Machine-made sand proportion control method and system based on PID adjustment

Abstract

The invention provides a machine-made sand proportion control method and system based on PID regulation, and the method comprises the steps: obtaining the mixing process response data of machine-made sand mixing equipment in a continuous operation period, carrying out the proportion state decoding processing of the mixing process response data, generating a proportion state vector representing the mixing uniformity under the current proportion, and carrying out the calculation of the proportion state vector. And performing mapping analysis on the matching state vector and a preset concrete compressive strength target value, calculating a deviation adjustment amount between the current matching state and a target strength requirement, and executing PID parameter dynamic setting operation based on the deviation adjustment amount. And adjusting the proportion adjustment coefficient, the integral adjustment coefficient and the differential adjustment coefficient according to the variation trend of the deviation adjustment quantity, and generating a proportion optimization instruction containing the adjustment direction and amplitude. The real-time performance and accuracy of machine-made sand proportion control are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and system for controlling the proportion of manufactured sand based on PID regulation. Background Technology

[0002] In the field of concrete production, manufactured sand, as a substitute for natural sand, directly affects the key properties of concrete such as compressive strength and workability through its mix proportion control. The manufactured sand mix proportion control method is a technology that adjusts the feeding ratio of components such as manufactured sand, stone powder, and water to ensure that the mixed concrete meets the preset performance requirements. In existing technologies, manufactured sand mix proportion control mostly adopts direct parameter measurement methods based on flow sensors or weighing sensors, combined with fixed parameter PID control to adjust the feeding amount. That is, the real-time feeding flow rate of each component is directly collected by sensors, compared with the target flow rate, and the adjustment amount is calculated and the actuator is controlled using preset PID parameters. However, this control method relies too much on the direct measurement accuracy of sensors, is easily affected by factors such as fluctuations in material moisture content and mechanical errors of equipment, and is difficult to adapt to dynamic conditions such as changes in sand particle size distribution and cement hydration reaction during the mixing process, resulting in lag or overshoot in mix proportion adjustment. At the same time, direct measurement data cannot reflect deeper state information such as mixing uniformity, making it difficult to achieve a precise correlation between mix proportion parameters and concrete strength target values. Summary of the Invention

[0003] In view of this, the present invention provides a method and system for controlling the proportion of manufactured sand based on PID regulation.

[0004] The technical solution of this invention is implemented as follows: In a first aspect, embodiments of the present invention provide a method for controlling the proportion of manufactured sand based on PID regulation. The method includes: collecting mixing process response data of a manufactured sand mixing equipment during a continuous operating cycle; the mixing process response data is acquired by a vibration sensor installed at the bottom of the mixing cylinder, and includes vibration waveform sequences under different proportion combinations and corresponding mixing timestamps, the mixing timestamps being used to mark the mixing stage corresponding to each vibration waveform sequence; performing proportion state decoding processing on the mixing process response data to generate a proportion state vector characterizing the mixing uniformity under the current proportion, the dimension of the proportion state vector corresponding to the combination parameters of sand particle size distribution, cement dosage, and water-cement ratio during the mixing process; and connecting the proportion state vector with… A mapping analysis is performed on the preset concrete compressive strength target value to calculate the deviation adjustment amount between the current mix proportion and the target strength requirement. The deviation adjustment amount is generated by comparing the differences between the features of each dimension in the mix proportion vector and the target strength feature library. Based on the deviation adjustment amount, a PID parameter dynamic tuning operation is performed. The proportional adjustment coefficient, integral adjustment coefficient, and derivative adjustment coefficient are adjusted according to the changing trend of the deviation adjustment amount to generate a mix proportion optimization command containing the adjustment direction and amplitude. According to the mix proportion optimization command, a parameter correction signal is sent to the feed control module of the manufactured sand mixing equipment to correct the feed flow parameters of manufactured sand, stone powder, and water, so that the deviation between the corrected mix proportion parameters and the concrete compressive strength target value is controlled within a preset range.

[0005] In a second aspect, the present invention provides a computer system including a memory and a processor, wherein the memory stores a computer program that can run on the processor, and the processor executes the program to implement the steps in the method described above.

[0006] The PID-based manufactured sand proportioning control method provided by this invention collects mixing process response data over a continuous operating cycle by installing a vibration sensor at the bottom of the mixing cylinder. This data includes vibration waveform sequences under different proportion combinations and corresponding mixing timestamps. This overcomes the limitations of traditional manufactured sand proportioning control that relies on direct parameter measurement or hysteresis detection, avoiding the problems of untimely proportion adjustments caused by insufficient sensor accuracy in direct measurement or feedback delays in hysteresis detection. By decoding the vibration waveform sequences to generate a proportion state vector characterizing mixing uniformity, the implicit vibration response of the mixing process is transformed into data related to sand particle size distribution and cement dosage. The method utilizes quantifiable state information corresponding to the water-cement ratio, solving the problem of difficulty in real-time perception of mix proportion uniformity in traditional methods. By mapping the mix proportion state vector to the target value of concrete compressive strength, a deviation adjustment amount is generated. Combined with the changing trend of the deviation adjustment amount, the PID parameters are dynamically tuned, realizing the coordinated adjustment of multi-dimensional parameters such as sand particle size distribution, cement dosage, and water-cement ratio. Finally, the feed flow rate parameter is corrected according to the mix proportion optimization command, so that the deviation between the corrected mix proportion and the target strength is controlled within the preset range. This effectively improves the real-time performance and accuracy of the manufactured sand mix proportion control, ensuring that the mix proportion parameters always remain dynamically adapted to the target value of concrete compressive strength during the mixing process. Attached Figure Description

[0007] Figure 1 This is a schematic diagram illustrating the implementation process of a PID-based controlled sand proportioning method according to an embodiment of the present invention.

[0008] Figure 2 This is a schematic diagram of the composition and structure of a manufactured sand proportioning control device provided in an embodiment of the present invention.

[0009] Figure 3 This is a schematic diagram of the hardware entity of a computer system provided in an embodiment of the present invention. Detailed Implementation

[0010] This invention provides a method for controlling the proportion of manufactured sand based on PID regulation, which can be executed by a processor of a computer system. The computer system can refer to a device with data processing capabilities, such as a laptop, tablet, or desktop computer.

[0011] Figure 1 This is a schematic diagram illustrating the implementation process of a PID-based controlled proportion method for manufactured sand, as provided in an embodiment of the present invention. Figure 1 As shown, the method includes: Step S100: Obtain the mixing process response data of the manufactured sand mixing equipment during the continuous operation cycle. The mixing process response data is collected by the vibration sensor installed at the bottom of the mixing cylinder. It includes vibration waveform sequences under different ratio combinations and corresponding mixing timestamps. The mixing timestamps are used to mark the mixing stage corresponding to each vibration waveform sequence.

[0012] In this embodiment of the invention, the manufactured sand mixing equipment performs manufactured sand mixing within a continuous operating cycle. The mixing process response data is crucial information reflecting the equipment status and material conditions during this process. Vibration sensors are installed at the bottom of the mixing cylinder. Their principle is to detect the vibration of the mixing cylinder during the mixing process, converting the mechanical vibration into an electrical signal, thereby obtaining a vibration waveform sequence. Different mix proportions lead to different material motion states and interactions during mixing, resulting in different vibration waveform sequences. A mixing timestamp is a time marker added to each vibration waveform sequence to identify which stage of the mixing process it belongs to. For example, a piezoelectric vibration sensor can be used, which features high sensitivity and wide frequency response, enabling accurate acquisition of the vibration signal from the mixing cylinder. During acquisition, the sensor samples the vibration signal according to a set sampling frequency, obtaining discrete vibration waveform sequences. Simultaneously, the system records the time corresponding to each sampling point, obtaining a mixing timestamp. These data allow for analysis of the characteristics and patterns of the mixing process under different mix proportions.

[0013] Step S200: Perform mix proportion state decoding on the mixing process response data to generate a mix proportion state vector characterizing the mixing uniformity under the current mix proportion. The dimension of the mix proportion state vector corresponds to the combination parameters of sand particle size distribution, cement dosage and water-cement ratio in the mixing process.

[0014] Mix proportion state decoding involves analyzing and transforming the collected mixing process response data to obtain information reflecting the mixing uniformity under the current mix proportion. The mix proportion state vector is a multi-dimensional vector, with each dimension corresponding to a key parameter in the mixing process: sand particle size distribution, cement content, and water-cement ratio. Different combinations of these parameters affect the mixing uniformity of manufactured sand, and the mix proportion state vector can comprehensively reflect the influence of these parameters on mixing uniformity.

[0015] In some embodiments, step S200 may specifically include the following steps S210~S260: Step S210: Based on the mixing timestamp, the vibration waveform sequence is divided into sub-waveform sequences that correspond one-to-one with the mixing stage. The duration of each sub-waveform sequence is equal to the duration of the mixing stage, and the start timestamp of the sub-waveform sequence is aligned with the start timestamp of the mixing stage.

[0016] By using mixing timestamps, the start and end times of each stage in the mixing process can be clearly identified, thus segmenting the vibration waveform sequence according to the mixing stages. Each sub-waveform sequence corresponds to a mixing stage, with its duration matching the stage's duration and its start timestamp being consistent with the stage's start timestamp. This segmentation aims to analyze the vibration characteristics of each mixing stage in greater detail, as different mixing stages may exhibit different material motion states and interactions, leading to variations in vibration characteristics. For example, the mixing process of manufactured sand includes a dry mixing stage, a wet mixing stage, and a pre-discharge stage. Using mixing timestamps, the vibration waveform sequence can be accurately segmented into sub-waveform sequences corresponding to these three stages.

[0017] In some embodiments, step S210 may specifically include the following steps S211 to S216: Step S211: Parse the stage identifier information in the mixing timestamp to determine the dry mixing stage, wet mixing stage and pre-discharge stage included in the mixing process. The time boundaries of each stage are divided by the start and end marks in the timestamp.

[0018] The mixing timestamp contains stage identifiers to distinguish different mixing stages. By parsing these stage identifiers, the dry mixing stage, wet mixing stage, and pre-discharge stage can be identified. The start and end markers in the timestamp determine the time boundaries of each stage. For example, in the process of manufactured sand mixing, the dry mixing stage is the stage where the material is added to the mixing tank and stirred without water; the wet mixing stage is the stage where water is added and stirred; and the pre-discharge stage is the stage where mixing is about to end and the material is ready to be discharged. By parsing the stage identifiers and time boundary markers in the mixing timestamp, these three stages can be accurately divided. The specific parsing process can use string parsing algorithms, such as regular expression matching algorithms. The stage identifiers and time boundary markers in the mixing timestamp are matched according to a preset format to extract the information of each stage.

[0019] Step S212: Extract continuous waveform data for the corresponding time period from the vibration waveform sequence according to the stage time boundary, and generate a sub-waveform sequence with the same number as the mixing stage. The starting point of the time axis of each sub-waveform sequence is the stage start timestamp, and the ending point is the stage end timestamp.

[0020] After determining the time boundaries of the mixing stages, continuous waveform data for the corresponding time periods can be extracted from the vibration waveform sequence based on these boundaries. Each extracted waveform data forms a sub-waveform sequence, and the number of sub-waveform sequences is the same as the number of mixing stages. The start and end points of the time axis of each sub-waveform sequence correspond to the start and end timestamps of that mixing stage, respectively. For example, in the process of manufactured sand mixing, the time boundaries of the dry mixing stage, wet mixing stage, and pre-discharge stage have been determined through step S211. Based on this, continuous waveform data corresponding to these three stages can be extracted from the vibration waveform sequence to obtain three sub-waveform sequences. The extraction process can be achieved by traversing the vibration waveform sequence and filtering and extracting data based on the range of timestamps.

[0021] Step S213: Perform data integrity verification on each sub-waveform sequence to check whether there is sampling loss or abnormal jump in the waveform data. If there is an abnormality, use the data interpolation method between the previous and next time points to repair it. The size of the interpolation window is inversely proportional to the sampling frequency.

[0022] Data integrity verification is essential to ensure the accuracy and reliability of sub-waveform sequences. During the acquisition of vibration waveform data, various reasons may lead to sample loss or abnormal jumps, which can affect subsequent analysis results. Therefore, data integrity verification is necessary for each sub-waveform sequence. The interpolation method uses normal data from preceding and following time points to estimate the values ​​of abnormal data. The interpolation window size is inversely proportional to the sampling frequency; that is, the higher the sampling frequency, the smaller the interpolation window, and vice versa. For example, in the process of mixing manufactured sand, if the vibration sensor used has a sampling frequency of 100Hz, the interpolation window can be set to a relatively small value. During data integrity verification, the difference between adjacent data points can be compared to determine if abnormal jumps exist. If the difference exceeds a preset threshold, an anomaly is considered to exist. For abnormal data points, the preceding and following time point interpolation method can be used for repair. Specifically, a linear interpolation algorithm can be used to calculate the estimated value of the abnormal data point based on the values ​​of the preceding and following normal data points and the time interval.

