A soymilk spray drying self-adaptive control method based on multi-parameter monitoring

Abstract

The application discloses a kind of soy milk spray drying self-adaptive control method based on multi-parameter monitoring, it is related to process control technical field, including, acquisition soy milk feed liquid parameter, atomization site parameter, drying environment parameter and powder state parameter, and carry out state calibration;Actual powder state is deviated from the deviation direction and deviation amplitude of qualified powder state, and atomization load decomposition is carried out, to determine the burden of feed liquid and atomization dispersion quantity, and conversion is atomization load control quantity;Positioning droplet water loss section, and tracking adhesion echo position, the position of uncompleted surface vitrification is connected with adhesion echo position, and the amount of droplet adhesion echo is formed;Atomization load control quantity is revised, and atomization site parameter and drying environment parameter of next control cycle are constrained, and self-adaptive control result is formed.The application is associated with the same position by droplet water loss state and soy milk powder body retention, avoid local drying imbalance in the process of droplet water loss by heating and continuously accumulate and pass to subsequent powder stage.

Description

Technical Field

[0001] This invention relates to the field of process control technology, and in particular to an adaptive control method for soybean milk spray drying based on multi-parameter monitoring. Background Technology

[0002] With the development of powdered processing of plant protein beverages, the spray preparation of soy milk powder is gradually shifting from manual experience-based regulation to online detection-based collaborative control. Existing methods typically set detection points around raw material flow characteristics, nozzle working status, drying tower heat and humidity conditions, and powder output performance to obtain the operating status during the spray preparation process. Controllers are then used to adjust operating conditions such as feed rate, spray pressure, hot air temperature, and exhaust humidity. Currently, related research mainly focuses on spray droplet formation, heat and mass transfer within the tower, powder moisture retention, and wall adhesion inhibition. Continuous production operation control is achieved through sensor data acquisition, operating condition identification, and parameter adjustment.

[0003] Because of the significant interactions between proteins, carbohydrates, and lipids in the soy milk system, droplets easily enter an intermediate state during the heat drying process, where the outer layer remains adhesive. This can induce powder retention, agglomeration, and wall adhesion. Existing control methods struggle to pinpoint the source of powder adhesion to specific intervals within the droplet drying process, resulting in a lack of targeted operational conditions for subsequent control adjustments to address the causes of adhesion. (Summary of the Invention)

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides an adaptive control method for soybean milk spray drying based on multi-parameter monitoring to solve the problem of lack of control regulation targeting the cause of adhesion in the operating conditions.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] This invention provides an adaptive control method for soybean milk spray drying based on multi-parameter monitoring, comprising:

[0008] Collect parameters of soybean milk liquid, atomization part, drying environment and powder state during the same control cycle of soybean milk spray drying, and perform state calibration to generate a drying condition state table.

[0009] Define the qualified powdering state, and find the deviation direction and deviation magnitude of the actual powdering state relative to the qualified powdering state from the drying condition state table. Then, decompose the atomization load of the soybean milk liquid parameters and the atomization part parameters to determine the liquid load when the soybean milk enters the atomization part and the atomization dispersion of the soybean milk by the atomization part. Finally, perform the same-scale difference calculation on the liquid load and the atomization dispersion to form the atomization load control quantity.

[0010] Locate the droplet dehydration section from the drying environment parameters, and track the adhesion and re-emergence position of the soybean milk powder based on the powder state parameters. Connect the position in the droplet dehydration section where the surface vitrification has not been completed with the adhesion and re-emergence position to form the droplet adhesion and re-emergence amount.

[0011] The atomization load control quantity is corrected by using the droplet adhesion and reconnection amount, and the corrected atomization load control quantity is used to constrain the atomization part parameters and drying environment parameters in the next control cycle to form an adaptive control result.

[0012] As a preferred embodiment of the adaptive control method for soybean milk spray drying based on multi-parameter monitoring described in this invention, the state calibration involves marking the source, collection time, and operating condition location of the soybean milk liquid parameters, atomization part parameters, drying environment parameters, and powdering state parameters within the same control cycle.

[0013] As a preferred embodiment of the adaptive control method for soybean milk spray drying based on multi-parameter monitoring described in this invention, the specific steps for defining a qualified powdering state and finding the deviation direction and magnitude of the actual powdering state relative to the qualified powdering state from the drying condition state table are as follows:

[0014] Extract the powder formation state parameters that indicate the powder formed during the spray drying process of soybean milk reaches the preset powder formation quality conditions from the drying condition state table, and define them as qualified powder formation state. Extract the actual powder formation state from the powder formation state parameters.

[0015] By comparing the qualified powder state with the actual powder state, the deviation direction of the actual powder state relative to the qualified powder state is determined as the deviation direction, and the difference between the actual powder state and the qualified powder state is determined as the deviation amplitude.

[0016] As a preferred embodiment of the adaptive control method for soybean milk spray drying based on multi-parameter monitoring described in this invention, the specific steps for determining the feed liquid load and the atomization dispersion amount of the soybean milk at the atomization section are as follows:

[0017] Track the feed blockage location in the soybean milk feed parameters that corresponds to the direction of deviation;

[0018] The load strength value is determined by the deviation range and the degree of deviation of the soybean milk liquid parameters at the feed stagnation position, and the load strength value is converted into the liquid load when the soybean milk enters the atomization part.

[0019] The dispersion change direction of the atomization part parameters is limited by the deviation direction, and the atomization part parameters are converted into a droplet size spectrum.

[0020] The droplet size spectrum is divided into a fine droplet distribution segment and a coarse droplet concentration segment, and the droplet dispersion breakpoint is located by the degree of aggregation of the coarse droplet concentration segment in the droplet size spectrum.

[0021] By using the droplet dispersion breakpoint as the dividing position, the continuous distribution state of the fine droplet distribution segment and the aggregation state of the coarse droplet concentration segment are distinguished, and the atomization dispersion amount is obtained.

[0022] As a preferred embodiment of the adaptive control method for soybean milk spray drying based on multi-parameter monitoring described in this invention, the specific steps for calculating the atomization load control quantity by performing a same-scale difference calculation between the feed liquid load and the atomization dispersion amount are as follows:

[0023] The difference between the feed load and the atomization dispersion is calculated on the same scale to obtain the atomization load difference.

[0024] The atomization load control quantity is obtained by constraining the adjustment direction and adjustment amplitude by the positive and negative directions and absolute magnitude of the atomization load difference.

[0025] As a preferred embodiment of the adaptive control method for soybean milk spray drying based on multi-parameter monitoring described in this invention, the specific steps for locating the droplet water loss segment from the drying environment parameters are as follows:

[0026] According to the order of the collection time points, the rise and fall trends of humidity and temperature changes in the dry environment parameters were checked one by one, and the collection time point when the humidity change changed from flat to rising and the temperature change decreased synchronously was taken as the starting point of droplet water loss.

[0027] Starting from the point where the droplet loses water, continue to track adjacent collection points and take the point where the rate of increase in humidity change begins to decrease and the rate of decrease in temperature change begins to decrease synchronously as the point where the droplet loses water.

[0028] The sampling time points and corresponding operating conditions within the range from the droplet water loss initiation point to the droplet water loss termination point are defined as the droplet water loss segment.

[0029] As a preferred embodiment of the adaptive control method for soybean milk spray drying based on multi-parameter monitoring described in this invention, the specific steps for tracking the adhesion and display position of soybean milk powder retention according to powder state parameters are as follows:

[0030] In the powder formation status parameters, the actual powder formation status is statistically analyzed item by item according to the corresponding working condition location, and the collection time points of continuous discharge in the actual powder formation status are connected into normal powder output segments.

[0031] Using the end and beginning of adjacent normal powder-producing segments as boundaries, the collection points within the boundaries that are not covered by normal powder-producing segments are extracted to form the powder-producing gap locations.

[0032] The powder outlet gap location is used to determine the stagnation status and eliminate equipment abnormalities. The working condition location where there is powder stagnation but no abnormality in the powder discharge part is classified as the powder stagnation location.

[0033] Based on the location of powder retention, the powder agglomeration state is distinguished item by item in adjacent collection time points to form the visible location of powder adhesion;

[0034] The positions of powder retention and powder adhesion are compared, and the positions that match are determined as the adhesion display positions.

[0035] As a preferred embodiment of the adaptive control method for soybean milk spray drying based on multi-parameter monitoring described in this invention, the incomplete surface vitrification refers to the state in which the moisture content of the soybean milk droplets decreases in the droplet dehydration section, but the surface viscosity has not yet disappeared, and a stable dry surface layer has not been formed.

[0036] As a preferred embodiment of the adaptive control method for soybean milk spray drying based on multi-parameter monitoring described in this invention, the specific steps of correcting the atomization load control quantity using droplet adhesion re-collection amount are as follows:

[0037] The direction of adjustment of the atomization load control quantity is corrected by using the operating condition position of the droplet adhesion return quantity;

[0038] When the operating condition of the droplet adhesion and reconnection amount is in the droplet dehydration section, the adjustment direction of the atomization load control amount is adjusted to the direction of increasing the atomization dispersion amount. When the operating condition of the droplet adhesion and reconnection amount is not in the droplet dehydration section, the adjustment direction of the atomization load control amount is maintained, thus forming the corrected adjustment direction.

[0039] The amount of droplet adhesion and reconnection is converted into an amplitude correction amount according to the corresponding operating condition position, and the amplitude correction amount is used to calibrate the adjustment amplitude of the atomization load control amount to obtain the corrected atomization load adjustment amplitude.

[0040] As a preferred embodiment of the adaptive control method for soybean milk spray drying based on multi-parameter monitoring described in this invention, the specific steps for constraining the atomization part parameters and drying environment parameters in the next control cycle are as follows:

[0041] The atomization dispersion adjustment direction of the atomization part parameter in the next control cycle is replaced with the corrected adjustment direction.

[0042] The adjustment range of the atomization part parameters in the next control cycle is limited by the modified atomization load adjustment range;

[0043] Obtain the number of locations where incomplete surface vitrification is formed in the next control cycle, and compare it with the number of locations where incomplete surface vitrification is formed in the current control cycle;

[0044] When the number of incomplete surface vitrification sites decreases in the next control cycle, maintain the atomization site parameters and drying environment parameters for the next control cycle.

