solves the response lag or excessive intervention caused by a single threshold judgment in traditional control methods, maintaining the continuity of rice milling operations while ensuring equipment safety.

[0167] This application also provides a rice mill control system based on voltage detection, please refer to Figure 9 The system includes: a memory 10, a processor 20, and a rice milling machine control program based on voltage detection stored in the memory 10 and executable on the processor 20. The rice milling machine control program based on voltage detection is configured to implement the steps of the rice milling machine control method based on voltage detection.

[0168] The rice mill control system based on voltage detection provided by this application, which adopts the rice mill control method based on voltage detection in the above-mentioned embodiment, can improve the processing effect of the rice mill. Compared with the existing technology, the beneficial effects of the rice mill control system based on voltage detection provided by this application are the same as those of the rice mill control method based on voltage detection provided by the above-mentioned embodiment. The other technical features of the rice mill control system based on voltage detection are the same as those disclosed in the above-mentioned embodiment and are not further described here.

[0169] The above description is only part of the embodiments of the present application and does not limit the patent scope of the present application. All equivalent structural transformations made by using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect application in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A rice mill control method based on voltage detection, characterized in that: The rice milling machine includes a grinding mechanism, and the grinding mechanism includes a grinding roller and a main motor drivingly connected to the grinding roller. The method includes: collecting terminal voltage signals of the main motor to generate raw voltage data; Preprocessing the raw voltage data to obtain corresponding preprocessed data; Extracting the voltage mean, voltage fluctuation amplitude, and high-frequency energy ratio from the preprocessed data to generate corresponding voltage characteristic data; Inputting the voltage characteristic data into a pre-trained load prediction model to output rice milling load state parameters; adjusting the feeding speed of the rice milling machine and the gap between the grinding rollers according to the rice milling load state parameters; Wherein, the rice milling load state parameters include the required level of milling pressure and the estimated value of rice moisture content, and the method further includes: Collecting a historical data set, the historical data set including voltage characteristic data, measured milling pressure level and rice moisture content data, to form a training sample set; A two-branch neural network model was constructed, including a fully connected network branch that outputs the required milling pressure level when inputting voltage mean data and fluctuation amplitude data, and a convolutional network branch that outputs an estimated rice moisture content when inputting high-frequency energy ratio data. Optimizing the parameters of the dual-branch neural network model by a back-propagation algorithm until the errors between the milling pressure requirement level and the estimated rice moisture content output by the dual-branch neural network model and the measured values ​​in the training sample set reach below a preset threshold; The optimized dual-branch neural network model is used as the load forecasting model.

2. The rice mill control method based on voltage detection according to claim 1, characterized in that: The step of preprocessing the original voltage data to obtain corresponding preprocessed data includes: The raw voltage data is smoothed using a Butterworth low-pass filter and segmented into fixed-length windows to extract the effective voltage range of each data segment. Normalize the segmented data to eliminate the dimensional differences in voltage amplitude caused by power supply fluctuations; The sensor sampling delay is compensated by the time series alignment algorithm to generate time-synchronized preprocessed data.

3. The rice mill control method based on voltage detection according to claim 1, characterized in that: The step of extracting the voltage mean, voltage fluctuation amplitude, and high-frequency energy ratio from the pre-processed data to generate corresponding voltage characteristic data includes: Calculating the arithmetic mean of the preprocessed data within a preset time window to generate a voltage mean; Calculating the difference between the maximum value and the minimum value of the pre-processed data within a preset time window to generate a voltage fluctuation amplitude; The pre-processed data is decomposed by wavelet transform to extract high-frequency components, and the energy value of the high-frequency components in the preset time window and the total energy value in the preset time window are calculated to determine the high-frequency energy ratio; The voltage characteristic data is formed by combining the voltage mean value, voltage fluctuation amplitude, and high-frequency energy proportion.

4. The rice mill control method based on voltage detection according to claim 1, characterized in that: The rice milling machine further includes a feeding motor; the step of adjusting the feeding speed of the rice milling machine and the gap between the grinding rollers according to the rice milling load state parameters includes: According to the milling pressure requirement level of the rice milling load state parameter, the preset pressure-speed mapping table is queried to generate the corresponding feed motor speed control signal to adjust the feed speed; According to the estimated value of rice moisture content in the rice milling load state parameter, the gap compensation amount of the grinding roller is determined, and a corresponding grinding roller displacement control signal is generated to adjust the gap of the grinding roller.

5. The rice mill control method based on voltage detection according to claim 4, characterized in that: The method further comprises: Determine the maximum allowable load current, and apply milling pressure at equal intervals from no-load to full-load on the rice mill, and record the steady-state load current of the feed motor at the corresponding milling pressure requirement level; Calculate the speed proportional coefficient corresponding to each grinding pressure requirement level according to the following formula: ； in, Indicates the rice milling pressure requirement level; Indicates Speed ​​proportional coefficient under the grinding pressure requirement level; Indicates the maximum allowable load current; Indicates The steady-state load current of the motor under the grinding pressure requirement level; The speed of the feed motor is determined according to the speed proportional coefficient corresponding to the grinding pressure requirement level, which specifically includes: Determine the speed of the feed motor according to the following formula: ； in, Indicates the speed of the feed motor; Indicates the speed proportional coefficient; Indicates the rated speed of the feed motor; A corresponding relationship between the grinding pressure requirement level and the feed motor speed is established to construct the preset pressure-speed mapping table.

6. The rice mill control method based on voltage detection according to claim 5, characterized in that: The method further comprises: When the measured load current is greater than the steady-state load current at a certain grinding pressure requirement level, the speed is corrected according to the following formula: ； in, Indicates the corrected speed; Indicates the speed of the feed motor; Indicates the measured load current, Indicates the maximum allowable load current; Indicates The steady-state load current of the motor under the grinding pressure requirement level.

7. The rice mill control method based on voltage detection according to claim 4, characterized in that: The step of determining the gap compensation amount of the grinding roller according to the estimated value of the rice moisture content in the rice milling load state parameter includes: When the estimated value of the rice moisture content is greater than a preset first threshold value and less than a preset second threshold value, determining the gap compensation amount of the grinding roller to be 0; When the estimated value of the rice moisture content is less than a preset first threshold, determining the gap compensation amount of the grinding roller to be 0.1 mm; When the estimated value of the rice moisture content is greater than a preset second threshold value and less than a preset third threshold value, determining the gap compensation amount of the grinding roller to be 0.2 mm; When the estimated rice moisture content is greater than a preset third threshold, the gap compensation amount of the grinding roller is determined to be 0.3 mm; wherein the preset first threshold is less than the preset second threshold, and the preset second threshold is less than the preset third threshold.

8. The rice mill control method based on voltage detection according to claim 4, characterized in that: The method further comprises: Calculating a voltage change rate based on the raw voltage data; Collect the current data of the main motor and calculate the current change rate; When the current change rate is greater than a preset current change rate and the voltage change rate is less than a preset voltage change rate, the feed motor speed is controlled to decrease by a preset speed, and the roller gap is controlled to increase by a preset distance; If, after a preset time period, the current change rate remains greater than the preset current change rate and the voltage change rate remains less than the preset voltage change rate, the grinding roller is controlled to reverse.

9. A rice mill control system based on voltage detection, characterized in that: The system includes: a memory, a processor, and a rice milling machine control program based on voltage detection stored in the memory and executable on the processor. The rice milling machine control program based on voltage detection is configured to implement the steps of the rice milling machine control method based on voltage detection according to any one of claims 1 to 8.

Images