[0023] Step S214: Sort the repaired sub-waveform sequences according to the order of the mixing stages to generate an ordered set of sub-waveform sequences. The order of each element in the set is consistent with the actual flow of the mixing process.

[0024] The purpose of sorting is to ensure that the order of the sub-waveform sequences matches the actual flow of the mixing process. After data repair, the sub-waveform sequences are arranged according to the chronological order of the mixing stages, resulting in a set of stage-ordered sub-waveform sequences. For example, in the process of manufactured sand mixing, the mixing stages are dry mixing, wet mixing, and pre-discharge. Therefore, the repaired sub-waveform sequences should also be arranged in this order. The sorting process can employ sorting algorithms, such as bubble sort or quicksort. The stage information of the sub-waveform sequences is used as the sorting key, and the sequences are sorted according to the chronological order of the stages.

[0025] Step S215: Add a stage attribute label to each sub-waveform sequence. The label content includes the stage name and duration. The stage attribute label is stored in association with the header metadata of the sub-waveform sequence.

[0026] Stage attribute tags are used to identify the mixing stage and duration of each sub-waveform sequence. Adding stage attribute tags to sub-waveform sequences facilitates data management and analysis. The stage attribute tags are stored in association with the header metadata of the sub-waveform sequence, allowing for quick retrieval of relevant information for each sub-waveform sequence in subsequent processing. For example, in the process of manufactured sand mixing, a stage attribute tag can be added to the dry mixing stage sub-waveform sequence, with the tag content "Dry mixing stage, duration: XX seconds". This tag is then stored in association with the header metadata of the sub-waveform sequence, allowing the stage information of the sub-waveform sequence to be retrieved by reading the header metadata. This association can be stored using a database or file, with the stage attribute tag and the header metadata of the sub-waveform sequence stored as a single record.

[0027] Step S216: Establish a mapping table between sub-waveform sequences and mixing stages, recording the timestamp range, data length, and corresponding stage index of each sub-waveform sequence. The mapping table is used for stage matching during subsequent feature extraction.

[0028] The mapping table facilitates subsequent feature extraction by enabling the rapid and accurate identification of the mixing stage corresponding to each sub-waveform sequence. By recording the timestamp range, data length, and corresponding stage index of each sub-waveform sequence, a one-to-one correspondence between sub-waveform sequences and mixing stages can be established. For example, in the process of manufactured sand mixing, for each sub-waveform sequence, its start timestamp, end timestamp, data length, and corresponding mixing stage index are recorded. During subsequent feature extraction, when it is necessary to analyze the characteristics of a specific mixing stage, the corresponding sub-waveform sequence can be quickly found using the mapping table. The mapping table can be stored as a two-dimensional array or a database table, where each row represents a sub-waveform sequence and each column represents a record item, such as the timestamp range, data length, and corresponding stage index.

[0029] Step S220: Perform Fourier transform processing on each sub-waveform sequence to convert the time-domain vibration signal into a frequency-domain signal, generating a preliminary frequency spectrum containing multiple frequency components. The frequency resolution of the preliminary frequency spectrum is proportional to the length of the sub-waveform sequence.

[0030] Fourier transform converts a vibration signal in the time domain into a frequency domain signal, thus obtaining a preliminary frequency spectrum containing multiple frequency components. Frequency resolution refers to the smallest frequency interval that can be distinguished in the frequency domain, and it is proportional to the length of the sub-waveform sequence. That is, the longer the sub-waveform sequence, the higher the frequency resolution, and the more accurately the frequency components in the signal can be analyzed. For example, in the process of mixing manufactured sand, each sub-waveform sequence can be processed using the Fast Fourier Transform (FFT) algorithm. The FFT algorithm can quickly calculate the frequency domain representation of the signal. By performing FFT processing on each sub-waveform sequence, a preliminary frequency spectrum is obtained. In practical applications, the FFT parameters can be adjusted according to the length of the sub-waveform sequence to obtain a suitable frequency resolution.

[0031] Step S230: Perform frequency band screening on the preliminary frequency spectrum through a bandpass filter, retain the target frequency band corresponding to the particle size distribution range of sand, and remove high-frequency noise and low-frequency interference signals. The upper and lower limits of the target frequency band are set according to the standard particle size range of manufactured sand production.

[0032] A bandpass filter is a filter used to filter signals within a specific frequency band. By processing the initial frequency spectrum, it can retain the target frequency band corresponding to the particle size distribution range of sand while removing high-frequency noise and low-frequency interference signals. The upper and lower limits of the target frequency band are set according to the standard particle size range for manufactured sand production. This is because sand particles of different sizes generate signals of different frequencies when vibrating. By selecting the frequency band corresponding to the standard particle size range, the sand particle size distribution can be analyzed more accurately. For example, in manufactured sand production, the target frequency band is determined to be 10Hz-100Hz based on the standard particle size range. Using a bandpass filter to process the initial frequency spectrum retains only the frequency components within this band, removing high-frequency noise above 100Hz and low-frequency interference signals below 10Hz. Bandpass filters can include Butterworth filters, Chebyshev filters, etc. In practical applications, it is necessary to select the appropriate filter type and parameters based on specific needs and signal characteristics.

[0033] In some embodiments, step S230 may specifically include the following steps S231 to S236: Step S231: Obtain the standard particle size distribution range for manufactured sand production, and convert the particle size range into the corresponding vibration frequency range. The conversion is based on the empirical correlation between the collision vibration frequency of sand particles and the particle size. The larger the particle size, the lower the corresponding frequency.

[0034] The standard particle size distribution range for manufactured sand production is determined based on production requirements and relevant standards. There is an empirical correlation between the collision vibration frequency of sand particles and their size; that is, the larger the particle size, the lower the collision vibration frequency. This correlation allows the standard particle size distribution range to be converted into a corresponding vibration frequency range. For example, in manufactured sand production, the standard particle size distribution range is 0.1mm-5mm. According to the empirical correlation, a 0.1mm particle size might correspond to a vibration frequency of 100Hz, while a 5mm particle size might correspond to a vibration frequency of 10Hz. Therefore, the corresponding vibration frequency range is 10Hz-100Hz. The standard particle size distribution range can be obtained by consulting relevant production standard documents or from the production management system. Converting the particle size range into a frequency range can be achieved using a pre-established empirical correlation table or mathematical model.

[0035] Step S232: Set the passband parameters of the bandpass filter according to the vibration frequency range obtained by conversion. The lower limit frequency of the passband corresponds to the vibration frequency of the largest particle size, and the upper limit frequency corresponds to the vibration frequency of the smallest particle size.

[0036] Once the vibration frequency range is obtained, the passband parameters of the bandpass filter can be set accordingly. The lower limit frequency of the passband corresponds to the vibration frequency of the largest particle size, and the upper limit frequency corresponds to the vibration frequency of the smallest particle size. This setting aims to retain only the frequency components corresponding to the standard particle size range. For example, in the previous example, the vibration frequency range was 10Hz-100Hz, so the lower limit frequency of the bandpass filter was set to 10Hz, and the upper limit frequency was set to 100Hz. By setting appropriate passband parameters, it can be ensured that the bandpass filter can accurately filter out signals in the target frequency band.

[0037] Step S233: Perform bandpass filtering based on Butterworth filter. The order of the filter is set according to the noise suppression requirements. The higher the order, the better the filtering effect, but the risk of signal distortion increases.

[0038] Butterworth filters have flat passband and stopband characteristics, making them a suitable choice for bandpass filtering. The filter order is set based on noise suppression requirements; higher orders generally offer better noise suppression but also increase the risk of signal distortion. For example, in manufactured sand production environments with significant noise interference, stronger noise suppression capabilities are needed, necessitating the selection of higher filter orders, such as 8th or 10th order. However, in practical applications, a trade-off must be struck based on specific circumstances to avoid signal distortion impacting subsequent analysis. The appropriate filter order can be determined experimentally or through simulation.

[0039] Step S234: Input the preliminary frequency spectrum into the configured bandpass filter, filter each frequency component, retain the frequency components within the passband, and attenuate the frequency components outside the passband.

[0040] After configuring the bandpass filter parameters, the initial frequency spectrum is input into the filter for filtering. The filter evaluates each frequency component, retaining those within the passband and attenuating those outside the passband. For example, if the bandpass of the previously configured bandpass filter is 10Hz-100Hz, for each frequency component in the initial frequency spectrum, if its frequency is between 10Hz and 100Hz, that component is retained; if the frequency is below 10Hz or above 100Hz, it is attenuated. The filtering process can be implemented using the filter's transfer function. The initial frequency spectrum is used as the input signal, and the filtered frequency spectrum is obtained after calculating the filter's transfer function.

[0041] Step S235: Calculate the energy retention rate of the frequency spectrum before and after filtering. The energy retention rate is the ratio of the total energy after filtering to the total energy before filtering. If the retention rate is lower than the preset threshold, readjust the filter parameters.

[0042] The energy retention rate reflects the loss of signal energy during filtering. By calculating the energy retention rate of the frequency spectrum before and after filtering, the performance of the filter can be evaluated. If the energy retention rate is lower than a preset threshold, it indicates excessive signal energy loss during filtering, which may affect subsequent analysis results. In this case, the filter parameters need to be readjusted. The method for calculating the energy retention rate is to calculate the total energy of the frequency spectrum before and after filtering, and then divide the total energy after filtering by the total energy before filtering to obtain the ratio. The total energy can be calculated by summing the energy of each frequency component in the frequency spectrum.

[0043] Step S236: Output the filtered target frequency spectrum. The target frequency spectrum contains only the effective frequency components related to the sand particle size distribution. The bandwidth is proportional to the span of the standard particle size distribution range.

[0044] After the preceding filtering and energy retention rate assessment, the target frequency spectrum is output. This spectrum contains only the effective frequency components related to sand grain gradation, and the bandwidth is proportional to the span of the standard particle size distribution range. For example, a larger standard particle size distribution range will result in a larger bandwidth for the corresponding target frequency spectrum. The target frequency spectrum can provide more accurate information for subsequent analysis, enabling further analysis of sand grain gradation.

[0045] Step S240: Perform inverse Fourier transform processing on the filtered frequency domain signal to reconstruct the time domain vibration signal and extract the amplitude change curve. The horizontal axis of the amplitude change curve is relative time, and the vertical axis is the vibration amplitude. The amplitude unit corresponds to the range of the vibration sensor.

[0046] The inverse Fourier transform (IFT) is the reverse process of the Fourier transform. By performing an IFT on the filtered frequency domain signal, the frequency domain signal can be converted back to a time domain signal, thus reconstructing the time-domain vibration signal. The amplitude variation curve is extracted from the reconstructed time-domain vibration signal, reflecting the change of the vibration signal's amplitude over time. The horizontal axis of the amplitude variation curve represents relative time, and the vertical axis represents the vibration amplitude, with the amplitude unit corresponding to the range of the vibration sensor. For example, in the process of mixing manufactured sand, if the vibration sensor's range is 0-10V, then the amplitude unit on the vertical axis of the amplitude variation curve is V. By analyzing the amplitude variation curve, we can understand the changing trend and characteristics of the vibration signal, and thus infer some aspects of the mixing process. The IFT can be implemented using the inverse fast Fourier transform (IFFT) algorithm, which is the inverse operation of the FFT algorithm. After obtaining the reconstructed time-domain vibration signal, we can extract the corresponding amplitude value by traversing each sampling point of the signal, and then plot the amplitude variation curve.

[0047] Step S250: Calculate the energy proportion index of each frequency band based on the frequency spectrum of the target frequency band, and generate the sand particle size distribution characteristic component through the nonlinear mapping relationship between the energy proportion index and the sand particle size distribution parameter. The energy proportion index is the ratio of the energy value of each frequency band to the total energy value.

[0048] The frequency spectrum of the target frequency band contains energy information for different frequency bands. By calculating the energy proportion index of each frequency band, the proportion of energy in the total energy can be understood. A nonlinear mapping relationship exists between the energy proportion index and sand particle size distribution parameters, which can be used to generate characteristic components of sand particle size distribution. For example, in manufactured sand production, the frequency spectrum of the target frequency band can be divided into multiple bands, such as low-frequency, mid-frequency, and high-frequency bands. The energy value of each frequency band is calculated and divided by the total energy value to obtain the energy proportion index of each band. Then, through a pre-established nonlinear mapping model, these energy proportion indices are converted into characteristic components of sand particle size distribution. This nonlinear mapping model can be trained using machine learning algorithms, using a large amount of historical data to establish the relationship between the energy proportion index and the sand particle size distribution parameters. The energy value of each frequency band can be calculated by summing the energy of the frequency components within each band in the frequency spectrum; the total energy value is the sum of the energy values ​​of all frequency bands.

