An intelligent management and scheduling system and method for a milk powder production workshop

Abstract

The present invention relates to the technical field of food production supervision, specifically to an intelligent management and scheduling system and method for a milk powder production workshop; the system includes: a data acquisition and processing unit that real-time collects milk slurry data, workshop environment data, and equipment operation data; a milk slurry drying monitoring unit that, based on the milk slurry data and workshop environment data, uses a milk slurry multi-variable stepwise concentration adjustment algorithm and combines it with a dynamic air flow control algorithm for a spray drying tower to jointly adjust the milk slurry concentration in real time, optimize the air flow distribution of the spray drying tower, and equipment operation data; a lactose crystallization monitoring unit that real-time monitors the lactose crystallization trend; an information management and traceability unit that collects and stores all data during the milk powder production process and conducts data report analysis. The present invention realizes the dynamic coupling adjustment of the milk slurry concentration and drying air flow by introducing a multi-variable stepwise concentration adjustment algorithm and a dynamic air flow control algorithm for a spray drying tower, and solves the problems of uneven spraying of high-viscosity milk slurry and out-of-control lactose crystallization.

Description

Technical Field

[0001] The present invention relates to the technical field of food production supervision, and in particular to an intelligent management and scheduling system and method for a milk powder production workshop. Background Art

[0002] In the milk powder production process, the spray drying process is the key link between liquid milk slurry and solid milk powder products, which has a decisive influence on the solubility, particle size distribution, lactose crystallization state and overall quality of milk powder. In current industrial practice, milk powder manufacturers generally use a variety of controllers and PLC systems to achieve single-point adjustment of equipment such as milk slurry concentration, sprayer operating parameters, and fan speed, and there are still major deficiencies in milk powder drying coordination and intelligent scheduling;

[0003] In the actual milk powder production workshop environment, the drying stability of milk slurry is synergistically affected by milk slurry concentration, milk slurry viscosity, temperature and lactose content. High-viscosity milk slurry usually first affects the fluidity and spraying effect of milk slurry during the spray drying process. If the viscosity of the milk slurry is too high, the milk slurry will become difficult to spray through the nozzle, and the uneven spraying effect will make the temperature and humidity in the spray tower and the air flow become uneven, forming a local airflow dead zone or eddy zone in the spray tower. In the later stage of drying, the crystallization rate of lactose is directly affected by the temperature and humidity, milk slurry viscosity and air flow in the drying tower. Due to the influence of uneven air flow and uneven temperature and humidity, some milk slurry droplets may not be fully dried, resulting in excessive humidity, lactose begins to crystallize and precipitate and form crystals in the particles. High-viscosity milk slurry will reduce the efficiency of lactose dissolution, making lactose easier to crystallize.

[0004] High-concentration milk slurry will affect the uneven temperature and humidity distribution and air flow stability in the drying tower, thereby affecting the evaporation rate of the milk slurry. Ultimately, the high-concentration milk slurry in some areas of the drying tower cannot be fully dried and will precipitate to form large particles. Summary of the invention

[0005] The purpose of the present invention is to provide an intelligent management and scheduling system and method for a milk powder production workshop, so as to solve the problem raised in the above background technology that high-concentration milk slurry will affect the uneven temperature and humidity distribution and air flow stability in the drying tower, thereby affecting the evaporation rate of the milk slurry, and ultimately causing the high-concentration milk slurry in some areas of the drying tower to fail to be fully dried and will precipitate to form large particles.

[0006] To achieve the above object, the present invention provides an intelligent management and scheduling system for a milk powder production workshop, comprising:

[0007] Data acquisition and processing unit: The data acquisition and processing unit uses sensors and monitoring equipment to collect milk slurry data, workshop environment data, and equipment operation data in real time, and performs preliminary processing on all collected data;

[0008] The milk slurry drying monitoring unit, based on milk slurry data and workshop environment data, uses the milk slurry multi-fractional variable-order concentration adjustment algorithm and combines it with the spray drying tower dynamic air flow control algorithm to jointly and real-time adjust the milk slurry concentration, optimize the air flow distribution in the spray drying tower, and equipment operation data;

[0009] The milk slurry drying monitoring unit includes a milk slurry concentration adjustment module and a drying tower air flow optimization module;

[0010] Among them, the milk slurry concentration adjustment module, based on milk slurry data and workshop environment data, uses the milk slurry multi-fractional variable-order concentration adjustment algorithm to combine and real-time adjust the milk slurry concentration before drying, and outputs the adjusted milk slurry concentration and the milk slurry concentration control residual to the drying tower air flow optimization module;

[0011] The drying tower air flow optimization module, based on the adjusted milk slurry concentration, workshop environment data, and equipment operation data, uses the spray drying tower dynamic air flow control algorithm to optimize the air flow distribution in the spray drying tower;

[0012] The lactose crystallization monitoring unit is used to real-time monitor the lactose crystallization trend;

[0013] The information management and traceability unit is used to collect and store all data during the milk powder production process and perform data report analysis.