[0045] If the number of incomplete surface vitrification locations formed in the next control cycle does not decrease, the drying environment parameters for the next control cycle are corrected in the direction of reducing the number of incomplete surface vitrification locations to form an adaptive control result.

[0046] The beneficial effects of this invention are as follows: By isotopically correlating the droplet dehydration state with the retention of soy milk powder, reverse localization between the adhesion formation location and abnormal powder formation phenomenon during the spray drying process of soy milk is achieved. This can effectively distinguish between powder adhesion problems caused by insufficient solidification of the droplet surface and powder formation deviations caused by fluctuations in normal operating conditions. It reduces the control lag caused by adjustments based on the final powder formation result, avoids the continuous accumulation of local drying imbalance during the droplet heating and dehydration process and its transmission to subsequent powder formation stages, enhances the traceability of the source of abnormality and the reliability of the adjustment basis, improves the accuracy, continuity and stability of process control, and helps maintain the consistency of the powder formation state during the spray drying process of soy milk. Attached Figure Description

[0047] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a flowchart of an adaptive control method for soybean milk spray drying based on multi-parameter monitoring.

[0049] Figure 2 A flowchart for generating atomization load control quantities.

[0050] Figure 3 A flowchart for generating droplet adhesion and reconnection volume.

[0051] Figure 4 A flowchart for generating adaptive control results. Detailed Implementation

[0052] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0053] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0054] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0055] Reference Figures 1-4 This is one embodiment of the present invention, which provides an adaptive control method for soybean milk spray drying based on multi-parameter monitoring, comprising the following steps:

[0056] S1. Collect parameters of soybean milk liquid, atomization part, drying environment and powder state during the same control cycle of soybean milk spray drying, and perform state calibration to generate a drying condition state table.

[0057] Using the same control cycle as the collection boundary, the parameters of the soybean milk liquid, the atomization part, the drying environment, and the powdering state are continuously recorded during the operation of soybean milk spray drying. During the recording process, batch tracking identifiers are configured for the parameters of the soybean milk liquid entering the atomization part. The batch tracking identifiers are retained as the same batch of soybean milk passes through the atomization part, the drying space, and the powdering position in sequence.

[0058] The same control cycle is the time range covered by the soybean milk spray dryer controller to complete one parameter reading, status calculation, and control result output. The control cycle is determined based on the data update time of the detection equipment, the residence time of the soybean milk from the atomization part into the drying space, and the response time required for the atomization pressure, atomization speed, feed flow rate, and drying temperature to complete one stable adjustment.

[0059] Based on the residence time required for soy milk to sequentially reach the atomization part, the drying space detection position, and the powdering position from the feeding position, the data collection time points of the parameters obtained at different detection positions are aligned.

[0060] When the spray drying equipment is put into production for the first time, when the nozzles are replaced, or when the drying tower is cleaned and maintained, the feed flow rate is adjusted three times in a step. For each step adjustment, the time when the feed flow rate begins to change, the time when the atomization parameters begin to stabilize, the time when the drying environment parameters begin to stabilize, and the time when the powder state parameters begin to stabilize are recorded.

[0061] The time interval between the point when the atomization parameters begin to stabilize and the point when the feed flow rate begins to change is defined as the atomization residence time; the time interval between the point when the drying environment parameters begin to stabilize and the point when the feed flow rate begins to change is defined as the drying residence time; and the time interval between the point when the powder formation state parameters begin to stabilize and the point when the feed flow rate begins to change is defined as the powder formation residence time. When there are differences in the time intervals between the three step adjustments, the median value of the three time intervals is used.

[0062] Using the feeding time as the baseline time, the parameters of the soybean milk liquid obtained at the baseline time, the parameters of the atomization part obtained by extending the atomization residence time from the baseline time, the parameters of the drying environment obtained by extending the drying residence time from the baseline time, and the parameters of the powdering state obtained by extending the powdering residence time from the baseline time are classified into the same record group under the same batch tracking identifier.

[0063] When there is no detection record at the target acquisition time point, the detection records of one acquisition time point before and after the target acquisition time point are selected for linear interpolation; when the time interval between the target acquisition time point and the adjacent detection record exceeds two acquisition intervals, the current record group is marked as a data missing record group, and the data missing record group is not used to form the deviation direction, deviation amplitude, liquid load, atomization dispersion amount and droplet adhesion reconnection amount.

[0064] Soy milk feed parameters are parameters of the state of the soy milk feed before it enters the atomization part, such as soy milk concentration, solid content, feed temperature, feed flow rate, and feed viscosity.

[0065] The atomization parameters are the parameters of the process of soy milk forming droplets in the atomization part, such as atomization pressure, atomization speed, atomization flow rate, atomization angle, and droplet size distribution.

[0066] Drying environment parameters are parameters related to the water loss and drying state of droplets within the drying space, such as drying temperature, exhaust temperature, ambient humidity, exhaust humidity, and airflow velocity.

[0067] Powder formation parameters are parameters of the powder formation state after the soybean milk has been dried, such as powder moisture content, powder particle size, powder flowability, actual powder formation state, qualified powder formation state, powder agglomeration state, powder discharge state, and powder adhesion state.

[0068] To further clarify, the qualified powder state is the powder state parameter that the powder formed during the spray drying process of soybean milk reaches the preset powder quality conditions; the powder state parameters of historical qualified batches serve as the basis for setting the qualified powder state, and the powder state parameters of the current control cycle serve as the actual powder state.

[0069] Preset powder quality conditions are powder state conditions determined in advance based on the powdering requirements of the target soy milk powder product before the production control of soy milk spray drying. They are used to determine whether the actual powder state formed in the current control cycle has reached the qualified powder state. Preset powder quality conditions include powder moisture content at the product's allowable level, powder particle size meeting the target specifications, powder flowability meeting packaging and conveying needs, powder being able to be continuously and normally discharged, powder not significantly sticking to the wall, and powder not showing abnormal agglomeration.

[0070] When setting the qualified powder formation state, the powder moisture content, powder particle size, powder flowability, powder discharge state, powder adhesion state, and powder agglomeration state are first read from the powder formation state parameters of historical qualified batches. These parameters are then combined with the target range of powder moisture content, the target range of powder particle size, the target boundary of powder flowability, the allowable agglomeration level number, the continuous discharge state, and the allowable adhesion state to jointly define the qualified powder formation state. The powder formation state parameters collected in the current control cycle are not included in the setting of the qualified powder formation state, but are only used to find the deviation direction and deviation magnitude between the actual powder formation state and the qualified powder formation state.

[0071] When calibrating the status, the source of each parameter is marked according to the generation process corresponding to each parameter. Specifically, the parameters of the soybean milk liquid are marked as the source of the liquid side, the parameters of the atomization part are marked as the source of the atomization side, the parameters of the drying environment are marked as the source of the drying side, and the parameters of the powdering state are marked as the source of the powdering side.

[0072] The actual recorded time corresponding to each parameter is retained as the data acquisition time point, and the operating condition position is determined according to the operating status of the data acquisition time point within the same control cycle. Specifically, the data acquisition time point in the feeding state is marked as the feeding operating condition position, the data acquisition time point in the atomizing state is marked as the atomizing operating condition position, the data acquisition time point in the drying state is marked as the drying operating condition position, and the data acquisition time point in the powdering state is marked as the powdering operating condition position.

[0073] Once the parameter source, acquisition time, and corresponding operating condition location are all labeled, record groups with the same operating condition location are grouped into the same operating condition category. All record groups formed within the same control cycle are then arranged in the order of acquisition time to form a drying operating condition status table.

[0074] It should be noted that the target range for powder moisture content is set based on the target moisture content of the target soy milk powder product and the fluctuation range of powder moisture content in historical qualified batches. The upper boundary of the target range for powder moisture content is determined by the upward floating position corresponding to the target moisture content and the fluctuation range of powder moisture content in historical qualified batches, and the lower boundary of the target range for powder moisture content is determined by the downward floating position corresponding to the target moisture content and the fluctuation range of powder moisture content in historical qualified batches. The example value of the fluctuation range of powder moisture content in historical qualified batches is one to two times the standard deviation of powder moisture content in historical qualified batches. When the standard deviation of powder moisture content in historical qualified batches is zero, the sampling resolution of the powder moisture content detection equipment is used as the fluctuation range of powder moisture content in historical qualified batches.

[0075] The target particle size range is set based on the target particle size of the target soy milk powder product and the fluctuation range of the particle size of historical qualified batches. The upper boundary of the target particle size range is determined by the upward floating position corresponding to the target particle size and the fluctuation range of the particle size of historical qualified batches, and the lower boundary of the target particle size range is determined by the downward floating position corresponding to the target particle size and the fluctuation range of the particle size of historical qualified batches. The example value of the fluctuation range of the particle size of historical qualified batches is one to two times the standard deviation of the particle size of historical qualified batches. When the standard deviation of the particle size of historical qualified batches is zero, the sampling resolution of the particle size detection equipment is used as the fluctuation range of the particle size of historical qualified batches.

[0076] The target boundary for powder flowability is set based on the packaging and conveying status of the target soy milk powder product and the fluctuation range of powder flowability in historical qualified batches. When powder flowability is expressed as a flowability value, the target boundary for powder flowability is determined by the target flowability value and the corresponding downward position of the fluctuation range of powder flowability in historical qualified batches. The example value of the fluctuation range of powder flowability in historical qualified batches is one to two times the standard deviation of powder flowability in historical qualified batches. When the standard deviation of powder flowability in historical qualified batches is zero, the sampling resolution of the powder flowability detection equipment is used as the fluctuation range of powder flowability in historical qualified batches.

[0077] The powder agglomeration state is ranked in the order of no agglomeration, slight agglomeration, obvious agglomeration, and blockage agglomeration to form an agglomeration level number; the allowable agglomeration level number is set according to the allowable agglomeration state of the target soy milk powder product and the powder agglomeration state of historical qualified batches. The allowable agglomeration level number is the agglomeration level number that is allowed to appear in the target soy milk powder product. The powder agglomeration state in historical qualified batches that does not exceed the allowable agglomeration level number is used as the basis for setting the allowable agglomeration state.