[0049] Step S260: Based on the peak point and decay rate of the amplitude change curve, and combined with the influence law of cement dosage and water-cement ratio on vibration damping characteristics, generate cement dosage characteristic components and water-cement ratio characteristic components. Then, perform vector splicing processing on the sand particle size distribution characteristic components, cement dosage characteristic components, and water-cement ratio characteristic components in a preset dimension order to generate a mix proportion state vector associated with mixing uniformity.

[0050] The peak points and decay rates of the amplitude variation curve reflect the dynamic characteristics of the vibration signal. Cement dosage and water-cement ratio affect the vibration damping characteristics, thus influencing the amplitude variation curve. By analyzing the peak points and decay rates of the amplitude variation curve, and combining this with the influence of cement dosage and water-cement ratio on vibration damping characteristics, characteristic components of cement dosage and water-cement ratio can be generated. For example, in the process of mixing manufactured sand, when the cement dosage increases, the vibration damping increases, and the decay rate of the amplitude variation curve may accelerate. Changes in the water-cement ratio also affect the vibration damping, thereby affecting the peak points and decay rates of the amplitude variation curve. By establishing corresponding mathematical models or empirical formulas, the characteristic components of cement dosage and water-cement ratio can be calculated based on the characteristics of the amplitude variation curve and the influence of cement dosage and water-cement ratio. Then, the previously generated characteristic components of sand particle size distribution, cement dosage, and water-cement ratio are vector-stitched according to a preset dimensional order to generate a comprehensive mix proportion state vector. This vector reflects the mixing uniformity under the current mix proportion. Vector concatenation can be achieved by arranging the three feature components sequentially to obtain a multidimensional vector.

[0051] Step S300: Map the mix proportion state vector to the preset concrete compressive strength target value, calculate the deviation adjustment amount between the current mix proportion state and the target strength requirement. The deviation adjustment amount is generated by comparing the difference between the features of each dimension in the mix proportion state vector and the target strength feature library.

[0052] Mapping analysis associates the mix proportion state vector with a preset target value for concrete compressive strength to identify the relationship between the two. The target strength feature library is a database containing feature information corresponding to different target values ​​for concrete compressive strength. By comparing the differences between the features of each dimension in the mix proportion state vector and the target strength feature library, the deviation adjustment amount between the current mix proportion state and the target strength requirement can be calculated. For example, in manufactured sand production, the preset target value for concrete compressive strength is C30. The target strength feature library stores feature information such as sand particle size distribution, cement content, and water-cement ratio corresponding to C30 strength. The current mix proportion state vector is compared with the C30 feature information in the target strength feature library to calculate the differences in each dimension of the features, and then a deviation adjustment amount is generated based on these differences. The deviation adjustment amount can be calculated using distance metrics such as Euclidean distance and Manhattan distance to measure the degree of difference between the mix proportion state vector and the feature information in the target strength feature library. Specifically, the target strength feature library pre-stores one or more corresponding standard target mix proportion state vectors for each preset target value for concrete compressive strength (such as C30). These target vectors were determined through statistical analysis, clustering, and optimization of the mix proportion state vectors corresponding to high-quality batches that consistently achieve the specified strength level from historical production data. They represent the optimal or typical combination of characteristics for each mix proportion parameter (sand gradation, cement content, water-cement ratio) to achieve the desired strength. During comparison, the target mix proportion state vector corresponding to the current production target (e.g., C30) is retrieved from the feature library. The process of calculating the degree of difference involves measuring the distance between the current mix proportion state vector (representing the current actual mixing state) and the target mix proportion state vector in a multi-dimensional space. Using Euclidean distance as an example, a scalar distance value is obtained by taking the square root of the sum of the squares of the differences between the two vectors in each corresponding dimension. This value quantitatively reflects the overall degree of deviation of the current mix proportion state from the target state. Simultaneously, by analyzing the component differences of the two vectors in each specific dimension (sand gradation, cement content, water-cement ratio), the individual deviation direction and magnitude of each parameter can be obtained, i.e., the degree of difference in characteristics of each dimension.

[0053] When generating deviation adjustment amounts based on the degree of difference, the deviation adjustment amount is a vector containing three components, indicating the direction and magnitude of adjustments needed for sand gradation, cement content, and water-cement ratio. The generation process incorporates technological transformations. For example, current calculations may show that the characteristic values ​​of sand gradation and cement content are both below the target (negative differences), while the water-cement ratio is above the target (positive difference). To reduce the overall difference, the deviation adjustment amount will indicate the need to increase the proportion of sand gradation and cement content while decreasing the water-cement ratio. The specific adjustment magnitude for each dimension is determined by the proportion of its corresponding difference in the overall difference, multiplied by a pre-set coefficient based on the effectiveness of the parameter adjustment. The final generated deviation adjustment amounts might be, for example: sand gradation adjustment component +6.5, cement content adjustment component +8.0, and water-cement ratio adjustment component -4.2. This result clearly provides the specific adjustment instructions for each feed parameter when correcting the current mix proportion towards the C30 target.

[0054] In some embodiments, step S300 may specifically include the following steps S310~S360: Step S310: Call the pre-trained mix proportion-strength prediction model to perform feature space transformation processing on the mix proportion state vector to generate the predicted value of concrete compressive strength corresponding to the current mix proportion. The mix proportion-strength prediction model is trained by historical mix proportion state vector samples and corresponding measured compressive strength value samples.

[0055] The mix proportion-strength prediction model is a trained machine learning model capable of predicting the corresponding compressive strength of concrete based on an input mix proportion state vector. This model is trained using a large number of historical mix proportion state vector samples and corresponding measured compressive strength samples, learning the relationship between mix proportion state and concrete compressive strength. For example, in manufactured sand production, mix proportion state vector samples and corresponding measured concrete compressive strength samples from different mix proportion states over a past period are collected. These samples are used to train the mix proportion-strength prediction model, enabling it to accurately predict the compressive strength of concrete under different mix proportion states. In practical applications, the current mix proportion state vector is input into the pre-trained mix proportion-strength prediction model. The model performs feature space transformation processing and outputs the predicted compressive strength value of the concrete corresponding to the current mix proportion. Feature space transformation processing is an internal computational process within the model, converting the input mix proportion state vector into a feature representation that the model can understand and process. Then, the predicted value is obtained through calculation using the model's learned parameters.

[0056] In some embodiments, step S310 may specifically include the following steps S311 to S316: Step S311: Input the matching state vector into the input layer of the matching-intensity prediction model. The number of neurons in the input layer is equal to the dimension of the matching state vector. Each neuron receives a feature component of one dimension of the matching state vector.

[0057] The input layer is the first layer of the mix proportion-intensity prediction model, responsible for receiving the input mix proportion state vector. The number of neurons in the input layer is equal to the dimension of the mix proportion state vector, so that each neuron can receive the feature components of one dimension of the mix proportion state vector. For example, if the mix proportion state vector is a three-dimensional vector corresponding to the three dimensions of sand particle size distribution, cement content, and water-cement ratio, then the input layer has three neurons, each receiving the feature components of these three dimensions. In this way, the information of the mix proportion state vector is passed to the model for subsequent processing.

[0058] Step S312: The input features are linearly transformed and nonlinearly activated through the first hidden layer of the ratio-intensity prediction model. The weight matrix of the linear transformation is obtained through optimization during the training process, and the ReLU function is used to enhance the nonlinear fitting ability of the model.

[0059] The first hidden layer is an intermediate layer in the model, performing linear transformations and non-linear activation on the input features. The linear transformation involves matrix multiplication of the input features with the weight matrix, linearly combining them. The weight matrix is ​​continuously adjusted and optimized during model training, enabling the model to better learn the relationship between input features and output. Non-linear activation processes the result of the linear transformation using an activation function, converting it into a non-linear output. In this embodiment, the ReLU (Rectified Linear Unit) activation function is used, which enhances the model's non-linear fitting ability, allowing it to handle more complex relationships. For example, for the result after the linear transformation, the ReLU function sets values ​​less than 0 to 0, while keeping values ​​greater than 0 unchanged, thus introducing non-linear characteristics.

[0060] Step S313: Input the output features of the first hidden layer into the second hidden layer for secondary feature extraction and dimensionality compression. The number of neurons in the second hidden layer is less than that in the first hidden layer. Feature aggregation enhances the expressive power of key influencing factors.

[0061] The second hidden layer further processes the output features of the first hidden layer, performing secondary feature extraction and dimensionality compression. Secondary feature extraction extracts more representative features from the output features of the first hidden layer, while dimensionality compression reduces the number of dimensions in the features. The second hidden layer has fewer neurons than the first hidden layer, allowing for enhanced representation of key influencing factors through feature aggregation. For example, in manufactured sand production, the output features of the first hidden layer may contain a lot of detailed information, but some of this information has a relatively small impact on the compressive strength of concrete. The second hidden layer can aggregate these features, highlighting key influencing factors while reducing feature dimensionality, thus improving the model's computational efficiency and generalization ability.

[0062] Step S314: Perform batch normalization on the output features of the second hidden layer to eliminate the dimensional differences between different feature dimensions. The normalization operation is based on the feature mean and variance statistically analyzed during training.

[0063] Batch normalization is used to eliminate dimensional differences between different feature dimensions, enabling the model to learn and train more stably. During training, the mean and variance of each feature dimension are calculated. Then, in the batch normalization operation, these statistics are used to normalize the output features of the second hidden layer. For example, for each feature dimension, the mean of that dimension is subtracted, and then divided by the standard deviation of that dimension, transforming the feature values ​​into a distribution with a mean of 0 and a variance of 1. This avoids certain feature dimensions having an excessive impact on model training due to their large dimensions, improving the model's convergence speed and stability.

[0064] Step S315: Input the normalized features into the third hidden layer. Use the dropout operation in the third hidden layer to prevent overfitting. The dropout ratio is dynamically adjusted according to the number of training samples.

[0065] The third hidden layer introduces dropout to prevent overfitting. Overfitting occurs when a model performs well on training data but poorly on test data. Dropout randomly sets the output of a subset of neurons to 0, preventing the model from over-relying on certain neurons during training and thus enhancing its generalization ability. The dropout ratio refers to the proportion of neurons set to 0, which is dynamically adjusted based on the number of training samples. For example, when the number of training samples is small, the dropout ratio can be increased to enhance the model's randomness; when the number of training samples is large, the dropout ratio can be decreased. In this way, the model's fitting ability and generalization ability are balanced.

[0066] Step S316: Linearly map the output features of the third hidden layer through the output layer of the mix proportion-strength prediction model to generate a single numerical value of the concrete compressive strength prediction. The bias term of the output layer is set according to the historical strength benchmark value.

[0067] The output layer is the final layer of the model. It performs a linear mapping on the output features of the third hidden layer to generate a single numerical predicted value for the concrete compressive strength. The linear mapping process involves linearly combining the weight matrix and bias terms with the output features of the third hidden layer to obtain the final predicted value. The bias terms of the output layer are set based on historical strength benchmark values, making the model's predicted values ​​more consistent with reality. For example, in manufactured sand production, a benchmark value for concrete compressive strength is obtained based on historical data, and this benchmark value is used as the bias term for the output layer.

[0068] Step S320: Compare the predicted value of concrete compressive strength with the target value of concrete compressive strength, and calculate the difference between the two as the basic value of strength deviation. When the predicted value is greater than the target value, the basic value of strength deviation is positive, and when the predicted value is less than the target value, the basic value of strength deviation is negative.

[0069] The baseline strength deviation is an indicator that measures the difference between the predicted compressive strength of concrete under the current mix design and the target strength. The baseline strength deviation is calculated by comparing the predicted compressive strength with the target compressive strength.

[0070] Step S330: Perform sensitivity analysis on the characteristics of each dimension in the mix proportion state vector to determine the sensitivity coefficients of the sand particle size distribution characteristic component, cement content characteristic component, and water-cement ratio characteristic component to the compressive strength of concrete. The sensitivity coefficient represents the amount of change in concrete compressive strength caused by a change in a unit characteristic component.

[0071] Sensitivity analysis aims to determine the influence of each dimension of the mix proportion state vector on the compressive strength of concrete. By performing sensitivity analysis on the characteristic components of sand gradation, cement content, and water-cement ratio, their sensitivity coefficients to the compressive strength of concrete can be obtained. The sensitivity coefficient represents the change in compressive strength of concrete caused by a unit change in a characteristic component.

[0072] In some embodiments, step S330 may specifically include the following steps S331 to S335: Step S331: Select a historical mix proportion state vector sample set. The sample set contains mix proportion state vectors under different mix proportion combinations and the corresponding measured values ​​of concrete compressive strength. The sample covers the entire adjustable range of mix proportion parameters.

[0073] The historical mix design vector sample set serves as the foundation for sensitivity analysis. This sample set contains mix design vectors for different mix design combinations and corresponding measured values ​​of concrete compressive strength, covering the entire adjustable range of mix design parameters. For example, in manufactured sand production, sand particle size distribution, cement content, and water-cement ratio all have certain adjustable ranges; the samples in the sample set should encompass various combinations of these parameters within these adjustable ranges. By selecting such a sample set, a more comprehensive understanding of the impact of different mix design combinations on concrete compressive strength can be achieved. The sample set can be selected and organized from historical production data to ensure the diversity and representativeness of the samples.