[0014] Preferably, the milk slurry data includes: milk slurry viscosity , milk slurry temperature , milk slurry concentration and lactose content ;

[0015] The workshop environment data includes: workshop temperature , workshop humidity and air flow rate ;

[0016] The equipment operation data includes: the fan speed of the spray drying tower and injection pressure ;

[0017] Among them, is time.

[0018] Preferably, the milk slurry multi-fractional variable-order concentration adjustment algorithm is implemented based on the multi-fractal dynamic fluctuation modeling technology and the variable-order fractional differential response mechanism, and is used to real-time dynamically adjust the milk slurry concentration before drying. The specific method of the milk slurry multi-fractional variable-order concentration adjustment algorithm is as follows:

[0019] S2.1.1. Input the milk slurry data and workshop environment data into the variable-order differential neural operator, calculate the quasi-variable-order quasi-fractal derivatives of all milk slurry data and workshop environment data , and combine and embed all the quasi-variable-order quasi-fractal derivatives to obtain a variable-order differential feature ;

[0020] Among them, in the quasi-variable-order quasi-fractal derivative , is the variable-order exponent, is the multi-scale quasi-fractal dimension, is the time; is the quasi-variable-order quasi-fractal derivative, used to represent the evolution trend of the milk slurry data and workshop environment data with the variable-order exponent and multi-scale quasi-fractal dimension characteristics in the time series;

[0021] S2.1.2. Use the Moran multifractal mapping to perform non-linear fusion on the variable-order differential feature , and calculate to obtain the milk slurry fluctuation intensity index ;

[0022] S2.1.3. Calculate the automatic adjustment control speed of the milk slurry based on the workshop environment data ;

[0023] S2.1.4. Use the variable-order Caputo derivative model to construct an adaptive milk slurry concentration control equation;

[0024] S2.1.5. According to the prediction result of the adaptive milk slurry concentration control equation, adjust the actual milk slurry concentration to the target milk slurry concentration in real time, and output the adjusted milk slurry concentration and the milk slurry concentration control residual , is an increment of a time period.

[0025] Preferably, in the S2.1.4, the adaptive milk slurry concentration control equation is the core equation of the milk slurry multi-fraction variable-order concentration adjustment algorithm, constructed based on the variable-order Caputo derivative model, and used to dynamically predict the milk slurry concentration trend. The specific form of the adaptive milk slurry concentration control equation is as follows:

[0026] ;

[0027] Among them, is the Caputo variable-order fractional derivative, representing the dynamic change rate of the milk slurry concentration at time ; is the automatic adjustment control speed of the milk slurry; is the milk slurry concentration; is at time The target milk slurry concentration; is the control gain coefficient; is the milk slurry fluctuation intensity index; is the fractal excitation gain coefficient.

[0028] Preferably, the dynamic air flow control algorithm of the spray drying tower is an air flow control algorithm based on persistent homology topological analysis and discrete differential manifold modeling, and is used to regulate the air flow distribution in the spray drying tower. The specific dynamic air flow control algorithm of the spray drying tower is as follows:

[0029] S2.2.1. Input the adjusted milk slurry concentration as well as the workshop environment data and equipment operation data into the lightweight CFD neural operator to construct a three-dimensional velocity vector field in the spray drying tower space ;

[0030] S2.2.2. Conduct persistent homology analysis on the three-dimensional velocity vector field and calculate the topological flux entropy ;

[0031] S2.2.3. According to the topological flux entropy , the adjusted milk slurry concentration , the lactose content as well as the milk slurry concentration control residual , calculate the adjusted fan speed and nozzle pressure;

[0032] S2.2.4. Based on the topological flux entropy and the workshop environment data, weighted update the automatic adjustment control speed of the milk slurry .

[0033] Preferably, in S2.2.3, according to the topological flux entropy , the adjusted milk slurry concentration , the lactose content as well as the milk slurry concentration control residual , calculate the adjusted fan speed and nozzle pressure. The specific method is as follows:

[0034] ;

[0035] ;

[0036] wherein, is the fan speed at the adjusted time ; is the nozzle pressure at the adjusted time ; is the weight of the fan change influence; is the weight of the milk slurry concentration deviation; is the residual coupling weight of the fan concentration adjustment; is the influence weight of the nozzle change; is the lactose deviation weight; is the residual coupling weight of the nozzle concentration adjustment; is the change rate of the topological flux entropy; is the adjusted milk slurry concentration; is time of the target milk slurry concentration; time of the lactose content; is time of the target lactose content.