[0078] The continuous discharge status is determined based on the collection time point of continuous powder discharge in historical qualified batches; the permissible adhesion status is determined based on the records of powder adhesion in historical qualified batches that did not affect continuous discharge.

[0079] For example, in one embodiment, the control cycle is set to sixty seconds, the parameter acquisition interval is set to one second, and each control cycle forms sixty acquisition time points. If the data update time of the detection device or the stable response time of the execution component exceeds sixty seconds, the control cycle is extended by the larger of the maximum data update time of the detection device and the maximum stable response time of the execution component.

[0080] The parameters of the soybean milk liquid are collected by concentration detection probe, temperature detection probe, flow detection probe and viscosity detection probe installed in the feed pipe of the atomization part; the parameters of the atomization part are collected by atomization pressure detection probe, atomizer speed encoder, atomization flow detection probe and laser particle size detection probe; the parameters of the drying environment are collected by temperature detection probe, humidity detection probe and airflow velocity detection probe installed at different heights in the drying space; the parameters of the powder state are collected by online moisture detection probe, powder particle size detection probe, powder flow detection probe and powder adhesion detection probe installed at the powder discharge position.

[0081] S2. Define the qualified powdering state, and find the deviation direction and deviation range of the actual powdering state relative to the qualified powdering state from the drying condition state table. Then, decompose the atomization load of the soybean milk liquid parameters and the atomization part parameters to determine the liquid load when the soybean milk enters the atomization part and the atomization dispersion of the soybean milk by the atomization part. Convert the liquid load and atomization dispersion into atomization load control quantity.

[0082] S2.1 Define the acceptable powdering state, and find the direction and magnitude of the deviation between the actual powdering state and the acceptable powdering state from the drying condition status table. Specifically,

[0083] S2.1.1 Extract the powder formation state parameters from the drying condition state table to indicate that the powder formed during the spray drying process of soybean milk meets the preset powder formation quality conditions, and define them as qualified powder formation states. Extract the actual powder formation state from the powder formation state parameters, specifically:

[0084] In the drying condition status table, within the same control cycle, find the corresponding condition position of the powdering condition parameter; mark the powdering condition parameter that meets the preset powdering quality condition as qualified powdering condition.

[0085] The historical qualified batches used to form the qualified powder state are the same target soy milk powder product specifications, the same spray drying equipment, and the same main raw material ratios used in the current control cycle. When the number of historical qualified batches is no less than ten, the 5th percentile and 95th percentile values ​​of powder moisture content, powder particle size, and powder flowability are respectively calculated, and their intersection with the quality allowable range of the target soy milk powder product is calculated to form the target range of powder moisture content, the target range of powder particle size, and the target boundary of powder flowability.

[0086] When the number of qualified batches in the past is less than ten, the target range for powder moisture content and the target range for powder particle size shall adopt the allowable range in the quality standard of the target product, and the target boundary for powder flowability shall adopt the minimum allowable flowability value that the packaging and conveying equipment can continuously operate.

[0087] The permissible agglomeration level is determined by the allowable percentage of oversized powder particles in the target soy milk powder product. In one embodiment, a percentage of powder particles exceeding the upper boundary of the target particle size range of less than 3% is considered no agglomeration; a percentage of 3% to less than 8% is considered slight agglomeration; a percentage of 8% to less than 15% is considered significant agglomeration; and a percentage of 15% or more is considered clogged agglomeration. The permissible agglomeration level is determined according to product quality standards.

[0088] The historical fluctuation range of each powder state parameter is calculated using the difference between the 95th percentile and the 5th percentile of the corresponding parameter in the historical qualified batches. When the fluctuation range of the historical qualified batches is less than the acquisition resolution of the detection equipment, the acquisition resolution of the detection equipment shall be used.

[0089] Under the same operating conditions, filter and collect parameters of powder formation status that belong to the current control cycle, and use the filtered parameters of powder formation status as the actual powder formation status.

[0090] S2.1.2. Compare the qualified powder state with the actual powder state. Determine the deviation direction of the actual powder state relative to the qualified powder state as the deviation direction, and determine the magnitude of the difference between the actual powder state and the qualified powder state as the deviation amplitude. Specifically,

[0091] Under the same working conditions, read the target range of powder moisture content, target range of powder particle size, target boundary of powder flowability, and allowable agglomeration level number in the qualified powdering state, and read the powder moisture content, powder particle size, powder flowability, and powder agglomeration state in the actual powdering state.

[0092] When the actual powder moisture content in the powdered state exceeds the upper boundary of the target powder moisture content range, the deviation direction of the sub-item is determined as the direction of being too wet, and the difference between the actual powder moisture content and the upper boundary of the target powder moisture content range is taken as the powder moisture content deviation. When the actual powder moisture content in the powdered state is lower than the lower boundary of the target powder moisture content range, the deviation direction of the sub-item is determined as the direction of being too dry, and the difference between the lower boundary of the target powder moisture content range and the actual powder moisture content is taken as the powder moisture content deviation. When the actual powder moisture content in the powdered state is within the target powder moisture content range, the powder moisture content deviation is taken as zero.

[0093] When the actual powder particle size exceeds the upper boundary of the target particle size range, the deviation direction is determined as the coarsening direction, and the difference between the actual powder particle size and the upper boundary of the target particle size range is taken as the powder particle size deviation. When the actual powder particle size is lower than the lower boundary of the target particle size range, the deviation direction is determined as the refining direction, and the difference between the lower boundary of the target particle size range and the actual powder particle size is taken as the powder particle size deviation. When the actual powder particle size is within the target particle size range, the powder particle size deviation is zero.

[0094] When the actual powder flowability in the powdered state does not reach the target boundary of powder flowability, the deviation direction of the sub-item is determined as the direction of insufficient flowability, and the difference between the target boundary of powder flowability and the actual powder flowability is taken as the powder flowability deviation amount; when the actual powder flowability in the powdered state reaches the target boundary of powder flowability, the powder flowability deviation amount is zero.

[0095] When the agglomeration level number corresponding to the actual powder agglomeration state exceeds the allowable agglomeration level number, the deviation direction of the sub-item is determined as the agglomeration enhancement direction, and the difference between the agglomeration level number corresponding to the actual powder agglomeration state and the allowable agglomeration level number is taken as the powder agglomeration state deviation amount; when the agglomeration level number corresponding to the actual powder agglomeration state does not exceed the allowable agglomeration level number, the powder agglomeration state deviation amount is taken as zero.

[0096] The deviations in powder moisture content, particle size, flowability, and agglomeration state were normalized using the corresponding historical qualified batch fluctuation ranges to form the moisture content deviation share, particle size deviation share, flowability deviation share, and agglomeration state deviation share. The historical qualified batch fluctuation range is the difference between the maximum and minimum records of the corresponding powdering state parameters for the same soybean milk powder target. When the historical qualified batch fluctuation range is zero, the collection resolution of the corresponding powdering state parameter is used as the normalization denominator, and the powder agglomeration state is normalized using the interval between adjacent agglomeration level numbers.

[0097] For the deviations in moisture content, particle size, flowability, and agglomeration, weights for powder formation parameters are assigned respectively. The weights for powder formation parameters are determined by the proportion of powder moisture content, particle size, flowability, and agglomeration in historical qualified batches that cause deviations in powder formation. If there are insufficient historical qualified batches, the weights for each powder formation parameter are determined according to the same proportion. The deviation proportions are then summarized after being corrected by the corresponding powder formation parameter weights to form the deviation range.

[0098] The formula for the deviation magnitude is:

[0099] ;

[0100] In the formula, Indicates the magnitude of the deviation; Indicates the first Normalized deviation values ​​of the powder state parameters; Indicates the first The weights of the powder formation state parameters are defined; the sum of the weights of each parameter is one.

[0101] The value range is from zero to one.

[0102] Will The deviation direction is determined by the direction of the largest deviation in the powder formation state parameter.

[0103] After adjusting the weights of each powder formation state parameter, the deviation proportions are arranged in descending order. The deviation direction of the sub-item corresponding to the powder formation state parameter with the sorting number 1 is taken as the deviation direction. When the deviation proportions of two or more powder formation state parameters are the same after adjusting the weights, the deviation direction is determined according to the powder formation control priority set before production control. The powder formation control priority is as follows: powder moisture content, powder agglomeration state, powder particle size, and powder flowability.

[0104] S2.2 The atomization load is decomposed based on the parameters of the soybean milk feed liquid and the atomization part, determining the feed liquid load when the soybean milk enters the atomization part and the atomization dispersion amount of the soybean milk by the atomization part. Specifically,

[0105] S2.2.1 Track the feed resistance location in the soybean milk feed parameters that corresponds to the direction of deviation, specifically:

[0106] Using the location of the deviation direction and the batch tracking mark as the positioning basis, trace back to the feeding location under the same batch tracking mark, and compile the changes in soybean milk concentration, liquid viscosity and liquid flow rate at continuous sampling points within the feeding location.

[0107] When the concentration of soy milk increases during continuous sampling points and the flow rate of the liquid decreases during the same continuous sampling points, the operating condition position of the sampling point is marked as the concentration blockage position; when the viscosity of the liquid continues to increase and the flow rate of the liquid decreases synchronously, the operating condition position of the sampling point is marked as the conveying blockage position.

[0108] Based on the direction of deviation, the candidate locations for concentration stagnation and conveying stagnation are assigned and organized. When the direction of deviation is the direction of moisture, the candidate location for concentration stagnation is determined as the concentration stagnation location; when the direction of deviation is the direction of coarsening, the candidate locations for concentration stagnation and conveying stagnation are determined as the concentration stagnation location and the conveying stagnation location, respectively.

[0109] When the deviation direction is the direction of insufficient fluidity, the candidate position of conveying blockage is determined as the conveying blockage position; when the deviation direction is the direction of enhanced agglomeration, the candidate positions of concentration blockage and conveying blockage are determined as concentration blockage and conveying blockage positions, respectively; when the deviation direction is the direction of dryness and the direction of refinement, the candidate positions of concentration blockage and conveying blockage are not converted into feed blockage positions.