[0074] Step S332: For each mix proportion state vector sample in the sample set, adjust the sand particle size distribution characteristic component, cement dosage characteristic component, and water-cement ratio characteristic component individually in sequence. Adjust one component at a time while keeping the other components unchanged. The adjustment range is a fixed proportion of the adjustable range of the component.

[0075] To determine the sensitivity of each feature component to the compressive strength of concrete, each mix design vector sample in the sample set needs to be adjusted individually. Only one feature component is adjusted at a time, while the other components remain unchanged, and the adjustment range is a fixed proportion of the adjustable range of that component.

[0076] Step S333: Calculate the change in the measured value of concrete compressive strength before and after adjustment. Use the ratio of the change to the adjustment range as the sensitivity coefficient of the characteristic component. The larger the absolute value of the sensitivity coefficient, the more significant the influence of the characteristic component on the strength.

[0077] After adjusting each characteristic component, the change in the measured value of concrete compressive strength before and after the adjustment is calculated. The ratio of this change to the adjustment range is used as the sensitivity coefficient of that characteristic component. The larger the absolute value of the sensitivity coefficient, the more significant the influence of that characteristic component on the concrete compressive strength. By calculating these sensitivity coefficients, the importance of each characteristic component in the formation of concrete compressive strength can be understood.

[0078] Step S334: Perform statistical averaging on the sensitivity coefficients of all samples, remove outlier samples that deviate from the preset range of the average value, and then calculate the average sensitivity coefficient. The criteria for judging outlier samples are the dispersion criteria in the statistical distribution.

[0079] To obtain a more accurate sensitivity coefficient, the sensitivity coefficients of all samples are statistically averaged. During this process, outliers that deviate from the preset range of the average are removed. Outliers are identified based on the dispersion criteria in statistical distributions, such as using the standard deviation. If the difference between the sensitivity coefficient of a sample and the mean exceeds a preset multiple of the standard deviation, then that sample is considered an outlier. After removing outliers, the average sensitivity coefficient of the remaining samples is calculated.

[0080] Step S335: Establish a correlation table between the influence weights and the proportion parameters, record the influence weights of sand particle size distribution, cement dosage and water-cement ratio and the corresponding sample statistics. The correlation table is used for weight lookup when calculating the subsequent deviation adjustment amount.

[0081] The correlation table between influence weights and mix proportion parameters records the influence weights of sand gradation, cement dosage, and water-cement ratio, along with corresponding sample statistics. Influence weights can be determined based on the average sensitivity coefficient; the higher the sensitivity coefficient, the greater the influence weight. This correlation table is used for weight lookup during subsequent deviation adjustment calculations, ensuring that adjustments to mix proportion parameters are made reasonably based on the importance of each parameter. For example, when calculating deviation adjustment amounts, the influence weights of sand gradation, cement dosage, and water-cement ratio are looked up from the correlation table, and then the adjustment amounts are allocated based on these weights. The correlation table can be stored as a two-dimensional array or a database table, where each row represents a mix proportion parameter, and each column represents a record item, such as influence weight and sample statistics.

[0082] Step S340: Based on the basic value of the strength deviation and the sensitivity coefficient of each characteristic component, calculate the deviation adjustment component corresponding to each characteristic component. The magnitude of the deviation adjustment component is the negative value of the ratio of the basic value of the strength deviation to the corresponding sensitivity coefficient.

[0083] The deviation adjustment component is an adjustment amount calculated to bring the current mix proportion closer to the target strength. Based on the baseline strength deviation value and the sensitivity coefficients of each characteristic component, the deviation adjustment component corresponding to each characteristic component can be calculated. The magnitude of the deviation adjustment component is the negative of the ratio of the baseline strength deviation value to the corresponding sensitivity coefficient. By calculating these deviation adjustment components, the direction and magnitude of adjustment required for each characteristic component can be determined to narrow the gap between the current mix proportion and the target strength.

[0084] Step S350: Verify the directional consistency of each deviation adjustment component to ensure that the adjustment directions of sand particle size distribution, cement dosage and water-cement ratio conform to the variation law of concrete strength.

[0085] Directional consistency verification is performed to ensure that the adjustment directions of sand gradation, cement dosage, and water-cement ratio conform to the changing patterns of concrete strength. For example, increasing cement dosage increases concrete strength. Therefore, during adjustment, if the base strength deviation is negative, indicating a need to increase strength, the cement dosage deviation adjustment component should be positive, signifying the need to increase cement dosage. By verifying the directional consistency of each deviation adjustment component, unreasonable adjustments can be avoided, making the adjustment process more scientific and effective. The verification process can be based on the fundamental principles and empirical rules of concrete strength, checking and adjusting the direction of each deviation adjustment component.

[0086] Step S360: Combine the verified deviation adjustment components in the order of the dimensions of the mix proportion state vector to generate a deviation adjustment amount that includes sand gradation adjustment component, cement dosage adjustment component and water-cement ratio adjustment component. The dimension of the deviation adjustment amount is the same as the dimension of the mix proportion state vector.

[0087] After completing the directional consistency verification, the verified deviation adjustment components are combined according to the dimensional order of the mix proportion state vector to generate the deviation adjustment amount. The deviation adjustment amount includes sand gradation adjustment components, cement dosage adjustment components, and water-cement ratio adjustment components, and its dimension is the same as that of the mix proportion state vector. For example, if the mix proportion state vector is a three-dimensional vector corresponding to sand gradation, cement dosage, and water-cement ratio, then the deviation adjustment amount is also a three-dimensional vector, corresponding to the adjustment components of these three parameters. In this way, the direction and magnitude of adjustment for each parameter can be clearly represented, providing accurate input for subsequent PID control.

[0088] Step S400: Perform dynamic tuning of PID parameters based on the deviation adjustment amount, adjust the proportional adjustment coefficient, integral adjustment coefficient and derivative adjustment coefficient according to the changing trend of the deviation adjustment amount, and generate a ratio optimization instruction containing the adjustment direction and amplitude.

[0089] Dynamic tuning of PID parameters involves adjusting the proportional, integral, and derivative control coefficients of the PID controller based on the changing trend of the deviation. The proportional control coefficient is used for rapid response to the current deviation, the integral control coefficient is used to eliminate steady-state errors, and the derivative control coefficient is used to predict the changing trend of the deviation and adjust in advance. By dynamically adjusting these coefficients, the PID controller can better adapt to different deviation adjustment situations, improving the accuracy and stability of the adjustment. For example, when the deviation changes rapidly, the derivative control coefficient is increased to adjust in advance; when there is a steady-state error, the integral control coefficient is increased to eliminate the error. Based on the adjusted PID parameters, a proportioning optimization command containing the adjustment direction and amplitude is generated to guide the feeding control of the manufactured sand mixing equipment.

[0090] In some embodiments, step S400 may specifically include the following steps S410~S470: Step S410: Decompose the deviation adjustment amount into sand particle size distribution adjustment component, cement dosage adjustment component and water-cement ratio adjustment component. Each component corresponds to the dimension order of the mix proportion state vector and constitutes an independent component adjustment sequence. Each component adjustment sequence contains the historical adjustment data of the current cycle and the previous two cycles. The sequence data of different components are stored independently and do not interfere with each other.

[0091] To analyze the adjustment of each parameter in more detail, the deviation adjustment is decomposed into sand gradation adjustment components, cement dosage adjustment components, and water-cement ratio adjustment components. Each component corresponds to the dimension order of the mix proportion state vector, forming an independent component adjustment sequence. Each component adjustment sequence includes historical adjustment data for the current period and the previous two periods, allowing observation of the trend of adjustment. The sequence data of different components are stored independently, without interference, facilitating subsequent individual analysis and processing of each component. For example, for the sand gradation adjustment component, an independent component adjustment sequence is established, recording the adjustment data for the current period, the previous period, and the two periods prior. Through this historical data, the trend of sand gradation adjustment can be analyzed, providing a basis for adjusting PID parameters.

[0092] Step S420: Calculate the rate of change index for each component adjustment sequence separately. The rate of change index is obtained by dividing the difference between the current period component value and the previous period component value by the difference between the previous period component value and the previous two period component values. It reflects the acceleration of the change in the adjustment amount of the component. The rate of change index of each component is calculated independently, and only the historical data of the same component is used in the calculation process.

[0093] The rate of change index reflects the acceleration of change in each component adjustment. It is obtained by dividing the difference between the current period component value and the previous period component value by the difference between the previous period component value and the values ​​of the two previous periods component values. For example, for the sand gradation adjustment component, if the current period component value is 10, the previous period component value is 8, and the values ​​of the two previous periods are 6, then the rate of change index is (10-8) / (8-6) = 1. By calculating the rate of change index for each component adjustment sequence, we can understand the trend and acceleration of the component adjustment, providing a reference for adjusting PID parameters. Each component rate of change index is calculated independently, using only historical data for the same component to ensure that each index accurately reflects the change of that component.

[0094] In some embodiments, step S420 may specifically include the following steps S421 to S427: Step S421: Select the target component from the sand particle size distribution adjustment component, cement dosage adjustment component, and water-cement ratio adjustment component. The target component is the component of the change rate index to be calculated. Process it in the order of sand particle size distribution, cement dosage, and water-cement ratio. The processing order is consistent with the dimensional order of the mix proportion state vector.

[0095] To calculate the rate of change index for each component adjustment sequence sequentially, target components need to be selected from the sand gradation adjustment component, cement dosage adjustment component, and water-cement ratio adjustment component. Processing in the order of sand gradation, cement dosage, and water-cement ratio, consistent with the dimensional order of the mix proportion state vector, ensures the continuity and consistency of the processing. For example, first select the sand gradation adjustment component as the target component and calculate its rate of change index, then select the cement dosage adjustment component, and finally select the water-cement ratio adjustment component. This sequential processing allows for the systematic calculation of the rate of change index for each component.

[0096] Step S422: Extract the component adjustment sequence of the target component. The sequence includes the component values ​​of the previous two cycles, the component value of the previous cycle, and the component value of the current cycle, arranged in chronological order. The unit of the component value is consistent with the unit of the adjustment amount of the component to ensure data dimension uniformity.

[0097] After selecting the target component, its corresponding component adjustment sequence is extracted. This sequence contains the component values ​​from the previous two periods, the previous period, and the current period, arranged in chronological order. The unit of the component value is consistent with the unit of its adjustment amount, thus ensuring data dimensional consistency and avoiding calculation errors caused by inconsistent units. For example, for the sand gradation adjustment component, if its adjustment amount is in percentage, then the component values ​​in the component adjustment sequence should also be percentages. By extracting an accurate component adjustment sequence, a reliable data foundation can be provided for subsequent calculations of the rate of change index.

[0098] Step S423: Calculate the difference between the previous period component value and the previous two period component values ​​of the target component, and use it as the first difference value. This difference value reflects the change of the target component during the previous two periods, and the sign of the difference value indicates the direction of change.

[0099] The first difference is the difference between the target component's value in the previous period and the value in the previous two periods. It reflects the amount of change in the target component during the previous two periods. The sign of the difference indicates the direction of change: a positive value indicates an increase, and a negative value indicates a decrease.

[0100] Step S424: Calculate the difference between the current period component value and the previous period component value of the target component, and use it as the second difference value. This difference value reflects the change of the target component in the most recent two periods, and the sign of the difference value indicates the direction of change.

[0101] The second difference is the difference between the target component's value in the current period and the component value in the previous period, reflecting the change in the target component over the most recent two periods. The sign of the difference also indicates the direction of change: a positive value indicates an increase, and a negative value indicates a decrease.

[0102] Step S425: If the first difference is zero, it indicates that the target component has not changed in the first two cycles. At this time, the rate of change index is set as the preset benchmark value. The benchmark value is set according to the historical adjustment trend of the component and is related to the component type. The benchmark values ​​of different components are set independently.

[0103] When the first difference is zero, it indicates that the target component has not changed in the first two cycles. In this case, the rate of change index cannot be calculated using the normal formula; therefore, a preset benchmark value is set for the rate of change index. The preset benchmark value is set based on the historical adjustment trend of the component, and the benchmark values ​​for different components are set independently.

[0104] Step S426: If the first difference is not zero, the ratio of the second difference to the first difference is used as the rate of change index. A positive ratio indicates that the target component is changing faster, and a negative ratio indicates that the change is slowing down. The absolute value of the ratio reflects the degree of acceleration or deceleration. The ratios of different components are calculated independently.