[0037] Preferably, the lactose crystallization monitoring unit is used to monitor the lactose crystallization trend in real time, specifically as follows:

[0038] Set the lactose content critical interval as , the critical milk slurry concentration fluctuation threshold and the critical milk slurry temperature fluctuation threshold ;

[0039] At time if the current lactose content is within the lactose content critical interval, the current milk slurry concentration fluctuation rate is greater than the critical milk slurry concentration fluctuation threshold, and the current milk slurry temperature fluctuation rate is greater than the critical milk slurry temperature fluctuation threshold, and the continuous duration exceeds the manually set duration, the production work shall be immediately stopped.

[0040] Preferably, the information management and traceability unit is used to collect and store all data during the milk powder production process, including the original milk slurry data, workshop environment data and equipment operation data in the data acquisition and processing unit, and also including the adjusted milk slurry concentration, milk slurry concentration control residual, topological flux entropy and milk slurry automatic adjustment control speed in the milk slurry drying monitoring unit, and also including the set lactose content critical interval, critical milk slurry concentration fluctuation threshold and critical milk slurry temperature fluctuation threshold in the lactose crystallization monitoring unit.

[0041] On the other hand, the present invention provides a method for intelligent management and scheduling of a milk powder production workshop, which is used for the above-mentioned intelligent management and scheduling system of a milk powder production workshop, and includes the following steps:

[0042] S10.1. Use sensors and monitoring devices to collect milk slurry data, workshop environment data and equipment operation data in real time, and perform preliminary processing on all the collected data;

[0043] S10.2. Based on the milk slurry data and the workshop environment data, the milk slurry multi-variable stepwise concentration adjustment algorithm is adopted and combined with the dynamic air flow control algorithm of the spray drying tower to jointly and real-time adjust the milk slurry concentration, optimize the air flow distribution of the spray drying tower, and the equipment operation data;

[0044] S10.3. Real-time monitor the lactose crystallization trend;

[0045] S10.4. Collect and store all data during the milk powder production process, and conduct data report analysis.

[0046] Compared with the prior art, the above technical solutions of the present invention have the following beneficial technical effects:

[0047] 1. In the present invention, through the milk slurry multi-variable stepwise concentration adjustment algorithm, the influence of the milk slurry viscosity, temperature and lactose content on the milk slurry concentration is sensed in real time, and the milk slurry concentration is adaptively adjusted under different workshop environment parameters, realizing high-precision and low-latency adjustment of the milk slurry concentration, thereby solving the problem of uneven spray distribution caused by poor fluidity of the high-viscosity milk slurry at the initial stage of drying;

[0048] 2. In the present invention, through the dynamic air flow control algorithm of the spray drying tower, the real-time quantification and control prediction adjustment of the complexity of the air flow topological structure in the spray drying tower are realized, and the milk slurry concentration control residual error is dynamically coupled with the topological flux entropy, so that the fan speed and the nozzle pressure can be adjusted in real-time linkage with the physical state of the milk slurry, alleviating the problems of non-uniform crystallization and large particle precipitation formed by the high-concentration milk slurry at the end stage of drying. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 is a principle block diagram of an embodiment proposed by the present invention;

[0050] Reference numerals: 1, data acquisition and processing unit; 2, milk slurry drying monitoring unit; 21, milk slurry concentration adjustment module; 22, drying tower air flow optimization module; 3, lactose crystallization monitoring unit; 4, information management and traceability unit. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0051] Embodiment 1, as Figure 1 shown, provides an intelligent management and scheduling system for a milk powder production workshop, including:

[0052] A data acquisition and processing unit 1, which uses sensors and monitoring devices to collect milk slurry data, workshop environment data and equipment operation data in real time, and preliminarily processes all the collected data;

[0053] The milk slurry data includes: milk slurry viscosity , milk slurry temperature , milk slurry concentration and lactose content ;

[0054] The workshop environmental data includes: workshop temperature , workshop humidity and air flow rate ;

[0055] The equipment operation data includes: the fan speed of the spray drying tower and injection pressure ;

[0056] Among them, is time;

[0057] In this embodiment, a temperature sensor is installed in front of the milk slurry conveying pipeline and the spray drying tower to monitor and collect the milk slurry temperature in real time, a viscosity sensor is used to monitor and collect the milk slurry viscosity, and a lactose content sensor is used to monitor the lactose content in the milk slurry in real time; then a temperature and humidity sensor is installed at multiple positions in the workshop to monitor the environmental temperature and humidity in the workshop in real time; an air flow velocity sensor is installed in the drying tower to monitor the air flow rate in real time to ensure uniform air flow distribution in the spray drying tower; a fan operating state sensor is used to monitor the working state of the spray drying tower fan, including wind speed and rotation speed; all sensors and monitoring devices upload the collected data to the PLC system in real time, and the data collection period is 50 ms;

[0058] The noise of the collected data is removed through a filtering algorithm, and the collected data is checked in real time, outliers are automatically identified and removed, outlier detection is performed through statistical algorithms and threshold determination to ensure that the subsequent used data is not affected by outliers, and the collected data is standardized;

[0059] The milk slurry viscosity directly affects the spray drying effect, the milk slurry temperature affects the lactose solubility and drying efficiency, and the lactose content affects the lactose solubility and crystallization situation.