[0110] The concentration blockage location and the conveying blockage location are classified as feed blockage locations.

[0111] S2.2.2. Determine the load strength value by utilizing the deviation range and the degree of deviation of the soybean milk liquid parameters at the feed stagnation position, and convert the load strength value into the liquid load amount when the soybean milk enters the atomization part. Specifically,

[0112] Calculate the deviation of each feed blockage position within the same control cycle, and then... Degree of deviation of each feed blockage position The ratio of the deviation from the sum of all feed obstruction positions is determined as the first value. The impact of each feed obstruction location on the percentage .

[0113] The formula for the degree of deviation of the feed obstruction position is:

[0114] ;

[0115] In the formula, Indicates the first The degree of deviation of each feed obstruction position; Indicates the first The portion of the soybean milk concentration that increases at the feed hindrance site This indicates the fluctuation range of soy milk concentration in historically qualified batches. Indicates the first The portion of the feed liquid viscosity that increases at the feed hindrance point This indicates the fluctuation range of the viscosity of the liquid material in historical qualified batches. Indicates the first The portion of the feed flow rate that decreases at each feed obstruction point This indicates the fluctuation range of liquid flow rate in historical qualified batches; An index variable representing the concentration of soy milk; An index variable representing the viscosity of the liquid feed; An index variable representing the flow rate of the liquid feed.

[0116] The formula for the percentage impact of feed obstruction location is:

[0117] ;

[0118] In the formula, Indicates the first The percentage affected by each feed obstruction location; Indicates the first The degree of deviation corresponding to each feed obstruction position; This indicates the number of feed obstruction locations within the same control cycle; An index variable indicating the location of feed obstruction; This represents the sum of deviations corresponding to all feed obstruction positions within the same control cycle; when the sum of deviations corresponding to all feed obstruction positions within the same control cycle is zero, the influence ratio of each feed obstruction position is determined according to the same share.

[0119] In the At each feed obstruction location, the soy milk concentration, liquid viscosity, and liquid flow rate are read. Using the soy milk concentration, liquid viscosity, and liquid flow rate of the corresponding qualified powder state in historical qualified batches as references, the increases in soy milk concentration, liquid viscosity, and liquid flow rate are recorded respectively. When the soy milk concentration is not higher than the reference soy milk concentration, the increase in soy milk concentration is recorded as zero; when the liquid viscosity is not higher than the reference liquid viscosity, the increase in liquid viscosity is recorded as zero; when the liquid flow rate is not lower than the reference liquid flow rate, the decrease in liquid flow rate is recorded as zero.

[0120] The concentration fluctuation range of historical qualified batches of soy milk is used as the concentration normalization denominator, the viscosity fluctuation range of historical qualified batches of liquid is used as the viscosity normalization denominator, and the flow rate fluctuation range of historical qualified batches of liquid is used as the flow rate normalization denominator. The fluctuation range of historical qualified batches is the difference between the maximum and minimum records of the corresponding parameters of the historical qualified batches under the same soy milk powder target. When the fluctuation range of historical qualified batches is zero, the resolution of the corresponding parameter acquisition sensor is used as the normalization denominator.

[0121] Based on the increased concentration of soy milk, increased viscosity of the feed liquid, decreased feed liquid flow rate, the fluctuation range of corresponding historical qualified batches, and the proportion of influence from the feed obstruction location, the feed liquid load when soy milk enters the atomization section is calculated using the following formula:

[0122] ;

[0123] In the formula, This indicates the amount of liquid feed required when the soy milk enters the atomization section.

[0124] S2.2.3. The dispersion variation of parameters in the atomization region is limited by the deviation direction. The parameters in the atomization region are then converted into a droplet size spectrum. Specifically,

[0125] Based on the direction of deviation, find the changes in atomization pressure, atomization speed, atomization flow rate, atomization angle, and droplet size distribution within the same operating condition location, and retain the corresponding batch tracking identifier and collection time point.

[0126] When the deviation direction is the direction of moistening, coarsening, or agglomeration enhancement, the change in droplet size distribution towards the concentration of subsequent droplet size intervals is determined as the coarse droplet concentration direction.

[0127] When the deviation direction is either the dryness direction or the refinement direction, the change in droplet size distribution diffused into the preceding particle size range is determined as the direction of fine droplet dispersion.

[0128] When the deviation direction is in the direction of insufficient fluidity, further examine the deviation direction of the powder state component; when the deviation direction of the powder particle size component is in the direction of coarsening or the deviation direction of the powder agglomeration state component is in the direction of enhanced agglomeration, determine the change in droplet particle size distribution towards the concentration of subsequent particle size intervals as the direction of coarse droplet concentration.

[0129] When the deviation of the powder particle size distribution is in the direction of refinement and the deviation of the powder agglomeration state is zero, the change in the droplet particle size distribution diffusing into the preceding particle size interval is determined as the direction of fine droplet dispersion.

[0130] Using the droplet size as the sorting criterion, the droplet distribution in each size range is arranged sequentially, the number and density of droplets in each size range are counted, and the sorting results of all size ranges are compiled into a droplet size spectrum.

[0131] The laser particle size detection probe outputs multiple particle size ranges and the number of droplets within each range at each acquisition point. The droplet size spectrum is arranged in ascending order of the upper boundary of the particle size range, and the number of droplets within each range is used as the spectral value of the range.

[0132] When the particle size ranges used at different collection points are inconsistent, the minimum particle size resolution of the particle size detection equipment is used as the uniform particle size interval, and the number of droplets is redistributed to the uniform particle size range.

[0133] When a single droplet spans two adjacent uniform particle size intervals, the number of droplets is allocated according to the distance between the droplet size and the center position of the two uniform particle size intervals.

[0134] To suppress detection fluctuations within a single particle size range, the droplet size spectrum is processed by averaging three adjacent particle size ranges.

[0135] The smoothed spectral value of the current particle size interval is formed by averaging the number of droplets in the previous particle size interval, the current particle size interval, and the next particle size interval; the beginning and end of the droplet particle size spectrum only use the adjacent particle size intervals that can be obtained.

[0136] When the number of droplets increases for three consecutive particle size intervals and then decreases for three consecutive particle size intervals, the interval where the increase turns into a decrease is determined as the aggregation and convergence position; the continuous increase or decrease occurs when the change in the number of droplets in adjacent particle size intervals reaches the state of repeated detection error of the particle size detection equipment.

[0137] S2.2.4. Divide the droplet size spectrum into a fine droplet distribution segment and a coarse droplet concentration segment, and locate the droplet dispersion breakpoint based on the degree of aggregation of the coarse droplet concentration segment in the droplet size spectrum. Specifically,

[0138] The droplet size ranges in the droplet size spectrum are arranged in order of size, and the number of droplets distributed in each size range is counted.

[0139] The particle size range that maintains a dispersed distribution state during the gradual increase in particle size is classified as the fine droplet distribution segment, and the region where the number of droplets continuously gathers in adjacent particle size ranges is classified as the coarse droplet concentration segment.

[0140] The changes in the number of droplets in the coarse droplet concentration segment are analyzed. The position where the number of droplets begins to increase continuously is taken as the aggregation start position, and the position where the number of droplets stops increasing and begins to decrease is taken as the aggregation convergence position.

[0141] When multiple coarse droplet clusters appear in the droplet size spectrum, the aggregation start position and aggregation end position of each coarse droplet cluster are marked, and the total percentage of droplet distribution in each coarse droplet cluster is calculated. The coarse droplet clusters are arranged in descending order of total percentage, and the coarse droplet cluster with the sort number 1 is selected as the target coarse droplet cluster. If two or more coarse droplet clusters have the same total percentage, the coarse droplet cluster with the particle size interval number corresponding to the aggregation end position located at the end of the droplet size spectrum is selected as the target coarse droplet cluster. The aggregation end position of the target coarse droplet cluster is determined as the droplet dispersion breakpoint.

[0142] When there is a single-point jump in the number of droplets in adjacent particle size intervals, the resolution of adjacent particle size intervals of the particle size detection device is used as the smoothing basis. The droplet distribution number of three consecutive adjacent particle size intervals is averaged and the droplet distribution number is reorganized. The aggregation and convergence position that still maintains a continuous increase after smoothing and then turns to decrease is determined as the droplet dispersion breakpoint.

[0143] When there is no region in the droplet size spectrum where the number of droplets is continuously concentrated in the adjacent subsequent size intervals, it is recorded as a segment where no coarse droplet concentration is formed. When no coarse droplet concentration is formed, the aggregation and convergence position is not marked. The boundary of the size interval with the sequence number at the end in the fine droplet distribution segment is determined as the droplet dispersion breakpoint. The atomization dispersion amount is determined according to the continuous state of the number of droplets in each size interval within the fine droplet distribution segment.

[0144] The aggregation region formed from the aggregation initiation position to the aggregation termination position represents the degree of aggregation of the coarse droplet concentration segment in the droplet size spectrum, and the aggregation termination position is determined as the droplet dispersion breakpoint.

[0145] S2.2.5. Using the droplet dispersion breakpoint as the dividing position, distinguish the continuous distribution state of the fine droplet distribution segment and the aggregation state of the coarse droplet concentration segment to obtain the atomization dispersion amount, specifically,

[0146] Using the droplet dispersion breakpoint as the boundary, count the number of droplets in the fine droplet distribution segment and the number of droplets in the coarse droplet concentration segment.

[0147] The number of droplets in each size range is arranged in ascending order of size to form a droplet size spectrum. The droplet size spectrum is then processed by averaging three consecutive adjacent size ranges to reduce the impact of detection jumps in a single size range.

[0148] When the number of droplets continuously increases and then continuously decreases in subsequent particle size intervals, the particle size interval from increasing to decreasing is determined as the droplet dispersion breakpoint.

[0149] A fine droplet distribution segment is formed at the front end of the droplet dispersion break point, and a coarse droplet concentration segment is formed at the rear end of the droplet dispersion break point.