[0105] When the first difference is not zero, the rate of change index is calculated according to the normal formula, that is, the ratio of the second difference to the first difference is used as the rate of change index. A positive ratio indicates that the change of the target component is accelerating, and a negative ratio indicates that the change is decelerating. The absolute value of the ratio reflects the degree of acceleration or deceleration. The ratios of different components are calculated independently to ensure that the rate of change index of each component can accurately reflect its own change.

[0106] Step S427: Limit the range of the calculated rate of change index. When the absolute value of the index exceeds the maximum allowable rate of change of the component, take the boundary value of the maximum rate of change as the final rate of change index. The boundary value is set according to the physical characteristics of the component to ensure that the index is within a reasonable physical range. The boundary values ​​of different components are set independently.

[0107] To ensure that the rate of change index remains within a reasonable physical range, the calculated rate of change index is subject to range limitations. When the absolute value of the index exceeds the maximum allowable rate of change for that component, the maximum rate of change boundary value is taken as the final rate of change index. The maximum rate of change boundary value is set based on the physical characteristics of the component, and the boundary values ​​for different components are set independently. For example, for the sand gradation adjustment component, based on its physical characteristics, the maximum rate of change boundary value is set to 3. If the calculated rate of change index is 4, then the final rate of change index is taken as 3. By limiting the range, unreasonable adjustments caused by excessively large rate of change in the index can be avoided.

[0108] Step S430: Adjust the corresponding proportional adjustment coefficient components according to the change rate index of each component. The change rate index of the sand particle size distribution adjustment component corresponds to the sand particle size distribution ratio coefficient, the change rate index of the cement dosage adjustment component corresponds to the cement dosage ratio coefficient, and the change rate index of the water-cement ratio adjustment component corresponds to the water-cement ratio ratio coefficient. The adjustment direction is consistent with the sign of the change rate index, and the adjustment range is proportional to the absolute value of the index.

[0109] Based on the rate of change index for each component, adjust the corresponding proportional adjustment coefficient. The rate of change index for the sand gradation adjustment component corresponds to the sand gradation proportional coefficient, the rate of change index for the cement dosage adjustment component corresponds to the cement dosage proportional coefficient, and the rate of change index for the water-cement ratio adjustment component corresponds to the water-cement ratio proportional coefficient. The adjustment direction is consistent with the sign of the rate of change index; that is, when the rate of change index is positive, increase the proportional coefficient; when the rate of change index is negative, decrease the proportional coefficient. The adjustment magnitude is directly proportional to the absolute value of the index; the larger the absolute value of the index, the larger the adjustment magnitude.

[0110] Step S440: Calculate the cumulative deviation index separately for each component adjustment sequence. The cumulative deviation index is the algebraic sum of the adjustment amounts of the component in the current period and the previous two periods, reflecting the cumulative deviation trend of the component. The cumulative deviation index of each component is calculated independently, and the statistical scope is limited to the continuous period data of the same component.

[0111] The cumulative deviation index reflects the cumulative deviation trend of each component. It is the algebraic sum of the adjustment amounts for the current period and the two previous periods for that component. The cumulative deviation index for each component is calculated independently, only counting continuous period data for the same component, ensuring that each index accurately reflects the cumulative deviation of that component. For example, for the sand gradation adjustment component, if the current period adjustment amount is 2, the previous period adjustment amount is -1, and the two previous periods adjustment amounts are 3, then the cumulative deviation index is 2 + (-1) + 3 = 4. By calculating the cumulative deviation index, the cumulative deviation of each component can be understood, providing a basis for adjusting the integral adjustment coefficient.

[0112] Step S450: Adjust the corresponding integral adjustment coefficient component based on the cumulative deviation index of each component. When the absolute value of the cumulative deviation index of a component is greater than the preset threshold of the component, increase the integral adjustment coefficient of the component; otherwise, decrease it. The adjustment direction is consistent with the sign of the cumulative deviation index. The preset threshold of different components is set separately according to their physical characteristics.

[0113] Based on the cumulative deviation index of each component, the corresponding integral control coefficient is adjusted. When the absolute value of the cumulative deviation index of a component exceeds the preset threshold for that component, it indicates a significant steady-state error, requiring an increase in the integral control coefficient to eliminate the error; otherwise, the integral control coefficient is decreased. The adjustment direction is consistent with the sign of the cumulative deviation index to ensure correct integral control. Preset thresholds for different components are set individually based on their physical characteristics to accommodate the control requirements of different components. For example, for the sand gradation control component, the preset threshold is 3. If the cumulative deviation index is 4, exceeding the threshold, the sand gradation integral control coefficient is increased; if the cumulative deviation index is 2, less than the threshold, the sand gradation integral control coefficient is decreased. In this way, the integral control coefficient is dynamically adjusted based on the cumulative deviation of each component, improving the control accuracy of the PID controller.

[0114] Step S460: Calculate the second-order difference value for each component adjustment sequence separately. The second-order difference value is obtained by subtracting the first-order difference value between the previous period and the two previous periods from the first-order difference value between the current period and the previous period. It reflects the change in the rate of change of the component adjustment amount. Adjust the differential adjustment coefficient of each component based on the second-order difference value. The adjustment magnitude is proportional to the absolute value of the second-order difference value, and the direction is consistent with the sign of the second-order difference value.

[0115] The second-order difference value reflects the change rate of each component's adjustment amount. It is obtained by subtracting the first-order difference between the previous period and the two periods prior from the first-order difference between the current period and the previous period. The first-order difference value is the difference between component values ​​in adjacent periods. For example, for the sand gradation adjustment component, if the current period's component value is 10, the previous period's component value is 8, and the two periods prior's component value is 6, then the first-order difference between the current period and the previous period is 10-8=2, the first-order difference between the previous period and the two periods prior is 8-6=2, and the second-order difference value is 2-2=0. The derivative adjustment coefficients of each component are adjusted based on the second-order difference value. The adjustment magnitude is proportional to the absolute value of the second-order difference value, and the direction is consistent with the sign of the second-order difference value. When the second-order difference value is positive, the derivative adjustment coefficient is increased; when the second-order difference value is negative, the derivative adjustment coefficient is decreased. In this way, the derivative adjustment coefficient is dynamically adjusted according to the change rate of each component's adjustment amount, enabling the PID controller to better predict the trend of deviation changes.

[0116] In some embodiments, step S460, calculating the second-order difference value for each component adjustment sequence individually, may specifically include the following steps S461~S467: Step S461: Select the target component from the component adjustment sequence. The target component is the component whose second-order difference value is to be calculated. Process it in the order of sand particle size distribution, cement dosage, and water-cement ratio, which is consistent with the calculation order of the rate of change index, to ensure the continuity of the processing logic.

[0117] To calculate the second-order difference value of each component adjustment sequence sequentially, a target component is selected from the component adjustment sequence. The process follows the order of sand particle size distribution, cement dosage, and water-cement ratio, consistent with the calculation order of the rate of change index, thus ensuring the continuity of the processing logic. For example, first, the sand particle size distribution adjustment component is selected as the target component, and its second-order difference value is calculated. Then, the cement dosage adjustment component is selected, and finally, the water-cement ratio adjustment component is selected. This sequential processing allows for the systematic calculation of the second-order difference value for each component.

[0118] Step S462: Based on the component adjustment sequence of the target component, obtain the component values ​​of the previous two periods, the component value of the previous period, and the component value of the current period. The unit of the component value is consistent with the unit of the adjustment amount of the component. The data acquisition range is limited to the historical adjustment data of the target component.

[0119] After selecting the target component, based on its component adjustment sequence, obtain the component values ​​for the previous two periods, the previous period, and the current period. The unit of the component value must be consistent with the unit of the adjustment amount for that component to ensure data dimension uniformity. The data acquisition scope is limited to the historical adjustment data of the target component to avoid introducing data interference from other components. For example, for the sand grain gradation adjustment component, if its adjustment amount is in percentage, then the component values ​​in the obtained component adjustment sequence should also be percentages.

[0120] Step S463: Calculate the difference between the previous period component value and the previous two period component values ​​of the target component, and use it as the first-order difference value to reflect the rate of change of the target component in the earlier period segment. The sign of the difference value indicates the direction of change.

[0121] The first-order difference value is the difference between the target component's value in the previous period and the value in the two periods prior, reflecting the rate of change of the target component in the earlier period. The sign of the difference indicates the direction of change, with positive values ​​indicating increase and negative values ​​indicating decrease.

[0122] Step S464: Calculate the difference between the current period component value and the previous period component value of the target component, and use it as the second-order difference value to reflect the rate of change of the target component in the most recent period segment. The sign of the difference value indicates the direction of change.

[0123] The second-order difference value is the difference between the target component's value in the current period and the value in the previous period, reflecting the rate of change of the target component in the most recent period. The sign of the difference also indicates the direction of change, with positive values ​​indicating increase and negative values ​​indicating decrease.

[0124] Step S465: Calculate the second-order difference value, which is the second-order difference value minus the first-order difference value. This reflects the change in the rate of change of the target component. A positive second-order difference value indicates that the rate of change is increasing, while a negative value indicates that the rate of change is decreasing. The magnitude of the value reflects the degree of change.

[0125] After obtaining the first and second order difference values, the difference between them is calculated as the second order difference value. This second order difference value reflects the change in the rate of change of the target component. If the second order difference value is positive, it indicates that the rate of change of the target component is increasing, meaning that the adjustment amount is changing faster and faster; if the second order difference value is negative, it indicates that the rate of change is decreasing, meaning that the change in the adjustment amount is gradually slowing down. The magnitude of the value reflects the degree of change; the larger the value, the more drastic the change. By calculating the second order difference value, a deeper understanding of the dynamic changes in the adjustment amount of the target component can be obtained.

[0126] Step S466: Take the absolute value of the second-order difference to obtain the absolute value of the second-order difference. The magnitude of this value reflects the degree of change of the target component. The larger the absolute value of the second-order difference, the more drastic the change, and the greater the adjustment range of the corresponding differential adjustment coefficient.

[0127] The absolute value of the second-order difference is the absolute value of the second-order difference, and it more directly reflects the degree of drastic change in the target component. When the absolute value of the second-order difference is large, it indicates that the rate of change of the target component is relatively drastic. In this case, a larger adjustment of the derivative control coefficient is needed so that the PID controller can respond to this drastic change more promptly and better predict the trend of the deviation. Conversely, if the absolute value of the second-order difference is small, it indicates that the change is relatively gradual, and the adjustment range of the derivative control coefficient can be reduced accordingly. By taking the absolute value of the second-order difference, the degree of drastic change in the target component can be accurately measured, thereby allowing for reasonable adjustment of the derivative control coefficient.

[0128] Step S467: Determine the adjustment direction of the differential adjustment coefficient based on the sign of the second-order difference value. When the second-order difference value is positive, the adjustment direction is to increase the differential adjustment coefficient. When the second-order difference value is negative, the adjustment direction is to decrease the differential adjustment coefficient. The adjustment direction is strictly consistent with the sign of the second-order difference value to ensure that the adjustment logic matches the trend of change.

[0129] The sign of the second-order difference value clearly defines the direction of change of the target component's rate of change. Using this sign to determine the adjustment direction of the derivative adjustment coefficient ensures that the PID controller's control logic matches the changing trend of the target component. When the second-order difference value is positive, it means the rate of change of the target component is increasing. To better cope with this change, the derivative adjustment coefficient needs to be increased to react to rapid changes in deviation in advance. When the second-order difference value is negative, it indicates the rate of change of the target component is decreasing. In this case, the derivative adjustment coefficient should be decreased to avoid over-adjustment. For example, for the water-ash ratio control component, if the second-order difference value is positive, it means the rate of change of the water-ash ratio control is accelerating, and the derivative adjustment coefficient for the water-ash ratio should be increased; if the second-order difference value is negative, the derivative adjustment coefficient for the water-ash ratio should be decreased. By strictly adjusting the derivative adjustment coefficient according to the sign of the second-order difference value, the control effect of the PID controller can be improved.

[0130] Step S470: Combine the adjusted proportional adjustment coefficient, integral adjustment coefficient, and derivative adjustment coefficient components in the original dimensional order to generate a ratio optimization instruction containing the adjustment direction and amplitude of each component. Each parameter in the instruction corresponds one-to-one with the component of the deviation adjustment amount to ensure that the adjustment parameters match the component type.

[0131] After adjusting the proportional, integral, and derivative adjustment coefficients of each component, these adjusted coefficients are combined in their original dimensional order to obtain the mix design optimization instruction. This instruction contains the adjustment direction and magnitude information for each component, and each parameter corresponds one-to-one with the deviation adjustment amount component, thus ensuring accurate matching between the adjustment parameters and component types. For example, the adjustment coefficient for the sand gradation adjustment component has a specific corresponding position in the instruction, as do the cement dosage and water-cement ratio adjustment components. The mix design optimization instruction generated in this way can clearly and accurately guide the feeding control of the manufactured sand mixing equipment, ensuring that the sand gradation, cement dosage, and water-cement ratio are adjusted in the appropriate direction and magnitude to achieve the purpose of optimizing the mix design.