[0060] The milk slurry drying monitoring unit 2, based on the milk slurry data and workshop environmental data, uses the milk slurry multi-variable order concentration adjustment algorithm and combines it with the spray drying tower dynamic air flow control algorithm to jointly adjust the milk slurry concentration in real time and optimize the air flow distribution of the spray drying tower and the equipment operation data;

[0061] The milk slurry drying monitoring unit 2 includes a milk slurry concentration adjustment module 21 and a drying tower air flow optimization module 22;

[0062] Among them, the milk slurry concentration adjustment module 21, based on the milk slurry data and workshop environmental data, uses the milk slurry multi-variable order concentration adjustment algorithm to combine and adjust the milk slurry concentration before drying in real time, and outputs the adjusted milk slurry concentration and the milk slurry concentration control residual to the drying tower air flow optimization module 22;

[0063] The drying tower air flow optimization module 22 optimizes the air flow distribution in the spray drying tower by using the spray drying tower dynamic air flow control algorithm based on the adjusted milk slurry concentration, workshop environment data, and equipment operation data.

[0064] In this embodiment, the milk slurry multi-fractional variable-order concentration adjustment algorithm uses the variable-order differential neural operator to extract multi-scale perturbations in the milk slurry data and environment data, forms a 128-dimensional feature vector, and constructs a milk slurry fluctuation intensity index through the Moran multifractal mapping to reflect the non-linear intensity of physical property fluctuations such as milk slurry viscosity and temperature.

[0065] The spray drying tower dynamic air flow control algorithm uses a lightweight CFD neural operator to predict the three-dimensional velocity field in the tower, and extracts its topological structure to calculate the topological flux entropy.

[0066] The milk slurry multi-fractional variable-order concentration adjustment algorithm is based on the multifractal dynamic fluctuation modeling technology and the variable-order fractional differential response mechanism, and is used to dynamically adjust the milk slurry concentration before drying in real time. The specific method of the milk slurry multi-fractional variable-order concentration adjustment algorithm is as follows:

[0067] In this embodiment, the multifractal dynamic fluctuation modeling technology is a method for describing the fluctuation structure at multiple scales in a complex system, and uses a set of exponential spectra of power-law distributions to describe the perturbation signal.

[0068] The variable-order fractional differential response mechanism is a variable-order fractional differential response mechanism constructed on the basis of the traditional fractional Caputo derivative model, which further allows the derivative order to vary with time, and is used to dynamically describe a control system with a memory effect.

[0069] By combining the multifractal dynamic fluctuation modeling technology and the variable-order fractional differential response mechanism, a milk slurry multi-fractional variable-order concentration adjustment algorithm for intelligently perceiving and controlling the milk slurry concentration is constructed in the present invention.

[0070] S2.1.1. Input the milk slurry data and workshop environment data into the variable-order differential neural operator, calculate the quasi-variable-order quasi-fractal derivatives of all milk slurry data and workshop environment data , and combine and embed all the quasi-variable-order quasi-fractal derivatives to obtain a variable-order differential feature ;

[0071] Among them, in the quasi-variable-order quasi-fractal derivative , is the variable-order exponent, is the multi-scale quasi-fractal dimension, is the time; is the quasi-variable-order quasi-fractal derivative, which is used to represent that the milk slurry data and workshop environment data have a variable-order exponent and a multi-scale quasi-fractal dimension Evolution trend of characteristics;

[0072] Extract the multi-scale dynamic change patterns of the original milk slurry data and workshop environment data through a variable-order pseudo-fractional derivative, and construct a mathematical representation of non-linear perturbation, cross-scale fluctuation and instantaneous response characteristics as the perception basis of the concentration regulation system. The variable-order pseudo-fractional derivative is a high-order dynamic feature extraction tool that combines fractional derivatives and fractal differentials; and because variables such as milk slurry concentration have non-stationarity and environmental dependence, the milk slurry viscosity fluctuates smoothly at high temperatures and oscillates violently at low temperatures, and the concentration regulation response is slow when the environmental humidity suddenly rises. Fixed orders cannot accommodate changes in fluctuation characteristics under different physical processes, so variable orders are used ;