[0150] Let the number of droplets in the fine droplet distribution segment be . The total number of droplets in the droplet size spectrum is Then the atomization dispersion amount is:

[0151] ;

[0152] In the formula, Indicates the amount of atomized dispersion; This indicates the number of droplets in the fine droplet distribution segment; This indicates the total number of droplets in the droplet size spectrum.

[0153] The range of values ​​for atomized dispersion is: ;

[0154] An increase in atomization dispersion indicates an increase in the proportion of droplets in the fine droplet distribution segment, while a decrease in atomization dispersion indicates that droplets are converging towards the coarse droplet concentration segment.

[0155] S2.3 Calculate the difference between the feed load and the atomization dispersion on the same scale to form the atomization load control value, specifically,

[0156] S2.3.1 Calculate the difference between the feed load and the atomization dispersion on the same scale to obtain the atomization load difference. The formula is as follows:

[0157] ;

[0158] In the formula, Indicates the difference in atomization load; Indicates the feed liquid load; This indicates the amount of atomized dispersion.

[0159] The feed liquid load reflects the degree of dispersion required when the soybean milk enters the atomization section; the atomization dispersion reflects the actual dispersion capacity provided by the atomization section.

[0160] S2.3.2. The atomization load control quantity is obtained by constraining the adjustment direction and adjustment amplitude by the positive and negative directions and absolute magnitude of the atomization load difference, respectively.

[0161] when When the feed load is higher than the atomization dispersion rate, the adjustment direction is limited to increasing the atomization dispersion rate, and the adjustment range is taken as... ;when When the feed load is lower than the atomization dispersion rate, the adjustment direction is limited to reducing the atomization dispersion rate, and the adjustment range is taken as... ;when At this time, the feed load and atomization dispersion are balanced, the adjustment direction remains in the current state, and the adjustment range is 0; the adjustment direction and adjustment range together constitute the atomization load control quantity.

[0162] S3. Locate the droplet dehydration section from the drying environment parameters, and track the adhesion and re-emergence position of the soybean milk powder based on the powder state parameters. Connect the position in the droplet dehydration section where the surface vitrification has not been completed with the adhesion and re-emergence position to form the droplet adhesion and re-emergence amount.

[0163] S3.1 Locate the water loss segment of the droplet from the drying environment parameters, specifically:

[0164] S3.1.1. According to the chronological order of the data collection time points, the moving average of the drying temperature and ambient humidity at the same operating condition location is calculated for three consecutive data collection time points. When the absolute value of the change in ambient humidity does not exceed twice the acquisition resolution of the humidity detection probe, it is recorded as a gradual humidity trend; when the increase in ambient humidity reaches twice the acquisition resolution of the humidity detection probe, it is recorded as a humidity rising trend; when the decrease in drying temperature reaches twice the acquisition resolution of the temperature detection probe, it is recorded as a temperature falling trend.

[0165] When the humidity change changes from a gradual trend to an increasing trend, and the temperature change maintains an downward trend for three consecutive sampling points, the first sampling point among the three consecutive sampling points is determined as the starting point of droplet water loss.

[0166] S3.1.2. Starting from the point where the droplet loses water, continue tracking adjacent data collection points. The point where the rate of increase in humidity begins to decrease and the rate of decrease in temperature begins to decrease synchronously is taken as the point where the droplet loses water. Specifically,

[0167] Starting from the next sampling point after the droplet begins to lose water, the data on the increase in ambient humidity and the decrease in drying temperature are collected. The increase in ambient humidity is the difference between the ambient humidity at the current sampling point and the ambient humidity at the previous sampling point, and the decrease in drying temperature is the difference between the drying temperature at the previous sampling point and the drying temperature at the current sampling point.

[0168] The points where the increase in ambient humidity changes from increasing to decreasing and the decrease in the decrease in drying temperature changes from increasing to decreasing in adjacent sampling time points are marked as droplet water loss convergence points.

[0169] S3.1.3 After determining the droplet water loss initiation point and droplet water loss termination point, retain the droplet water loss initiation point, droplet water loss termination point, and the sampling time points within the section between the droplet water loss initiation point and droplet water loss termination point, and retain the corresponding operating condition position of each sampling time point; the sampling time points are used to limit the time boundary of the droplet water loss segment, and the corresponding operating condition position is used to limit the operating position of the droplet water loss segment. The sampling time points and corresponding operating condition positions within the droplet water loss initiation point to droplet water loss termination point are defined as the droplet water loss segment.

[0170] S3.2. Track the adhesion and display position of the soybean milk powder based on the powder state parameters. Specifically,

[0171] S3.2.1 In the powder formation status parameters, the actual powder formation status is statistically analyzed item by item according to the corresponding working condition location. The collection time points of continuous discharge in the actual powder formation status are connected to form a normal powder output segment. Specifically,

[0172] In the powder formation status parameters, the actual powder formation status is organized according to the name of the working condition location. The actual powder formation status whose working condition location name is the powder formation working condition location is included in the powder formation status, and the actual powder formation status whose working condition location name is not the powder formation working condition location is not included in the powder formation status.

[0173] Within the powder formation location, collection points where the powder discharge status shows continuous discharge and the powder agglomeration status does not show abnormal agglomeration are listed as normal powder discharge collection points; when the powder discharge status of adjacent normal powder discharge collection points both show continuous discharge, the sequential connection position of adjacent normal powder discharge collection points is retained, and the positions of consecutive adjacent normal powder discharge collection points are connected according to the sequential order of collection points, and the connected continuous collection point segment forms a normal powder discharge segment.

[0174] S3.2.2 In adjacent normal powder output segments, the last acquisition time point of the previous normal powder output segment is recorded as the segment tail end, and the first acquisition time point of the next normal powder output segment is recorded as the segment head end. The segment tail end and segment head end of adjacent normal powder output segments are used as boundaries. Acquisition time points within the boundaries that are not covered by normal powder output segments are extracted. Acquisition time points that are not included in normal powder output segments are retained as powder output interruption acquisition time points.

[0175] The location of the working condition where the powder output interruption data collection point is located, together with the powder output interruption data collection point, is used as the powder output gap location.

[0176] S3.2.3. Determine the powder retention status and troubleshoot equipment malfunctions at the powder discharge gap location. Classify the operating conditions where powder retention exists but no malfunction occurs at the powder discharge point as powder retention locations. Specifically,

[0177] Using the time point of powder discharge interruption and the corresponding working condition location at the powder discharge gap as the limiting range, the powder discharge state and powder adhesion state in the powder formation state parameters within the powder discharge gap location are distinguished.

[0178] If the same powder formation state parameter contains a powder discharge state field, and the powder discharge state is recorded as any of the states of interrupted discharge, intermittent discharge, or no discharge, and the powder adhesion state is recorded as any of the states of powder retention, powder attachment, or powder accumulation, then the powder discharge gap position is listed as a candidate position for retention. If the powder discharge state is recorded as continuous discharge, or the powder adhesion state is recorded as no adhesion or normal detachment, then the powder discharge gap position is not listed as a candidate position for retention.

[0179] The abnormalities of the candidate locations for stagnation are eliminated. The operating status information of the powder discharge location is formed by the powder discharge valve opening status, powder discharge valve action status, cyclone separator differential pressure status, cyclone separator discharge status and tower wall vibration status in the operation record of the soybean milk spray drying equipment.

[0180] Using the time point of powder discharge interruption and the corresponding operating condition position of the candidate position of stagnation as the exclusion range, check the opening status and operation status of the powder discharge valve. If either the opening status of the powder discharge valve does not reach the powder discharge opening position or the operation status of the powder discharge valve is interrupted, the candidate position of stagnation will be classified as an abnormal position of the powder discharge valve and will not be regarded as a powder stagnation position.

[0181] If the powder discharge valve does not show any abnormalities, continue to check the differential pressure status and discharge status of the cyclone separator. If either the differential pressure status of the cyclone separator exceeds the allowable range for equipment operation or the discharge status of the cyclone separator is interrupted, the candidate position for stagnation will be classified as an abnormal position of the cyclone separator and will not be considered as a powder stagnation position.

[0182] If neither the powder discharge valve nor the cyclone separator shows any abnormality, check the tower wall rapping status. If any of the following conditions are present: rapping not performed, rapping interval exceeds the allowable interval for equipment operation, or powder discharge status has not been restored to continuous discharge after rapping, the candidate position for stagnation is classified as an abnormal position of tower wall rapping and is not considered a position for powder stagnation.

[0183] If the candidate location for stagnation is not classified as an abnormal location of the powder discharge valve, cyclone separator, or tower wall rapping, the candidate location for stagnation will be determined as a stagnation-type powder discharge gap location; the operating condition location of the stagnation-type powder discharge gap location will be classified as a powder stagnation location.

[0184] The allowable ranges for the opening status of the discharge valve, the operation time of the discharge valve, the differential pressure of the cyclone separator, and the interval of the tower wall rapping should preferably be read from the rated operating parameters of the soybean milk spray dryer. If the rated operating parameters of the equipment do not provide allowable ranges, the allowable ranges should be formed using the 5th percentile to the 95th percentile values ​​from at least ten historical fault-free production batches.

[0185] If any of the following situations occur within the powder discharge gap position: the powder discharge valve opening does not reach the lower limit, the powder discharge valve action time exceeds the upper limit, the cyclone separator pressure difference exceeds the upper limit, or the tower wall rapping is not performed within the predetermined interval, the powder discharge gap position will be classified as an abnormal equipment position.

[0186] The abnormal equipment location does not participate in the formation of powder retention location, powder adhesion display location, adhesion feedback location, and droplet adhesion reconnection amount, and outputs equipment inspection marks.

[0187] S3.2.4. Based on the powder retention location, the powder agglomeration state is distinguished item by item in adjacent collection time points to form the visible location of powder adhesion. Specifically,

[0188] After determining the sampling time point for the powder retention location, check the powder agglomeration state and powder discharge state at adjacent sampling time points.

[0189] When the powder agglomeration status shows that agglomerated particles are discharged, and the powder discharge status changes from interruption to discharge, the adjacent collection time point is marked as the agglomeration powder discharge collection time point; the working condition location of the agglomeration powder discharge collection time point and the powder agglomeration status together serve as the external location of powder adhesion.