[0132] Step S500: Based on the mix proportion optimization instruction, send a parameter correction signal to the feed control module of the manufactured sand mixing equipment to correct the feed flow parameters of manufactured sand, stone powder and water, so that the deviation between the corrected mix proportion parameters and the target value of concrete compressive strength is controlled within the preset range.

[0133] The mix design optimization command provides a basis for correcting the feed flow parameters of manufactured sand, stone powder, and water. Based on this command, a parameter correction signal is sent to the feed control module of the manufactured sand mixing equipment. Upon receiving the signal, the feed control module adjusts the feed flow parameters of manufactured sand, stone powder, and water accordingly. This adjustment brings the mix proportions of manufactured sand, stone powder, and water closer to the preset target value for concrete compressive strength, keeping deviations within the allowable preset range. For example, when the mix design optimization command requires increasing the sand particle size distribution, the feed control module will correspondingly increase the feed flow of manufactured sand; if it requires adjusting the water-cement ratio, it will adjust the feed flow of water. By continuously adjusting the feed flow parameters, the mix design is gradually optimized to ensure that the produced concrete meets the target compressive strength requirements.

[0134] In some embodiments, step S500 may specifically include the following steps S510~S590: Step S510: Analyze the adjustment direction and amplitude information in the proportion optimization instruction, and determine the adjustment requirements corresponding to the feed flow parameters of manufactured sand, stone powder and water. The adjustment requirements include the adjustment direction and amplitude ratio.

[0135] The proportioning optimization command contains information on the adjustment direction and magnitude of manufactured sand, stone powder, and water. This information needs to be analyzed to clarify the specific adjustment requirements for each feed flow rate parameter. The adjustment direction is either increase or decrease, and the magnitude indicates the degree of adjustment required. For example, analyzing the command reveals that the feed flow rate of manufactured sand needs to be increased by a certain percentage, while the feed flow rate of stone powder needs to be decreased by a certain percentage.

[0136] Step S520: Based on the adjustment requirements and the correspondence between each feed component and the proportion state vector, the proportion optimization command is converted into the flow correction coefficient of each component, the sand particle size distribution adjustment range corresponding to manufactured sand, the cement dosage adjustment range corresponding to stone powder, and the water-cement ratio adjustment range corresponding to water.

[0137] After determining the adjustment requirements for each feed flow rate parameter, the proportioning optimization command is converted into flow rate correction coefficients for each component based on the correspondence between the feed components and the proportioning state vector. Since manufactured sand is related to sand particle size distribution, stone powder is related to cement dosage, and water is related to the water-cement ratio, the adjustment range can be directly converted into flow rate correction coefficients based on these correspondences. For example, if the adjustment range for sand particle size distribution in the proportioning optimization command is to increase, then the flow rate correction coefficient for manufactured sand will also be positive, indicating that the feed flow rate of manufactured sand needs to be increased. Similarly, the flow rate correction coefficients for stone powder and water are determined based on the adjustment ranges for cement dosage and water-cement ratio. Through this conversion, abstract adjustment commands can be transformed into specific flow rate correction coefficients, facilitating subsequent calculations of feed flow rate parameters.

[0138] In some embodiments, step S520 may specifically include the following steps S521 to S526: Step S521: Determine the correspondence rules between each component of the mix proportion state vector and the feed components by querying the pre-stored mix proportion-component association table. The mix proportion-component association table records the mapping relationship between the sand particle size distribution characteristic component and the amount of manufactured sand feed, the cement dosage characteristic component and the amount of stone powder feed, and the water-cement ratio characteristic component and the amount of water feed.

[0139] The pre-stored mix proportion-component correlation table is an important data reference, clearly defining the correspondence between each component of the mix proportion state vector and the feed components. By querying this correlation table, the feed components corresponding to the sand particle size distribution characteristic component, cement content characteristic component, and water-cement ratio characteristic component can be accurately located, namely manufactured sand, stone powder, and water. For example, the correlation table clearly shows the correspondence between the sand particle size distribution characteristic component and the manufactured sand feed amount, providing an accurate basis for subsequently converting the adjustment range into a flow correction coefficient. By querying the correlation table, the accuracy of the correspondence between each component and the feed components can be ensured during the conversion process.

[0140] Step S522: Extract the sand particle size distribution adjustment range, cement dosage adjustment range, and water-cement ratio adjustment range from the mix design optimization command. Each adjustment range is the relative change ratio of the corresponding characteristic component.

[0141] The mix design optimization instructions contain information on the adjustment ranges of sand particle size distribution, cement dosage, and water-cement ratio, which need to be extracted from the instructions. These adjustment ranges are presented as the relative change ratios of the corresponding characteristic components. By accurately extracting this adjustment range information, specific data can be provided for subsequent calculation of the flow correction coefficient.

[0142] Step S523: The adjustment range of sand particle size distribution is directly used as the flow correction coefficient of manufactured sand, the adjustment range of cement dosage is directly used as the flow correction coefficient of stone powder, and the adjustment range of water-cement ratio is directly used as the flow correction coefficient of water. The sign of the correction coefficient is consistent with the sign of the adjustment range.

[0143] Based on the correspondence determined by the proportion-component correlation table, the extracted adjustment range is directly converted into the flow correction coefficient for each feed component. Since sand particle size distribution corresponds to manufactured sand, cement dosage corresponds to stone powder, and water-cement ratio corresponds to water, the adjustment range can be directly used as the corresponding flow correction coefficient. Furthermore, the sign of the correction coefficient is consistent with the sign of the adjustment range; that is, when the adjustment range increases, the correction coefficient is positive; when the adjustment range decreases, the correction coefficient is negative.

[0144] Step S524: Perform sign verification on the flow correction coefficients of each component. When the adjustment requirement is to increase the feed flow rate, the correction coefficient is positive; when the adjustment requirement is to decrease the feed flow rate, the correction coefficient is negative, ensuring that the sign is consistent with the adjustment direction.

[0145] To ensure that the sign of the flow correction coefficient is consistent with the adjustment direction, the sign of the flow correction coefficient for each component needs to be checked. If the adjustment requirement is to increase the feed flow rate, the corresponding flow correction coefficient should be positive; if the adjustment requirement is to decrease the feed flow rate, the correction coefficient should be negative. For example, if the proportioning optimization instruction requires increasing the feed flow rate of manufactured sand, but the calculated flow correction coefficient for manufactured sand is negative, it indicates an error in the sign, which needs to be corrected. By checking the sign, errors in the feed flow rate adjustment direction due to incorrect sign can be avoided, ensuring the accuracy of the adjustment.

[0146] Step S525: Based on the physical properties of manufactured sand, stone powder and water, limit the range of values ​​for the flow correction coefficients of each component. The absolute value of the correction coefficient shall not exceed the maximum allowable adjustment ratio of the equipment, and any excess shall be treated as boundary values.

[0147] The physical properties of each feed component and the adjustment capability of the equipment determine the range of values ​​for the flow correction coefficient. To ensure the normal operation of the equipment and the rationality of its adjustment, the range of values ​​for the flow correction coefficient needs to be limited based on the physical properties of the manufactured sand, stone powder, and water. The absolute value of the correction coefficient cannot exceed the maximum allowable adjustment ratio of the equipment; if it exceeds this range, the excess portion will be treated as a boundary value.

[0148] Step S526: Generate a set of flow correction coefficients that includes correction coefficients for manufactured sand, stone powder, and water. Each element in the set corresponds one-to-one with the feed components and is used for subsequent calculation of preliminary correction flow parameters.

[0149] After calculating and verifying the flow correction coefficients for each component, the correction coefficients for manufactured sand, stone powder, and water are combined into a flow correction coefficient set. Each element in the set corresponds one-to-one with the corresponding feed component. This allows for convenient adjustment of the feed flow rate based on the coefficients in this set when calculating the initial corrected flow parameters. For example, in the initial corrected flow rate calculation for computer-controlled sand making, the manufactured sand correction coefficient in the set can be used directly. By generating the flow correction coefficient set, the flow correction coefficients for each feed component can be managed uniformly, facilitating subsequent calculations.

[0150] Step S530: Obtain the current feed flow parameters of the feed control module, including the current flow rate of manufactured sand, the current flow rate of stone powder and the current flow rate of water. The current flow parameters are collected through the real-time feedback signal of the feed control module.

[0151] Real-time feedback signals from the feed control module are crucial for obtaining the current feed flow rate parameters. By collecting these feedback signals, the current feed flow rates of manufactured sand, stone powder, and water can be obtained. These current flow rate parameters form the basis for subsequent calculations and preliminary corrections of the flow rate parameters. For example, the feed control module may monitor the flow rates of manufactured sand, stone powder, and water in real time through sensors and transmit this data in the form of feedback signals. By processing and analyzing these feedback signals, the current feed flow rate parameters can be accurately obtained.

[0152] Step S540: Generate preliminary corrected flow parameters by combining the current flow parameters and the flow correction coefficient. The generation method is to adjust the current flow parameters according to the proportion of the flow correction coefficient, and the adjustment range is a multiple of the flow correction coefficient of the current flow parameters.

[0153] After obtaining the current flow rate parameters and the flow rate correction factor, the two are combined to generate preliminary corrected flow rate parameters. Specifically, the current flow rate parameters are adjusted proportionally to the flow rate correction factor, with the adjustment amount being the current flow rate parameter multiplied by the flow rate correction factor. For example, if the current flow rate of manufactured sand is a fixed value and the flow rate correction factor for manufactured sand is +10%, then the preliminary corrected flow rate for manufactured sand is the current flow rate multiplied by (1 + 10%). In this way, the current feed flow rate can be initially adjusted based on the flow rate correction factor to obtain preliminary corrected flow rate parameters, providing a basis for subsequent compensation processing.

[0154] Step S550: Call the pre-built coupling effect compensation model to compensate for the preliminary corrected flow parameters. The coupling effect compensation model is constructed in the following way: using the manufactured sand flow rate, stone powder flow rate, water flow rate and corresponding mixing uniformity test results in the historical mix ratio correction data as training samples, the ratio of sand particle size distribution to stone powder dosage as input features, the correlation between water-cement ratio and mixing uniformity as label values, and the generation of model parameters through gradient descent algorithm.

[0155] The coupling effect compensation model is constructed to consider the interactions between manufactured sand, stone powder, and water. It uses the flow rates of manufactured sand, stone powder, and water from historical mix design data, along with the corresponding mixing uniformity test results, as training samples. When constructing the model, the ratio of sand particle size distribution to stone powder dosage is used as an input feature, as this ratio affects the mixing effect. The correlation between the water-cement ratio and mixing uniformity is used as a label value, reflecting the important role of the water-cement ratio in the mixing process. The model parameters are continuously optimized using a gradient descent algorithm, enabling the model to accurately predict and compensate for the coupling effects between the various feed components. For example, adjusting the flow rates of manufactured sand and stone powder may affect the effect of water; the coupling effect compensation model can consider this effect and provide corresponding compensation. After obtaining the initial corrected flow rate parameters, this pre-built model is used for compensation processing to improve the mixing uniformity and the quality of the final product.

[0156] In some embodiments, step S550 may specifically include the following steps S551 to S556: Step S551: Receive the preliminary flow rates of manufactured sand, stone powder, and water in the preliminary corrected flow parameters through the coupling effect compensation model. The input data format is consistent with the input format during model training, and the flow rate unit is converted to the standard unit required by the model.

[0157] The coupling effect compensation model requires the initial flow rates of manufactured sand, stone powder, and water from the preliminary corrected flow parameters as input. To ensure the model can correctly process this data, the input data format must be consistent with the input format used during model training, and the flow units need to be converted to the standard units required by the model. For example, if the flow unit used during model training is cubic meters per second, while the actual collected flow unit may be liters per minute, then the units of the preliminary corrected flow parameters need to be converted.

[0158] Step S552: The first processing layer inside the compensation model is compensated by coupling effect. Based on the gradation coupling law, the initial flow rate of manufactured sand and the initial flow rate of stone powder are correlated and calculated. The gradation coupling law is obtained by statistical analysis of historical mixing data. It is manifested as the influence of the ratio change of manufactured sand flow rate and stone powder flow rate on the gradation complementarity. The calculation result is used as the intermediate feature vector.

[0159] The first processing layer of the coupling effect compensation model is responsible for calculating the correlation between the initial flow rates of manufactured sand and stone powder. The gradation coupling law is derived from historical mixing data statistics, reflecting the impact of changes in the ratio of manufactured sand flow rate to stone powder flow rate on gradation complementarity. For example, when the ratio of manufactured sand flow rate to stone powder flow rate is within a certain range, the gradation complementarity is good, which can improve the mixing effect. In the first processing layer, the initial flow rates of manufactured sand and stone powder are calculated according to this gradation coupling law, and the results are used as intermediate feature vectors. These intermediate feature vectors contain the correlation information between the manufactured sand and stone powder flow rates, providing important intermediate data for subsequent further processing.