[0073] For each variable data, calculate the variable-order pseudo-fractional derivative and then consider their characteristic structures at different time windows and scales. Encode all the results of the variable-order pseudo-fractional derivative into a set of feature dimensions through a group of learnable embedding mappings, which are the variable-order differential features ;

[0074] In this embodiment, the variable-order differential features are 128-dimensional, specifically as follows:

[0075] The milk slurry data includes four types of data: milk slurry viscosity , milk slurry temperature , milk slurry concentration and lactose content ; The workshop environment data includes three types of data: workshop temperature , workshop humidity and air flow rate ; There are a total of 7 types of data, that is, the variable-order differential features take dimensions;

[0076] S2.1.2. Use Moran multifractal mapping to non-linearly fuse the variable-order differential features and calculate the milk slurry fluctuation intensity index ;

[0077] Moran multifractal mapping is a non-linear fusion method for measuring the intensity of multi-scale perturbation structures in complex systems. Based on the idea of multifractal spectrum analysis in statistical physics, map the local sharp changes and global structural differences in the variable-order differential features into a single-valued index representing the perturbation intensity of the system, which is the milk slurry fluctuation intensity index , specifically as follows:

[0078] ;

[0079] Among them, is the index of the variable-order and quasi-fractal derivative for each data; is the weight of the variable-order differential feature of the th data; is the Moran power exponent, which is used to amplify severe perturbations; is the perturbation amplitude of the th variable-order and quasi-fractal derivative of the data;

[0080] S2.1.3. Calculate the automatic adjustment control speed of the milk slurry based on the workshop environment data ;

[0081] S2.1.4. Use the variable-order Caputo derivative model to construct an adaptive milk slurry concentration control equation;

[0082] S2.1.5. According to the prediction result of the adaptive milk slurry concentration control equation, adjust the actual milk slurry concentration to the target milk slurry concentration in real time, and output the adjusted milk slurry concentration and the milk slurry concentration control residual , is an increment of a time period.

[0083] In the above S2.1.4, the adaptive milk slurry concentration control equation is the core equation of the milk slurry multi-variable-order concentration adjustment algorithm, which is constructed based on the variable-order Caputo derivative model and is used to dynamically predict the trend of the milk slurry concentration. The specific form of the adaptive milk slurry concentration control equation is as follows:

[0084] ;

[0085] Among them, is the Caputo variable-order fractional derivative, representing the dynamic change rate of the milk slurry concentration at time ; is the automatic adjustment control speed of the milk slurry; is the milk slurry concentration; is the target milk slurry concentration at time ; is the control gain coefficient; is the milk slurry fluctuation intensity index; is the fractal excitation gain coefficient.

[0086] In this embodiment, the Caputo derivative model is a fractional calculus tool used to describe dynamic processes with memory effects and non-locality. The variable-order Caputo derivative model further extends the Caputo derivative model, and its order is the automatic adjustment control speed of the milk slurry , which changes dynamically over time, indicating different degrees of response inertia or memory strength at different moments;

[0087] Under the control system of the milk slurry concentration, the milk slurry concentration disturbance is non-stationary. A fixed control coefficient cannot adapt to the characteristics of fluctuating quickly and slowly, and there is an inertial memory effect in the milk slurry control. Since the first derivative has no memory, a variable-order Caputo derivative model is used;

[0088] In the formula of the adaptive milk slurry concentration control equation, represents the steady-state attraction of traditional closed-loop regulation: if the current concentration deviates from the target concentration, it will be pulled back; represents the disturbance response enhancement term: if the system is currently in a strong disturbance environment, such as viscosity mutation, humidity fluctuation, etc., the regulation should be increased;

[0089] Finally, the adjusted milk slurry concentration is output and the milk slurry concentration control residual .

[0090] The dynamic air flow control algorithm of the spray drying tower is an air flow control algorithm based on persistent homology topological analysis and discrete differential manifold modeling, which is used to regulate the air flow distribution in the spray drying tower. The specific dynamic air flow control algorithm of the spray drying tower is as follows:

[0091] In this embodiment, persistent homology analysis is a computational topology data analysis technique in topology; in the spray drying tower, the three-dimensional air flow velocity field is a tensor field, which includes: complex forms such as local swirling eddy current structures, local air flow dead zones or fracture surfaces, and asymmetric air mass channels. Using persistent homology analysis, the following topological features can be extracted: the number of connected components, one-dimensional holes, and two-dimensional cavities; by calculating the difference between the birth time and the death time of different features, the topological complexity quantification index of this three-dimensional air flow velocity field can be obtained, which is the topological flux entropy;