[0190] S3.2.5. Verify the location of powder retention with the location of powder adhesion, and determine the location that matches the verification as the adhesion display location. Specifically,

[0191] Check whether the working condition position in the powder retention position and the working condition position in the powder adhesion manifestation position are the same, and verify whether the collection time point in the powder retention position is earlier than the collection time point in the powder adhesion manifestation position.

[0192] When the working condition position in the powder retention location is the same as the working condition position in the powder adhesion display location, and the collection time point in the powder retention location is earlier than the collection time point in the powder adhesion display location, the powder retention location and the powder adhesion display location are assigned to the same adhesion display group.

[0193] Within the same adhesion feedback group, the powder retention position indicates that the powder stops at the powder forming condition position, and the powder adhesion external display position indicates that the stopped powder is discharged in an agglomerated powder discharge state. The corresponding working condition positions and agglomerated powder discharge states in the same adhesion feedback group are organized into adhesion feedback positions.

[0194] S3.3, Connect the areas in the droplet's dehydration section where the surface vitrification is not yet complete with the adhesion re-emergence area to form the droplet adhesion re-emergence amount, specifically,

[0195] When a drop of soy milk is placed in the dehydration zone of the droplet, the moisture decreases but the surface stickiness has not yet disappeared, and a stable, dry surface has not yet formed. This is marked as the location where the surface vitrification has not been completed.

[0196] Within the droplet dehydration section, the droplet surface temperature and surface moisture content in the drying environment parameters are determined based on the same batch tracking identifier, and the powder adhesion detection value in the powder formation state parameters is determined based on the same batch tracking identifier.

[0197] During the product manufacturing calibration stage, multiple soybean milk powder samples with known surface moisture content were selected, and the glass transition temperature of each sample was obtained by differential scanning calorimetry. The surface moisture content and glass transition temperature of each sample were arranged from low to high to form a glass transition temperature-surface moisture content calibration curve.

[0198] When the current surface moisture content is the same as a certain calibration surface moisture content in the calibration curve, the corresponding glass transition temperature is read directly; when the current surface moisture content is between two adjacent calibration surface moisture contents, the two adjacent glass transition temperatures are linearly interpolated according to the position ratio of the current surface moisture content between the two adjacent calibration surface moisture contents to form the current glass transition temperature.

[0199] The glass transition moisture content boundary was determined using the 95th percentile of the surface moisture content in historically qualified batches that had formed a stable, dry surface. The droplet adhesion critical value was determined using the 95th percentile of the powder adhesion detection value in historically qualified batches where no powder retention occurred.

[0200] When the droplet surface temperature is higher than the glass transition temperature corresponding to the current surface moisture content, the surface moisture content reaches the glass transition moisture content boundary, and the powder adhesion detection value reaches the droplet adhesion critical value, the corresponding collection time point and the location of the corresponding working condition will be determined as the location where the surface glass transition has not been completed.

[0201] If any of the three conditions is not met, the corresponding data acquisition time point and the associated operating condition location will not be determined as the location where surface vitrification has not been completed. If the detection record is missing or marked as abnormal, the corresponding data acquisition time point will not be used in the judgment.

[0202] If the droplet surface temperature is higher than the glass transition temperature at the corresponding surface moisture content, the surface moisture content is not lower than the glass transition moisture content boundary, and the powder adhesion detection value reaches the droplet adhesion critical value, the operating condition location corresponding to the collection time point is determined as the location where the surface glass transition has not been completed. If any one of the following conditions is met: the droplet surface temperature is not higher than the glass transition temperature at the corresponding surface moisture content, the surface moisture content is lower than the glass transition moisture content boundary, or the powder adhesion detection value does not reach the droplet adhesion critical value, the operating condition location corresponding to the collection time point is not determined as the location where the surface glass transition has not been completed.

[0203] Establish a time-series mapping relationship between the location where surface vitrification is incomplete and the adhesion echo location within the same batch; using the same batch tracking identifier as the mapping boundary, determine the center time point of powder echo by extending the collection time point corresponding to the location where surface vitrification is incomplete by drying to the powder dwell time; the drying to powder dwell time is determined by the duration between the end point of the droplet water loss segment under the same batch tracking identifier and the powder formation time point; use the fluctuation of the powder output time point under the same batch tracking identifier as the window width to form the powder echo time point window, with the window width example taking values ​​of one to two collection cycles.

[0204] If the acquisition time of the adhesion echo position is within the powder formation echo time window, and the adhesion echo position and the position where surface vitrification is incomplete belong to the same batch tracking identifier, the position where surface vitrification is incomplete is identified as the droplet adhesion reconnection position; if the acquisition time of the adhesion echo position is not within the powder formation echo time window, or if the adhesion echo position and the position where surface vitrification is incomplete do not belong to the same batch tracking identifier, the position where surface vitrification is incomplete is not identified as the droplet adhesion reconnection position.

[0205] The number of droplet adhesion reconnection locations was counted, and the number of acquisition time points between each droplet adhesion reconnection location and the adhesion display location was read one by one.

[0206] The number of droplet adhesion reconnection sites indicates the frequency of occurrence at locations where surface vitrification has not been completed. The number of data collection points indicates the hysteresis distance between the location where surface vitrification has not been completed and the powder adhesion reconnection. After organizing the occurrence frequency and hysteresis distance according to their respective operating conditions, the droplet adhesion reconnection amount is obtained.

[0207] It should be noted that the drying environment parameters also include droplet surface temperature and surface moisture content. The droplet surface temperature is recorded by an infrared thermometer placed at the detection location in the drying space, performing non-contact temperature measurement on the soybean milk droplets passing through that location. The surface moisture content is recorded by converting the surface moisture response of the soybean milk droplets using a near-infrared moisture detector placed at the same detection location. The droplet surface temperature and surface moisture content are recorded as drying environment parameters under the same collection time, the same operating condition, and the same batch tracking identifier.

[0208] The powder adhesion state in the powder formation state parameters is recorded numerically using powder adhesion detection values.

[0209] The powder adhesion detection value is collected by a piezoelectric adhesion detection probe installed at the detection position on the drying tower wall or at the powder discharge position. The pressure response, capacitance response or vibration response generated by the powder contacting the detection position is converted into the powder adhesion detection value by the detection equipment and written into the powder formation status parameter under the same batch tracking mark.

[0210] Therefore, the droplet surface temperature and surface moisture content are specific characterization values ​​of the dry environment parameters on the droplet surface, while the powder adhesion detection value is a numerical record of the powder adhesion state among the powder formation state parameters.

[0211] S4. The atomization load control quantity is corrected by the droplet adhesion and reconnection quantity, and the corrected atomization load control quantity is used to constrain the atomization part parameters and drying environment parameters in the next control cycle to form an adaptive control result.

[0212] S4.1, Correct the atomization load control amount using the droplet adhesion and return amount, specifically as follows:

[0213] S4.1.1. The adjustment direction of the atomization load control quantity is corrected based on the operating condition position of the droplet adhesion and reconnection quantity. When the operating condition position of the droplet adhesion and reconnection quantity is in the droplet dehydration section, the adjustment direction of the atomization load control quantity is adjusted to increase the atomization dispersion. When the operating condition position of the droplet adhesion and reconnection quantity is not in the droplet dehydration section, the adjustment direction of the atomization load control quantity is maintained, forming the corrected adjustment direction. Specifically...

[0214] When the operating condition of the droplet adhesion reconnection amount is in the droplet dehydration section, it means that the powder adhesion can be traced back to the position where the surface vitrification was not completed during the droplet dehydration process. The adjustment direction of the atomization load control amount should be adjusted to the direction of increasing the atomization dispersion amount.

[0215] The increase in atomized dispersion can be achieved by increasing atomization pressure, increasing atomization speed, and decreasing atomization flow rate, thereby shifting the coarse droplet concentration segment in the droplet size spectrum towards the fine droplet distribution segment.

[0216] When the operating condition of the droplet adhesion return quantity is not in the droplet water loss section, maintain the original adjustment direction of the atomization load control quantity.

[0217] When the droplet size spectrum does not show a concentrated segment of coarse droplets and the exhaust humidity exceeds the upper boundary of the exhaust humidity of historical qualified batches, the adjustment direction of the atomization load control amount is maintained, and an exhaust humidity correction mark is generated. The exhaust humidity correction mark is used to correct the drying environment parameters in the next control cycle.

[0218] The adjustment direction in the atomization load control is retained. The working condition position and droplet adhesion reconnection position in the droplet adhesion reconnection amount are used as the correction objects. The droplet adhesion reconnection position is mapped to the droplet water loss section, droplet size spectrum, adhesion reconnection position and drying environment parameters that have been formed under the same batch tracking mark to determine the adhesion cause.

[0219] When the operating condition position in the droplet adhesion recovery range and the operating condition position in the droplet water loss range are in the same location segment, the process enters the stage of adhesion cause differentiation, specifically:

[0220] When the droplet adhesion re-attachment location is in the coarse droplet concentration segment of the droplet size spectrum, the adhesion cause is determined to be due to the coarse droplets not being dried, and the adjustment direction in the atomization load control is corrected to the direction of increasing atomization dispersion. When the droplet adhesion re-attachment location is in the fine droplet distribution segment of the droplet size spectrum, and the adhesion re-appearance location is the powder adhesion external appearance location, the adhesion cause is determined to be due to the fine powder re-adhesion, and the adjustment direction in the atomization load control is corrected to the direction of decreasing atomization dispersion. When the droplet size spectrum does not point to the abnormal distribution of the coarse droplet concentration segment and the fine droplet distribution segment, and the exhaust humidity in the drying environment parameters exceeds the exhaust humidity boundary of the historical qualified batches, the adhesion cause is determined to be due to the exhaust humidity being too high, the adjustment direction in the atomization load control remains unchanged, and the cause of the exhaust humidity being too high is used as the basis for correcting the drying environment parameters in the next control cycle.

[0221] When the operating condition position in the droplet adhesion return quantity is not in the operating position limited by the droplet water loss section, the adjustment direction in the atomization load control quantity continues in the original direction.