[0160] Step S553: ​​The second processing layer inside the compensation model is used to calculate the cross-influence of the initial water flow rate and the intermediate feature vector based on the correlation law. The correlation law is obtained by fitting experimental data and is shown as the influence trend of water-cement ratio and mixing uniformity. The calculation result is used as the basic value of the compensation coefficient.

[0161] The second treatment layer of the coupling effect compensation model calculates the cross-influence of the initial water flow rate and the intermediate eigenvector. The correlation law is obtained through fitting experimental data and reflects the influence trend between the water-cement ratio and the mixing uniformity. For example, changes in the water-cement ratio will affect the mixing uniformity of manufactured sand and stone powder. In the second treatment layer, based on this correlation law, the initial water flow rate and the intermediate eigenvector are calculated, and the result is used as the base value of the compensation coefficient.

[0162] Step S554: The basic values ​​are dynamically adjusted through the output layer of the coupling effect compensation model. The adjustment is based on the deviation between the current proportion state vector and the historical optimal proportion state vector. The larger the deviation, the larger the adjustment range. Finally, the coupling compensation coefficients of each component are output.

[0163] The output layer of the coupling effect compensation model dynamically adjusts the base values ​​of the compensation coefficients. The adjustment is based on the deviation between the current mix proportion state vector and the historical optimal mix proportion state vector. The historical optimal mix proportion state vector represents the mix proportion state that achieved good mixing results in previous production. When the deviation between the current mix proportion state vector and the historical optimal mix proportion state vector is large, it indicates a significant difference between the current mix proportion state and the optimal state, requiring a larger adjustment of the compensation coefficients; conversely, the adjustment can be smaller. Through this dynamic adjustment, the coupling compensation coefficients of each component are ultimately output. These coefficients can more accurately compensate for the coupling effect between the feed components, improving the mixing quality.

[0164] Step S555: Compare the output coupling compensation coefficient with the preset compensation coefficient threshold. If the absolute value of the compensation coefficient exceeds the threshold, it is truncated according to the threshold. The threshold is set according to the equipment adjustment accuracy so that the compensated flow parameters are within the controllable range of the equipment.

[0165] To ensure that the compensated flow parameters remain within the controllable range of the equipment, the output coupling compensation coefficient needs to be compared with a preset compensation coefficient threshold. The compensation coefficient threshold is set based on the equipment's adjustment precision, defining the maximum allowable range for the compensation coefficient. If the absolute value of the coupling compensation coefficient exceeds the threshold, it is truncated according to the threshold. This method avoids situations where the equipment cannot accurately adjust the flow parameters due to an excessively large compensation coefficient, ensuring the normal operation of the equipment and the effectiveness of its adjustment.

[0166] Step S556: Output the truncated coupling compensation coefficient, which serves as the basis for generating the compensated flow parameters and corresponds one-to-one with the preliminary corrected flow parameters of manufactured sand, stone powder and water.

[0167] After completing the threshold comparison and truncation of the coupling compensation coefficients, the truncated coupling compensation coefficients are output. These coefficients correspond one-to-one with the preliminary corrected flow parameters of manufactured sand, stone powder, and water, and they will serve as an important basis for generating subsequent compensated flow parameters. For example, the coupling compensation coefficient of manufactured sand will be used to compensate for the preliminary corrected flow rate of manufactured sand, and the same applies to stone powder and water. By outputting the truncated coupling compensation coefficients, it can be ensured that the subsequently generated compensated flow parameters take into account both the coupling effect between the various feed components and remain within the controllable range of the equipment, thereby improving the quality and stability of manufactured sand mixing.

[0168] Step S560: Input the preliminary flow rates of manufactured sand, stone powder, and water in the preliminary corrected flow parameters into the coupling effect compensation model, and output the coupling compensation coefficients of each component. The magnitude of the coupling compensation coefficients is positively correlated with the strength of gradation complementarity and the degree of cross-influence, and the sign indicates the compensation direction.

[0169] The initial flow rates of manufactured sand, stone powder, and water from the preliminary corrected flow parameters are input again into the coupling effect compensation model. After processing by the model, the coupling compensation coefficients of each component are output. The magnitude of the coupling compensation coefficient is positively correlated with the strength of gradation complementarity and the degree of cross-influence; that is, the stronger the gradation complementarity and the greater the degree of cross-influence, the larger the value of the coupling compensation coefficient; conversely, the smaller the value. The sign indicates the direction of compensation: a positive sign indicates that the flow rate needs to be increased, and a negative sign indicates that the flow rate needs to be decreased. For example, if the gradation complementarity of manufactured sand and stone powder is good, and the cross-influence with water is large, then the output coupling compensation coefficient may be large. By outputting the coupling compensation coefficient, the preliminary corrected flow parameters can be accurately compensated, taking into account the mutual influence between the various feed components.

[0170] Step S570: Combine the preliminary corrected flow parameters and the coupling compensation coefficient to generate the compensated flow parameters. The generation method is to make additional adjustments to the preliminary corrected flow parameters according to the ratio of the coupling compensation coefficient, and the adjustment direction is consistent with the sign of the coupling compensation coefficient.

[0171] After obtaining the coupling compensation coefficient, it is combined with the preliminary corrected flow parameters to generate the compensated flow parameters. Specifically, the preliminary corrected flow parameters are adjusted proportionally to the coupling compensation coefficient, with the adjustment direction consistent with the sign of the coefficient. This method further adjusts the preliminary corrected flow parameters, taking into account the coupling effect between the various feed components, resulting in more reasonable final feed flow parameters and improving the quality and effectiveness of manufactured sand mixing.

[0172] Step S580: Perform a mix proportion balance verification on the compensated flow parameters. The verification process involves substituting the sand particle size distribution characteristic component, cement dosage characteristic component, and water-cement ratio characteristic component into the pre-constructed mix proportion balance function. If the function output value is within the preset balance threshold range, the verification passes.

[0173] The pre-constructed mix proportioning equilibrium function is used to evaluate the balance relationship between sand particle size distribution, cement content, and water-cement ratio. Substituting the characteristic components of sand particle size distribution, cement content, and water-cement ratio into this function yields an output value. The preset equilibrium threshold range is determined based on experience and experiments. When the function output value falls within this range, it indicates that the current mix proportion is balanced and the verification passes; otherwise, further adjustments to the flow parameters are needed. For example, if the mix proportioning equilibrium function considers the interrelationship between sand particle size distribution, cement content, and water-cement ratio, and the output value obtained after substituting these characteristic components into the function falls within the preset equilibrium threshold range, it indicates that the feed flow ratio of manufactured sand, stone powder, and water is reasonable and can ensure a good mixing effect. Through mix proportioning balance verification, it can be ensured that the final feed flow parameters meet the requirements of mix proportioning balance.

[0174] Step S590: Convert the verified compensated flow parameters into a parameter correction signal in a standard communication protocol format and send it to the feed control module via the industrial bus. The correction signal includes the target flow value and adjustment time window information.

[0175] After the compensated flow parameters pass the mix balance verification, they are converted into parameter correction signals in a standard communication protocol format. This standard protocol format ensures accurate signal transmission over the industrial bus and proper reception and processing by the feed control module. The parameter correction signal contains the target flow value and adjustment time window information. The target flow value specifies the required flow rates for manufactured sand, stone powder, and water, while the adjustment time window specifies the timeframe for completing the flow adjustment. This parameter correction signal is sent to the feed control module via the industrial bus. The feed control module adjusts the feed flow rates of manufactured sand, stone powder, and water based on the information in the signal, ultimately optimizing the manufactured sand mix and ensuring that the produced concrete meets the target compressive strength requirements.

[0176] It is understood that the various algorithms involved in the above descriptions of the embodiments of the present invention can all be obtained from relevant content in the prior art. To save space, they will not be elaborated on in the embodiments of the present invention. In addition, those skilled in the art can supplement the details based on common knowledge in the art when implementing the solutions of the present invention. For example, they can use normalization to eliminate dimensional conflicts before feature fusion, use interpolation to eliminate dimensional differences, reasonably set thresholds based on historical data, experience or business scenario requirements, train the model based on a general model training method, set the number of layers in the model structure based on actual needs, select activation functions, etc. The present invention will not provide redundant descriptions of overly detailed implementation processes here.

[0177] Figure 2 This is a schematic diagram of the composition structure of a manufactured sand proportioning control device provided in an embodiment of the present invention, as shown below. Figure 2 As shown, the manufactured sand proportioning control device 200 includes: The data acquisition module 210 is used to acquire the mixing process response data of the manufactured sand mixing equipment during a continuous operating cycle. The mixing process response data is acquired by a vibration sensor installed at the bottom of the mixing cylinder and includes vibration waveform sequences under different ratio combinations and corresponding mixing timestamps. The mixing timestamps are used to mark the mixing stage corresponding to each vibration waveform sequence. Vector generation module 220 is used to perform proportion state decoding processing on the mixing process response data to generate a proportion state vector characterizing the mixing uniformity under the current proportion. The dimension of the proportion state vector corresponds to the combination parameters of sand particle size distribution, cement dosage and water-cement ratio in the mixing process. The deviation calculation module 230 is used to perform mapping analysis between the mix proportion state vector and the preset concrete compressive strength target value, and calculate the deviation adjustment amount between the current mix proportion state and the target strength requirement. The deviation adjustment amount is generated by comparing the difference between the features of each dimension in the mix proportion state vector and the target strength feature library. The instruction generation module 240 is used to perform dynamic tuning of PID parameters based on the deviation adjustment amount, adjust the proportional adjustment coefficient, integral adjustment coefficient and derivative adjustment coefficient according to the changing trend of the deviation adjustment amount, and generate a ratio optimization instruction containing the adjustment direction and amplitude. The parameter correction module 250 is used to send a parameter correction signal to the feed control module of the manufactured sand mixing equipment according to the proportion optimization instruction, and correct the feed flow parameters of manufactured sand, stone powder and water, so that the deviation between the corrected proportion parameters and the target value of concrete compressive strength is controlled within a preset range.

[0178] The descriptions of the apparatus embodiments above are similar to those of the method embodiments above, and have similar beneficial effects. In some embodiments, the functions or modules included in the apparatus provided by the present invention can be used to perform the methods described in the method embodiments above. For technical details not disclosed in the apparatus embodiments of the present invention, please refer to the descriptions of the method embodiments of the present invention for understanding.

[0179] Figure 3 A hardware entity diagram of a computer system provided as an embodiment of the present invention, such as... Figure 3 As shown, the hardware entity of the computer system 1000 includes a processor 1001 and a memory 1002, wherein the memory 1002 stores a computer program that can run on the processor 1001, and the processor 1001 executes the program to implement the steps in the method of any of the above embodiments.

[0180] The memory 1002 stores computer programs that can run on the processor. The memory 1002 is configured to store instructions and applications that can be executed by the processor 1001. It can also cache data to be processed or already processed (e.g., image data, audio data, voice communication data, and video communication data) of the processor 1001 and various modules in the computer system 1000. It can be implemented by flash memory or random access memory (RAM).

[0181] When the processor 1001 executes the program, it implements the steps of the PID-based controlled sand proportioning method described above. The processor 1001 typically controls the overall operation of the computer system 1000.

Claims

1. A method for controlling the proportion of machine-made sand based on PID adjustment, characterized in that, The method includes: The mixing process response data of the manufactured sand mixing equipment during a continuous operating cycle is obtained. The mixing process response data is collected by a vibration sensor installed at the bottom of the mixing cylinder and includes vibration waveform sequences under different ratio combinations and corresponding mixing timestamps. The mixing timestamps are used to mark the mixing stage corresponding to each vibration waveform sequence. The mixing process response data is processed by ratio state decoding to generate a ratio state vector characterizing the mixing uniformity under the current ratio. The dimension of the ratio state vector corresponds to the combination parameters of sand particle size distribution, cement dosage and water-cement ratio in the mixing process. The mix proportion state vector is mapped and analyzed with the preset concrete compressive strength target value to calculate the deviation adjustment amount between the current mix proportion state and the target strength requirement. The deviation adjustment amount is generated by comparing the difference between the features of each dimension in the mix proportion state vector and the target strength feature library. Based on the deviation adjustment amount, perform dynamic tuning of PID parameters, adjust the proportional adjustment coefficient, integral adjustment coefficient and derivative adjustment coefficient according to the changing trend of the deviation adjustment amount, and generate a ratio optimization instruction containing the adjustment direction and amplitude. According to the proportion optimization instruction, a parameter correction signal is sent to the feeding control module of the manufactured sand mixing equipment to correct the feeding flow parameters of manufactured sand, stone powder and water, so that the deviation between the corrected proportion parameters and the target value of concrete compressive strength is controlled within a preset range.