[0092] Discrete differential manifold modeling transforms the continuous differential geometry definition into a numerical form that can be calculated on a discrete grid, and is used to deal with the evolution law of complex physical fields in the discrete space; in the spray drying tower, there are: conical lower necking geometry, multi-nozzle angle distribution, and non-uniform boundary heat transfer and reflection inside; these will all lead to the geometric non-linear propagation and local structural deformation of the air flow in the spray drying tower; compared with the traditional CFD which is a black box simulation, this method is more interpretable in structure;

[0093] By combining persistent homology analysis with discrete differential manifold modeling, a dynamic air flow control algorithm for spray drying towers is constructed. This algorithm can not only extract and predict the local structural characteristics of the air flow velocity field inside the spray drying tower, but also quantitatively measure the complexity and unevenness of the flow field from the perspective of the global topological structure, realizing the linkage prediction scheduling of the fan and nozzle control, and avoiding problems such as air flow dead zones, uneven heat distribution, or premature generation of lactose crystal nuclei;

[0094] S2.2.1. Input the adjusted milk slurry concentration along with the workshop environment data and equipment operation data into the lightweight CFD neural operator to construct a three-dimensional velocity vector field within the spray drying tower space ;

[0095] The lightweight CFD neural operator is more suitable for real-time feedback control compared to traditional CFD solvers. By using the lightweight CFD neural operator to construct the air flow velocity field at any position inside the spray drying tower, that is, the three-dimensional velocity vector field , the air flow information in any region of the drying tower space can be obtained in real time;

[0096] S2.2.2. Conduct persistent homology analysis on the three-dimensional velocity vector field and calculate the topological flux entropy ;

[0097] Regard the three-dimensional velocity vector field as a three-dimensional tensor field, analyze its topological structure characteristics through persistent homology, calculate the Betti numbers, including the number of connected components, one-dimensional holes, and two-dimensional cavities, and construct the topological flux entropy :

[0098] ;

[0099] Among them, is the topological structure feature index; is the topological structure feature in the proportion of the three-dimensional velocity vector field;

[0100] S2.2.3. According to the topological flux entropy , the adjusted milk slurry concentration , the lactose content and the milk slurry concentration control residual , calculate the adjusted fan speed and nozzle pressure;

[0101] S2.2.4. Based on the topological flux entropy and the workshop environment data, weighted update the automatic adjustment control speed of the milk slurry .

[0102] In this embodiment, based on the topological flux entropy Weighted update of the control speed of the milk slurry automatically according to the workshop environment data , specifically as follows:

[0103] ;

[0104] Among them, is the control speed reference value; is the temperature difference influence weight coefficient; is the humidity influence weight coefficient; is the topological entropy coupling coefficient; is the workshop temperature; is the milk slurry temperature; is the workshop humidity;

[0105] In the above S2.2.3, according to the topological flux entropy , the adjusted milk slurry concentration , the lactose content and the milk slurry concentration control residual , calculate the adjusted fan speed and nozzle pressure, and the specific method is as follows:

[0106] ;

[0107] ;

[0108] Among them, is the fan speed at the adjusted time ; is the nozzle pressure at the adjusted time ; is the influence weight of the fan change; is the weight of the milk slurry concentration deviation; is the coupling weight of the fan concentration adjustment residual; is the influence weight of the nozzle change; is the lactose deviation weight; is the coupling weight of the nozzle concentration adjustment residual; is the change rate of the topological flux entropy; is the adjusted milk slurry concentration; is the time of the target milk slurry concentration; Time of the lactose content; is the time of the target lactose content.

[0109] Lactose crystallization monitoring unit 3, which is used to monitor the lactose crystallization trend in real time;

[0110] In this embodiment, the lactose crystallization monitoring unit 3 is used to monitor the lactose crystallization trend in real time, specifically as follows:

[0111] Set the critical range of lactose content as , the critical milk slurry concentration fluctuation threshold and the critical milk slurry temperature fluctuation threshold ;

[0112] At time if the current lactose content is within the critical range of lactose content, the current milk slurry concentration fluctuation rate is greater than the critical milk slurry concentration fluctuation threshold, and the current milk slurry temperature fluctuation rate is greater than the critical milk slurry temperature fluctuation threshold, and the duration exceeds the manually set duration, the production work shall be immediately stopped.

[0113] The information management and traceability unit 4 is used to collect and store all data during the production process of milk powder and perform data report analysis;

[0114] In this embodiment, the information management and traceability unit 4 is used to collect and store all data during the production process of milk powder, including the original milk slurry data, workshop environment data and equipment operation data in the data acquisition and processing unit 1, and also including the adjusted milk slurry concentration, milk slurry concentration control residual, topological flux entropy and milk slurry automatic adjustment control speed in the milk slurry drying monitoring unit 2, and also including the set critical range of lactose content, critical milk slurry concentration fluctuation threshold and critical milk slurry temperature fluctuation threshold in the lactose crystallization monitoring unit 3.