[0222] When both coarse droplet undried cause and fine powder re-adhesion cause appear simultaneously under the same batch tracking identifier, the number of droplet adhesion re-attachment locations in the coarse droplet concentration section and the number of droplet adhesion re-attachment locations in the fine droplet distribution section are sorted in descending order. The adhesion cause with the sorting number 1 is used as the basis for adjustment direction correction. When the sorting number of coarse droplet undried cause is 1, the adjustment direction is corrected to the direction of increasing atomization dispersion. When the sorting number of fine powder re-adhesion cause is 1, the adjustment direction is corrected to the direction of decreasing atomization dispersion. When the two are the same, the original adjustment direction in the atomization load control is maintained.

[0223] S4.1.2. Convert the droplet adhesion and reconnection amount into an amplitude correction amount according to the corresponding operating condition and position. The formula is as follows:

[0224] ;

[0225] In the formula, Indicates the amplitude correction amount. Indicates the amplitude conversion factor. This indicates the number of droplet adhesion re-attachment sites. This indicates the time interval between the droplet adhesion reconnection position and the adhesion display position.

[0226] The adjustment amplitude of the atomization load control quantity is calibrated using the amplitude correction amount to obtain the corrected atomization load adjustment amplitude, as shown in the formula.

[0227] ;

[0228] In the formula, Indicates the difference in atomization load. This indicates the adjusted atomization load range.

[0229] Number of droplet adhesion re-attachment sites When increased, amplitude correction amount Increase; data collection time interval When increased, amplitude correction amount Reduce; Amplitude conversion factor The single adjustment step distance is set by the parameters of the atomization part. In the example, the value is one percent of the single atomization pressure adjustment step distance.

[0230] Amplitude conversion factor It is a dimensionless numerical value used to convert the ratio of the number of droplet adhesion reconnection sites to the sampling time interval into a numerical scale of atomization load difference.

[0231] When the spray drying equipment is operating stably and the actual powder formation reaches the qualified powder formation state, a minimum adjustment step is executed for the atomization pressure, atomization speed, and atomization flow rate, while keeping the other atomization parameters and drying environment parameters unchanged. The change range of atomization dispersion in one control cycle is recorded. The minimum adjustment step is normalized to the corresponding equipment allowable range, and the relationship between the change range of atomization dispersion and the normalized adjustment step is used to form a single amplitude conversion value. The median value of the single amplitude conversion value obtained from no less than three single-parameter step tests is used as the amplitude conversion coefficient.

[0232] When the nozzle is replaced, the atomizer is maintained, or the soybean milk liquid formula changes, the single-parameter step test should be repeated.

[0233] S4.2. Using the corrected atomization load control quantity to constrain the atomization part parameters and drying environment parameters for the next control cycle, an adaptive control result is formed, specifically,

[0234] S4.2.1, Replace the atomization dispersion adjustment direction of the atomization part parameters in the next control cycle with the corrected adjustment direction, specifically,

[0235] When the corrected adjustment direction is the direction of increasing atomization dispersion, the atomization parameters for the next control cycle are processed within the allowable ranges of atomization pressure, atomization speed, and atomization flow rate. The allowable range of atomization pressure is jointly determined by the nozzle rated pressure boundary and the pressure pump operating boundary; the allowable range of atomization speed is determined by the atomizer rated speed boundary; and the allowable range of atomization flow rate is jointly determined by the feed pump stable feeding boundary and the nozzle anti-clogging boundary.

[0236] The single-cycle adjustment step is determined based on the corrected atomization load adjustment range. The atomization pressure and atomization speed are adjusted upward according to the single-cycle adjustment step, while the atomization flow rate is adjusted downward according to the single-cycle adjustment step. The adjusted atomization pressure must not exceed the upper boundary of the allowable atomization pressure range, the adjusted atomization speed must not exceed the upper boundary of the allowable atomization speed range, and the adjusted atomization flow rate must not be lower than the lower boundary of the allowable atomization flow rate range. After the boundary limits are met, the enhanced dispersion state of the atomization parameters in the next control cycle is formed.

[0237] When the corrected adjustment direction is to decrease the atomization dispersion, the atomization pressure and atomization speed in the atomization parameters of the next control cycle will decrease, while the atomization flow rate will increase, causing the atomization parameters of the next control cycle to enter a weakened dispersion state.

[0238] The corrected adjustment direction is to maintain the current state, and the atomization parameters in the next control cycle will continue the atomization dispersion adjustment trend of the current control cycle.

[0239] S4.2.2, By limiting the adjustment range of the atomization part parameters in the next control cycle through the corrected atomization load adjustment range, the number of incomplete surface vitrification locations formed in the next control cycle is obtained and compared with the number of incomplete surface vitrification locations formed in the current control cycle. The drying environment parameters for the next control cycle are then corrected in the direction of reducing the number of incomplete surface vitrification locations. Specifically,

[0240] Once the corrected atomization load adjustment range is determined, it will be allocated to the atomization pressure adjustment, atomization speed adjustment, and atomization flow rate adjustment in the next control cycle.

[0241] When the adjusted atomization load range increases, the adjustment range of atomization pressure and atomization speed in the next control cycle is increased, while the adjustment range of atomization flow rate in the next control cycle is decreased.

[0242] When the adjusted atomization load decreases, the adjustment amount of atomization pressure and atomization speed in the next control cycle is reduced, while the adjustment amount of atomization flow rate in the next control cycle is increased.

[0243] The number of locations in the droplet dehydration section that have not completed surface vitrification is counted and compared with the number of locations in the droplet dehydration section of the previous control cycle. When the number of locations that have not completed surface vitrification decreases, the adjustment status of drying temperature, air intake dehumidification rate, exhaust rate and feed rate is maintained for the next control cycle.

[0244] When the number of incomplete surface vitrification sites decreases in the next control cycle, maintain the atomization site parameters and drying environment parameters for the next control cycle.

[0245] If the number of locations where surface vitrification has not been completed does not decrease, adjust the drying environment parameters for the next control cycle in the order of exhaust volume, inlet dehumidification volume, drying temperature, and feed volume.

[0246] First, increase the exhaust volume by one minimum adjustment step, while keeping the intake dehumidification volume, drying temperature, and feed volume unchanged; if the exhaust humidity still exceeds the upper limit of the exhaust humidity of historical qualified batches in the next control cycle, increase the intake dehumidification volume by one minimum adjustment step.

[0247] If the exhaust humidity falls within the historical acceptable range for batches, but the number of areas where surface vitrification has not been completed has not decreased, increase the drying temperature by a minimum adjustment step. Stop increasing the drying temperature when the powder moisture content reaches the lower boundary of the target moisture content range.

[0248] If, after adjusting the exhaust volume, inlet dehumidification volume, and drying temperature, the number of locations where surface vitrification has not been completed does not decrease for two consecutive control cycles, the feed rate is reduced by a minimum adjustment step. Once the feed rate reaches the lower boundary of the stable feeding by the feed pump, the reduction in feed rate is stopped.

[0249] If the number of locations where surface vitrification has not been completed is less than that of the previous control cycle, the current control parameters are maintained and the adjustment range is not further increased; if the number of locations where surface vitrification has not been completed does not decrease for three consecutive control cycles, the automatic increase of the adjustment range is stopped, and inspection information such as nozzle blockage, detection probe abnormality, powder discharge location abnormality, and material liquid status abnormality is output.

[0250] When any parameter of the atomization part or the drying environment parameter reaches the allowable boundary of the equipment operation, the boundary value of the corresponding parameter shall be maintained.

[0251] When the atomization pressure, atomization speed, or atomization flow rate reaches the allowable operating boundary of the equipment, stop adjusting the parameters of the atomization part that has reached the allowable operating boundary, and continue adjusting the parameters of other atomization parts that are still within the allowable operating boundary of the equipment according to the corrected adjustment direction.

[0252] When the exhaust volume, inlet dehumidification volume, drying temperature, or feed rate reach the allowable operating boundary of the equipment, the adjustment of the drying environment parameters that have reached the allowable operating boundary of the equipment is stopped, and the next parameter in the sequence of exhaust volume, inlet dehumidification volume, drying temperature, and feed rate is entered; when all parameters reach the allowable operating boundary of the equipment, the operating boundary information is output.

[0253] The adjustment results of the atomization part parameters in the next control cycle and the correction results of the drying environment parameters in the next control cycle are used together as the adaptive control results.

[0254] In this embodiment,

[0255] Select a soy milk powder spray drying production line, set the control cycle to 60 seconds, and the parameter acquisition interval to 1 second. The target range for the moisture content of the powder formed from historical qualified batches is 3.5% to 4.5%, the target range for the particle size is 90 micrometers to 140 micrometers, the target boundary for the powder flowability is a Hausner ratio of no more than 1.25, and the allowable agglomeration level is slight agglomeration.

[0256] The actual powder moisture content collected during the current control cycle is 5.1%, the median particle size is 162 micrometers, the Haussner ratio is 1.31, and there is intermittent discharge. Based on the normalized deviation results of each powder formation parameter relative to the qualified powder formation state, the powder moisture content has the largest deviation after weight correction, and the deviation direction is determined to be the moistening direction.

[0257] Between the eighth and eighteenth sampling points in the current control cycle, the soybean milk concentration increased by 0.8% compared to the historical qualified batches, the liquid viscosity increased by 12 mPa·s, and the liquid flow rate decreased by 6%. The corresponding operating condition position of the sampling point was determined as the feed blockage position, and the liquid load was formed based on the normalized deviation of the soybean milk concentration, liquid viscosity, and liquid flow rate.

[0258] The droplet size spectrum formed by the laser particle size detection probe shows a droplet dispersion breakpoint at the 160-micrometer position. The number of droplets in the fine droplet distribution segment at the front of the droplet dispersion breakpoint accounts for 34% of the total number of droplets, and the atomization dispersion amount is taken as 0.34%. When the feed load is higher than the atomization dispersion amount, the atomization load difference is positive, and the adjustment direction of the atomization load control amount is determined to be the direction of increasing the atomization dispersion amount.

[0259] At the 17th sampling point of the current control cycle, the ambient humidity increased and the drying temperature decreased simultaneously. At the 35th sampling point, the increase in ambient humidity and the decrease in drying temperature decreased simultaneously. The period from the 17th to the 35th sampling point is defined as the droplet water loss segment.