2. The method of claim 1, wherein, The step of decoding the mixing process response data to generate a mixing state vector characterizing the mixing uniformity under the current mixing ratio includes: Based on the mixing timestamp, the vibration waveform sequence is divided into sub-waveform sequences corresponding to the mixing stages. The duration of each sub-waveform sequence is equal to the duration of the mixing stage, and the start timestamp of the sub-waveform sequence is aligned with the start timestamp of the mixing stage. Each sub-waveform sequence is subjected to Fourier transform processing to convert the time-domain vibration signal into a frequency-domain signal, generating a preliminary frequency spectrum containing multiple frequency components. The frequency resolution of the preliminary frequency spectrum is proportional to the length of the sub-waveform sequence. The preliminary frequency spectrum is subjected to frequency band screening to retain the target frequency bands corresponding to the particle size distribution range of sand. The upper and lower limits of the target frequency bands are set according to the standard particle size range of manufactured sand production. The filtered frequency domain signal is processed by inverse Fourier transform to reconstruct the time domain vibration signal and extract the amplitude change curve. The horizontal axis of the amplitude change curve is relative time, and the vertical axis is vibration amplitude. The amplitude unit corresponds to the range of the vibration sensor. The energy proportion index of each frequency band is calculated based on the frequency spectrum of the target frequency band. The characteristic components of sand particle size distribution are generated through the nonlinear mapping relationship between the energy proportion index and the sand particle size distribution parameters. The energy proportion index is the ratio of the energy value of each frequency band to the total energy value. Based on the peak point and decay rate of the amplitude variation curve, and combined with the influence of cement dosage and water-cement ratio on vibration damping characteristics, characteristic components of cement dosage and water-cement ratio are generated. The characteristic components of sand gradation, cement dosage, and water-cement ratio are then vector-joined in a preset dimensional order to generate a proportioning state vector associated with mixing uniformity.

3. The method of claim 2, wherein, The step of dividing the vibration waveform sequence into sub-waveform sequences corresponding one-to-one with the mixing stage based on the mixing timestamp includes: The stage identifier information in the mixing timestamp is analyzed to determine the dry mixing stage, wet mixing stage, and pre-discharge stage included in the mixing process. The time boundaries of each stage are divided by the start and end marks in the timestamp. According to the stage time boundary, continuous waveform data corresponding to the time period are extracted from the vibration waveform sequence to generate the same number of sub-waveform sequences as the mixing stage. The time axis of each sub-waveform sequence starts at the stage start timestamp and ends at the stage end timestamp. For each sub-waveform sequence, data integrity is checked to see if there is any sampling loss or abnormal jump in the waveform data. If there is an abnormality, the data interpolation method between the previous and next time points is used to repair it. The size of the interpolation window is inversely proportional to the sampling frequency. The repaired sub-waveform sequences are sorted according to the order of the mixing stages to generate a set of stage-ordered sub-waveform sequences. The order of each element in the set is consistent with the actual process of mixing. Add a stage attribute label to each sub-waveform sequence. The label content includes the stage name and duration. The stage attribute label is stored in association with the header metadata of the sub-waveform sequence. Establish a mapping table between sub-waveform sequences and mixing stages, and record the timestamp range, data length, and corresponding stage index of each sub-waveform sequence.

4. The method according to claim 2, characterized in that, The step of performing frequency band screening on the preliminary frequency spectrum to retain the target frequency band corresponding to the particle size distribution range of the sand includes: Obtain the standard particle size distribution range for manufactured sand production, and convert the particle size range into the corresponding vibration frequency range. The conversion is based on the empirical correlation between the collision vibration frequency of sand particles and the particle size. The larger the particle size, the lower the corresponding frequency. The passband parameters of the bandpass filter are set according to the vibration frequency range obtained by conversion. The lower limit frequency of the passband corresponds to the vibration frequency of the largest particle size, and the upper limit frequency corresponds to the vibration frequency of the smallest particle size. Bandpass filtering is performed based on Butterworth filters, and the order of the filters is set according to the noise suppression requirements. The preliminary frequency spectrum is input into the configured bandpass filter to filter each frequency component, retaining the frequency components within the passband and attenuating the frequency components outside the passband. Calculate the energy retention rate of the frequency spectrum before and after filtering. The energy retention rate is the ratio of the total energy after filtering to the total energy before filtering. If the retention rate is lower than the preset threshold, readjust the filter parameters. Output the filtered target frequency spectrum, which contains only the effective frequency components related to sand particle size distribution, and the bandwidth is proportional to the span of the standard particle size distribution range.

5. The method according to claim 1, characterized in that, The step of mapping the mix proportion state vector to a preset concrete compressive strength target value and calculating the deviation adjustment amount between the current mix proportion state and the target strength requirement includes: The pre-trained mix proportion-strength prediction model is invoked to perform feature space transformation on the mix proportion state vector to generate the predicted value of concrete compressive strength corresponding to the current mix proportion. The mix proportion-strength prediction model is obtained through training on historical mix proportion state vector samples and corresponding measured compressive strength value samples. The predicted value of concrete compressive strength is compared with the target value of concrete compressive strength, and the difference between the two is calculated as the basic value of strength deviation. When the predicted value is greater than the target value, the basic value of strength deviation is positive, and when the predicted value is less than the target value, the basic value of strength deviation is negative. Sensitivity analysis is performed on the characteristics of each dimension in the mix proportion state vector to determine the sensitivity coefficients of the sand particle size distribution characteristic component, cement content characteristic component and water-cement ratio characteristic component to the compressive strength of concrete. The sensitivity coefficients represent the amount of change in compressive strength of concrete caused by a unit change in characteristic component. Based on the basic value of the strength deviation and the sensitivity coefficient of each characteristic component, the deviation adjustment component corresponding to each characteristic component is calculated. The magnitude of the deviation adjustment component is the negative value of the ratio of the basic value of the strength deviation to the corresponding sensitivity coefficient. The directional consistency of each deviation adjustment component is checked to ensure that the adjustment directions of sand particle size distribution, cement dosage and water-cement ratio conform to the variation law of concrete strength. The verified deviation adjustment components are combined in order of dimension of the mix proportion state vector to generate a deviation adjustment amount that includes sand gradation adjustment component, cement dosage adjustment component and water-cement ratio adjustment component. The dimension of the deviation adjustment amount is the same as the dimension of the mix proportion state vector.

6. The method according to claim 5, characterized in that, The step of calling the pre-trained mix proportion-strength prediction model to perform feature space transformation processing on the mix proportion state vector to generate the predicted value of concrete compressive strength corresponding to the current mix proportion includes: The matching state vector is input into the input layer of the matching-intensity prediction model. The number of neurons in the input layer is equal to the dimension of the matching state vector, and each neuron receives a feature component of one dimension of the matching state vector. The input features are linearly transformed and nonlinearly activated through the first hidden layer of the ratio-intensity prediction model. The weight matrix of the linear transformation is obtained through optimization during the training process, and the ReLU function is used to enhance the nonlinear fitting ability of the model. The output features of the first hidden layer are input into the second hidden layer for secondary feature extraction and dimensionality compression. The number of neurons in the second hidden layer is less than that in the first hidden layer. Feature aggregation enhances the expressive power of key influencing factors. The output features of the second hidden layer are batch normalized to eliminate the dimensional differences between different feature dimensions. The normalized features are input into the third hidden layer, and the dropout operation in the third hidden layer is used to prevent overfitting. The dropout ratio is dynamically adjusted according to the number of training samples. The output features of the third hidden layer are linearly mapped through the output layer of the mix proportion-strength prediction model to generate a single numerical value of concrete compressive strength prediction. The bias term of the output layer is set according to the historical strength benchmark value.

7. The method according to claim 5, characterized in that, The sensitivity analysis of the characteristics of each dimension in the mix proportion state vector, determining the sensitivity coefficients of the sand particle size distribution characteristic component, cement content characteristic component, and water-cement ratio characteristic component to the compressive strength of concrete, includes: A historical mix design state vector sample set was selected. The sample set contains mix design state vectors under different mix combinations and corresponding measured values ​​of concrete compressive strength. The sample covers the entire adjustable range of mix parameters. For each mix proportion state vector sample in the sample set, the sand particle size distribution characteristic component, cement content characteristic component, and water-cement ratio characteristic component are adjusted individually in sequence. One component is adjusted each time while the other components remain unchanged. The adjustment range is a fixed proportion of the adjustable range of that component. Calculate the change in the measured value of concrete compressive strength before and after adjustment, and use the ratio of the change to the adjustment range as the sensitivity coefficient of the characteristic component. The larger the absolute value of the sensitivity coefficient, the more significant the influence of the characteristic component on the strength. The sensitivity coefficients of all samples are statistically averaged, and the average sensitivity coefficient is calculated after removing outlier samples that deviate from the preset range of the average value. Establish a correlation table between influence weights and mix proportion parameters, and record the influence weights of sand particle size distribution, cement dosage and water-cement ratio and their corresponding sample statistical information.

8. The method according to claim 1, characterized in that, The step of performing dynamic PID parameter tuning based on the deviation adjustment amount involves adjusting the proportional control coefficient, integral control coefficient, and derivative control coefficient according to the changing trend of the deviation adjustment amount, and generating a proportioning optimization instruction that includes the adjustment direction and amplitude, including: The deviation adjustment amount is decomposed into sand particle size distribution adjustment component, cement dosage adjustment component and water-cement ratio adjustment component. Each component corresponds to the dimension order of the mix proportion state vector and constitutes an independent component adjustment sequence. Each component adjustment sequence contains the historical adjustment data of the current cycle and the previous two cycles. The rate of change index is calculated separately for each component adjustment sequence. The rate of change index is obtained by dividing the difference between the current period component value and the previous period component value by the difference between the previous period component value and the previous two period component values. Adjust the corresponding proportional adjustment coefficients according to the rate of change of each component. The rate of change of the sand particle size distribution adjustment component corresponds to the sand particle size distribution proportional coefficient, the rate of change of the cement dosage adjustment component corresponds to the cement dosage proportional coefficient, and the rate of change of the water-cement ratio adjustment component corresponds to the water-cement ratio proportional coefficient. The adjustment direction is consistent with the sign of the rate of change index, and the adjustment range is proportional to the absolute value of the index. The cumulative deviation index is calculated separately for each component adjustment sequence. The cumulative deviation index is the algebraic sum of the adjustment amounts of the component in the current period and the previous two periods. Based on the cumulative deviation index of each component, the corresponding integral adjustment coefficient is adjusted. When the absolute value of the cumulative deviation index of a certain component is greater than the preset threshold of that component, the integral adjustment coefficient of that component is increased; otherwise, it is decreased. The second-order difference value is calculated separately for each component adjustment sequence. The second-order difference value is obtained by subtracting the first-order difference value between the previous period and the two previous periods from the first-order difference value between the current period and the previous period. It reflects the change in the rate of change of the adjustment amount of the component. The differential adjustment coefficient of each component is adjusted based on the second-order difference value. The adjustment magnitude is proportional to the absolute value of the second-order difference value, and the direction is consistent with the sign of the second-order difference value. The proportional adjustment coefficient, integral adjustment coefficient, and derivative adjustment coefficient are combined in their original dimensional order to generate a ratio optimization instruction that includes the adjustment direction and amplitude of each component.

9. The method according to claim 8, characterized in that, The calculation of the rate of change index for each component-adjusted sequence includes: Select the target component from the sand particle size distribution adjustment component, cement dosage adjustment component and water-cement ratio adjustment component. The target component is the component of the change rate index to be calculated. Process it in the order of sand particle size distribution, cement dosage and water-cement ratio. The processing order is consistent with the dimensional order of the mix proportion state vector. Extract the component adjustment sequence of the target component. The sequence includes the component values ​​of the previous two periods, the component value of the previous period, and the component value of the current period. They are arranged in chronological order, and the unit of the component value is consistent with the unit of the adjustment amount of the component to ensure data unit consistency. The difference between the target component value in the previous period and the component value in the previous two periods is calculated as the first difference. This difference reflects the change of the target component in the previous two periods, and the sign of the difference indicates the direction of change. The difference between the current period component value and the previous period component value of the target component is calculated as the second difference value. This difference value reflects the change of the target component in the most recent two periods, and the sign of the difference value indicates the direction of change. If the first difference is zero, it indicates that the target component has not changed in the first two cycles. At this time, the rate of change index is set as the preset benchmark value. The benchmark value is set according to the historical adjustment trend of the component and is related to the component type. The benchmark values ​​of different components are set independently. If the first difference is not zero, the ratio of the second difference to the first difference is used as the rate of change index. A positive ratio indicates that the target component is changing faster, and a negative ratio indicates that the change is slowing down. The absolute value of the ratio reflects the degree of acceleration or deceleration. The ratios of different components are calculated independently. The calculated rate of change index is subject to range restrictions. When the absolute value of the index exceeds the maximum allowable rate of change for that component, the boundary value of the maximum rate of change is taken as the final rate of change index.

10. A computer system comprising a memory and a processor, the memory storing a computer program executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the method according to any one of claims 1 to 9.

Images