[0115] Embodiment 2, the present invention proposes an intelligent management and scheduling method for a milk powder production workshop, which is used for an intelligent management and scheduling system of a milk powder production workshop in the above Embodiment 1, and includes the following steps:

[0116] S10.1. Use sensors and monitoring devices to collect milk slurry data, workshop environment data and equipment operation data in real time, and perform preliminary processing on all the collected data;

[0117] S10.2. Based on the milk slurry data and workshop environment data, adopt the milk slurry multi-variable order concentration adjustment algorithm and combine it with the spray drying tower dynamic air flow control algorithm to jointly adjust the milk slurry concentration in real time, optimize the air flow distribution of the spray drying tower and the equipment operation data;

[0118] S10.3. Monitor the lactose crystallization trend in real time;

[0119] S10.4. Collect and store all data during the production process of milk powder and perform data report analysis.

[0120] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited thereto, and various changes can be made without departing from the spirit of the present invention within the scope of knowledge possessed by those skilled in the art.

Claims

1. An intelligent management and scheduling system for a milk powder production workshop, characterized in that, Including: A data acquisition and processing unit (1) which uses sensors and monitoring devices to collect milk slurry data, workshop environment data, and equipment operation data in real time, and preliminarily processes all the collected data. A milk slurry drying monitoring unit (2) which, based on the milk slurry data and workshop environment data, uses a milk slurry multi-fractional order concentration adjustment algorithm and combines it with a spray drying tower dynamic airflow control algorithm to jointly adjust the milk slurry concentration in real time, optimize the airflow distribution in the spray drying tower, and equipment operation data. The milk slurry drying monitoring unit (2) includes a milk slurry concentration adjustment module (21) and a drying tower airflow optimization module (22). Among them, the milk slurry concentration adjustment module (21) is based on the milk slurry data and workshop environment data, uses a milk slurry multi-fractional order concentration adjustment algorithm to combine and dynamically adjust the milk slurry concentration before drying in real time, and outputs the adjusted milk slurry concentration and the milk slurry concentration control residual to the drying tower airflow optimization module (22). The drying tower airflow optimization module (22) is based on the adjusted milk slurry concentration, workshop environment data, and equipment operation data, and uses a spray drying tower dynamic airflow control algorithm to optimize the airflow distribution in the spray drying tower. A lactose crystallization monitoring unit (3) which is used to monitor the lactose crystallization trend in real time. An information management and traceability unit (4) which is used to collect and store all the data in the milk powder production process, and perform data report analysis.

2. The intelligent management and scheduling system for the milk powder production workshop according to claim 1, wherein, The milk slurry data includes: milk slurry viscosity , milk slurry temperature , milk slurry concentration and lactose content ; Workshop environmental data includes: workshop temperature , workshop humidity and air velocity ; The equipment operation data includes: the fan speed of the spray drying tower and the injection pressure ; wherein, is time.

3. The intelligent management and scheduling system for the milk powder production workshop according to claim 2, wherein, The milk slurry multi-fractional order concentration adjustment algorithm is realized based on the multi-fractal dynamic fluctuation modeling technology and the variable order fractional differential response mechanism, and is used to dynamically adjust the milk slurry concentration before drying in real time. The specific method of the milk slurry multi-fractional order concentration adjustment algorithm is as follows: S2.1.

1. Input the milk slurry data and workshop environment data into the variable-order differential neural operator to calculate the quasi-variable-order quasi-fractal derivatives of all the milk slurry data and workshop environment data and combine and embed all the quasi-variable-order quasi-fractal derivatives to obtain a variable-order differential feature ; Among them, the order-varying and fractal-like derivative wherein is the order-varying exponent, is the multi-scale fractal-like dimension, is time; is the order-varying and fractal-like derivative, which is used to represent the evolving trend of the characteristics of the milk slurry data and the workshop environment data with the order-varying exponent and the multi-scale fractal-like dimension ; S2.1.

2. Use Moran multifractal mapping to perform non - linear fusion on variable - order differential features and calculate the milk slurry fluctuation intensity index ; S2.1.

3. Calculate the automatic adjustment control speed of the milk slurry based on the workshop environment data ; S2.1.

4. Use a variable order Caputo derivative model to construct an adaptive milk slurry concentration control equation. S2.1.

5. According to the prediction results of the adaptive milk slurry concentration control equation, the actual milk slurry concentration is adjusted to the target milk slurry concentration in real time, and the adjusted milk slurry concentration is output and the milk slurry concentration control residual , is an increment of a time period.