[0260] Based on the glass transition temperature-surface moisture content calibration curve, the nineteenth, twenty-fourth, twenty-eighth, and thirty-first sampling points in the droplet dehydration segment were identified as locations where surface vitrification was incomplete. The drying-to-powdering residence time for the current batch was forty-one seconds, and the powder formation feedback window width was two sampling intervals. Three adhesion feedback positions were obtained within the corresponding powder formation feedback window, forming three droplet adhesion re-attachment positions.

[0261] The operating condition corresponding to the droplet adhesion and reconnection amount is in the droplet dehydration section. Therefore, the adjustment direction of the atomization load control amount is corrected to the direction of increasing the atomization dispersion. Based on the amplitude conversion coefficient and the number of droplet adhesion and reconnection locations, the atomization pressure for the next control cycle is increased by 0.03 MPa, the atomization speed is increased by 40 rpm, and the atomization flow rate is decreased by 0.05 liters per minute.

[0262] In the next control cycle, the number of incomplete surface vitrification locations is reduced from four to two, the powder moisture content is reduced to 4.4 percent, and the median powder particle size is reduced to 138 micrometers. All of these are within the allowable range for qualified powder formation. The current adjustment state is maintained, and the adjustment range is not further increased.

[0263] In actual production, the control cycle, detection boundary, and adjustment step are determined based on the rated parameters of the spray drying equipment, the target soy milk powder product specifications, and historical production records.

[0264] This embodiment also provides a computer device applicable to the adaptive control method for soybean milk spray drying based on multi-parameter monitoring, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the adaptive control method for soybean milk spray drying based on multi-parameter monitoring as proposed in the above embodiment.

[0265] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0266] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the adaptive control method for soybean milk spray drying based on multi-parameter monitoring as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0267] In summary, this invention achieves reverse localization between the adhesion formation location and abnormal powder formation phenomenon during the spray drying of soy milk by isotopically correlating the droplet dehydration state with the soy milk powder retention. This effectively distinguishes between powder adhesion problems caused by insufficient surface solidification of droplets and powder formation deviations caused by fluctuations in normal operating conditions. It reduces control lag caused by adjustments based on the final powder formation result, avoids the continuous accumulation of local drying imbalances during the droplet heating and dehydration process and their transmission to subsequent powder formation stages, enhances the traceability of the source of abnormalities and the reliability of adjustment basis, improves the accuracy, continuity and stability of process control, and helps maintain the consistency of the powder formation state during the spray drying of soy milk.

[0268] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. An adaptive control method for soybean milk spray drying based on multi-parameter monitoring, characterized in that: include, Collect parameters of soybean milk liquid, atomization part, drying environment and powder state during the same control cycle of soybean milk spray drying, and perform state calibration to generate a drying condition state table. Define the qualified powdering state, and find the deviation direction and deviation magnitude of the actual powdering state relative to the qualified powdering state from the drying condition state table. Then, decompose the atomization load of the soybean milk liquid parameters and the atomization part parameters to determine the liquid load when the soybean milk enters the atomization part and the atomization dispersion of the soybean milk by the atomization part. Finally, perform the same-scale difference calculation on the liquid load and the atomization dispersion to form the atomization load control quantity. Locate the droplet dehydration section from the drying environment parameters, and track the adhesion and re-emergence position of the soybean milk powder based on the powder state parameters. Connect the position in the droplet dehydration section where the surface vitrification has not been completed with the adhesion and re-emergence position to form the droplet adhesion and re-emergence amount. The atomization load control quantity is corrected by using the droplet adhesion and reconnection amount, and the corrected atomization load control quantity is used to constrain the atomization part parameters and drying environment parameters in the next control cycle to form an adaptive control result.

2. The adaptive control method for soybean milk spray drying based on multi-parameter monitoring as described in claim 1, characterized in that: The state calibration refers to labeling the source, collection time, and corresponding working condition location of the soybean milk liquid parameters, atomization part parameters, drying environment parameters, and powdering state parameters within the same control cycle.

3. The adaptive control method for soybean milk spray drying based on multi-parameter monitoring as described in claim 1, characterized in that: The steps for defining a qualified powder-forming state and finding the direction and magnitude of the deviation between the actual powder-forming state and the qualified powder-forming state from the drying condition status table are as follows: Extract the powder formation state parameters that indicate the powder formed during the spray drying process of soybean milk reaches the preset powder formation quality conditions from the drying condition state table, and define them as qualified powder formation state. Extract the actual powder formation state from the powder formation state parameters. By comparing the qualified powder state with the actual powder state, the deviation direction of the actual powder state relative to the qualified powder state is determined as the deviation direction, and the difference between the actual powder state and the qualified powder state is determined as the deviation amplitude.

4. The adaptive control method for soybean milk spray drying based on multi-parameter monitoring as described in claim 1, characterized in that: The specific steps for determining the feed volume of the soy milk when it enters the atomization section and the atomization dispersion amount of the soy milk by the atomization section are as follows: Track the feed blockage location in the soybean milk feed parameters that corresponds to the direction of deviation; The load strength value is determined by the deviation range and the degree of deviation of the soybean milk liquid parameters at the feed stagnation position, and the load strength value is converted into the liquid load when the soybean milk enters the atomization part. The dispersion change direction of the atomization part parameters is limited by the deviation direction, and the atomization part parameters are converted into a droplet size spectrum. The droplet size spectrum is divided into a fine droplet distribution segment and a coarse droplet concentration segment, and the droplet dispersion breakpoint is located by the degree of aggregation of the coarse droplet concentration segment in the droplet size spectrum. By using the droplet dispersion breakpoint as the dividing position, the continuous distribution state of the fine droplet distribution segment and the aggregation state of the coarse droplet concentration segment are distinguished, and the atomization dispersion amount is obtained.

5. The adaptive control method for soybean milk spray drying based on multi-parameter monitoring as described in claim 1, characterized in that: The specific steps for calculating the atomization load control value by performing a same-scale difference calculation between the feed liquid load and the atomization dispersion amount are as follows: The difference between the feed load and the atomization dispersion is calculated on the same scale to obtain the atomization load difference. The atomization load control quantity is obtained by constraining the adjustment direction and adjustment amplitude by the positive and negative directions and absolute magnitude of the atomization load difference.

6. The adaptive control method for soybean milk spray drying based on multi-parameter monitoring as described in claim 1, characterized in that: The specific steps for locating the water loss segment of the droplet from the drying environment parameters are as follows: According to the order of the collection time points, the rise and fall trends of humidity and temperature changes in the dry environment parameters were checked one by one, and the collection time point when the humidity change changed from flat to rising and the temperature change decreased synchronously was taken as the starting point of droplet water loss. Starting from the point where the droplet loses water, continue to track adjacent collection points and take the point where the rate of increase in humidity change begins to decrease and the rate of decrease in temperature change begins to decrease synchronously as the point where the droplet loses water. The sampling time points and corresponding operating conditions within the range from the droplet water loss initiation point to the droplet water loss termination point are defined as the droplet water loss segment.

7. The adaptive control method for soybean milk spray drying based on multi-parameter monitoring as described in claim 1, characterized in that: The specific steps for tracking the adhesion and display position of soybean milk powder based on powder state parameters are as follows: In the powder formation status parameters, the actual powder formation status is statistically analyzed item by item according to the corresponding working condition location, and the collection time points of continuous discharge in the actual powder formation status are connected into normal powder output segments. Using the end and beginning of adjacent normal powder-producing segments as boundaries, the collection points within the boundaries that are not covered by normal powder-producing segments are extracted to form the powder-producing gap locations. The powder discharge gap location is assessed for stagnation and equipment malfunctions are eliminated. The working condition location where powder stagnation occurs but no malfunction occurs at the powder discharge location is classified as a powder stagnation location. Based on the location of powder retention, the powder agglomeration state is distinguished item by item in adjacent collection time points to form the visible location of powder adhesion; The positions of powder retention and powder adhesion are compared, and the positions that match are determined as the adhesion display positions.

8. The adaptive control method for soybean milk spray drying based on multi-parameter monitoring as described in claim 1, characterized in that: The incomplete surface vitrification refers to the state in which the moisture content of the soybean milk droplet decreases within the water loss zone, but the surface viscosity has not yet disappeared, and a stable, dry surface layer has not yet been formed.

9. The adaptive control method for soybean milk spray drying based on multi-parameter monitoring as described in claim 1, characterized in that: The specific steps for correcting the atomization load control amount using droplet adhesion and reconnection amount are as follows: The adjustment direction of the atomization load control quantity is corrected by using the operating condition position of the droplet adhesion return quantity; When the operating condition of the droplet adhesion and reconnection amount is in the droplet dehydration section, the adjustment direction of the atomization load control amount is adjusted to the direction of increasing the atomization dispersion amount. When the operating condition of the droplet adhesion and reconnection amount is not in the droplet dehydration section, the adjustment direction of the atomization load control amount is maintained, thus forming the corrected adjustment direction. The amount of droplet adhesion and reconnection is converted into an amplitude correction amount according to the corresponding operating condition position, and the amplitude correction amount is used to calibrate the adjustment amplitude of the atomization load control amount to obtain the corrected atomization load adjustment amplitude.

10. The adaptive control method for soybean milk spray drying based on multi-parameter monitoring as described in claim 1, characterized in that: The specific steps for constraining the atomization parameters and drying environment parameters in the next control cycle are as follows: The atomization dispersion adjustment direction of the atomization part parameter in the next control cycle is replaced with the corrected adjustment direction. The adjustment range of the atomization part parameters in the next control cycle is limited by the modified atomization load adjustment range; Obtain the number of locations where incomplete surface vitrification is formed in the next control cycle, and compare it with the number of locations where incomplete surface vitrification is formed in the current control cycle; When the number of incomplete surface vitrification sites decreases in the next control cycle, maintain the atomization site parameters and drying environment parameters for the next control cycle. If the number of incomplete surface vitrification locations formed in the next control cycle does not decrease, the drying environment parameters for the next control cycle are corrected in the direction of reducing the number of incomplete surface vitrification locations to form an adaptive control result.

Images