4. The intelligent management and scheduling system for milk powder production workshop according to claim 3, wherein In the above S2.1.4, the adaptive milk slurry concentration control equation is the core equation of the milk slurry multi-fractional order concentration adjustment algorithm, which is constructed based on the variable order Caputo derivative model and is used to dynamically predict the milk slurry concentration trend. The specific adaptive milk slurry concentration control equation is as follows: ； Among them, is the Caputo variable-order fractional derivative, representing the milk slurry concentration at time of the dynamic change rate; is the automatic adjustment control speed of the milk slurry; is the milk slurry concentration; is the target milk slurry concentration at time ; is the control gain coefficient; is the milk slurry fluctuation intensity index; is the fractal excitation gain coefficient.

5. The intelligent management and scheduling system for the milk powder production workshop according to claim 4, wherein The spray drying tower dynamic airflow control algorithm is an airflow control algorithm based on persistent homology topological analysis and discrete differential manifold modeling, and is used to regulate the airflow distribution in the spray drying tower. The specific spray drying tower dynamic airflow control algorithm is as follows: S2.2.

1. Input the adjusted milk slurry concentration along with the workshop environment data and equipment operation data into the lightweight CFD neural operator to construct a three-dimensional velocity vector field within the spray drying tower space ; S2.2.

2. Perform persistent homology analysis on the three-dimensional velocity vector field and calculate the topological flux entropy ; S2.2.

3. According to the topological flux entropy and the adjusted milk slurry concentration , lactose content as well as the milk slurry concentration control residual , calculate the adjusted fan speed and nozzle pressure; S2.2.

4. Based on topological flux entropy and update the automatic adjustment control speed of the milk slurry weighted by the workshop environment data .

6. The intelligent management and scheduling system for the milk powder production workshop according to claim 5, characterized in that, In S2.2.3, according to the topological flux entropy , the adjusted milk slurry concentration , the lactose content and the milk slurry concentration control residual , the adjusted fan speed and nozzle pressure are calculated as follows: ； ； Among them, is the fan speed under the adjusted time; is the nozzle pressure under the adjusted time; is the adjusted time; is the nozzle pressure under the adjusted time; is the weight of the influence of fan change; is the weight of the deviation of milk slurry concentration; is the coupling weight of the residual of fan concentration adjustment; is the weight of the influence of nozzle change; is the weight of lactose deviation; is the coupling weight of the residual of nozzle concentration adjustment; is the change rate of topological flux entropy; is the adjusted milk slurry concentration; is the time; is the target milk slurry concentration at time; time; is the lactose content at time; is the time; is the target lactose content at time.

7. The intelligent management and scheduling system for the milk powder production workshop according to claim 6, wherein The lactose crystallization monitoring unit (3) is used to monitor the lactose crystallization trend in real time, specifically as follows: Set the critical lactose content range as , the critical milk slurry concentration fluctuation threshold and the critical milk slurry temperature fluctuation threshold ; At time If the current lactose content is within the lactose content critical range, the current milk slurry concentration fluctuation rate is greater than the critical milk slurry concentration fluctuation threshold, and the current milk slurry temperature fluctuation rate is greater than the critical milk slurry temperature fluctuation threshold, and the duration exceeds the manually set duration, then immediately stop the production work.

8. The intelligent management and scheduling system for the milk powder production workshop according to claim 7, wherein, The information management and traceability unit (4) is used to collect and store all the data in the milk powder production process, including the original milk slurry data, workshop environment data, and equipment operation data in the data acquisition and processing unit (1), and also includes the adjusted milk slurry concentration, milk slurry concentration control residual, topological flux entropy, and milk slurry automatic adjustment control speed in the milk slurry drying monitoring unit (2), and also includes the set lactose content critical interval, critical milk slurry concentration fluctuation threshold, and critical milk slurry temperature fluctuation threshold in the lactose crystallization monitoring unit (3).

9. An intelligent management and scheduling method for a milk powder production workshop, which is used for an intelligent management and scheduling system of a milk powder production workshop as described in any one of claims 1-8, and is characterized in that: Including the following steps: S10.

1. Use sensors and monitoring devices to collect milk slurry data, workshop environment data, and equipment operation data in real time, and perform preliminary processing on all the collected data; S10.

2. Based on the milk slurry data and workshop environment data, adopt the milk slurry multi-resolution concentration adjustment algorithm and combine it with the spray drying tower dynamic air flow control algorithm to jointly adjust the milk slurry concentration in real time, optimize the air flow distribution of the spray drying tower, and the equipment operation data; S10.

3. Monitor the lactose crystallization trend in real time; S10.

4. Collect and store all the data in the milk powder production process, and conduct data report analysis.

Images